Analysis and Modelling of the Energy Consumption

Research Collection

Doctoral Thesis

Analysis and Modelling of the Energy Consumption of ChemicalBatch Plants

Author(s): Bieler, Patric S.

Publication Date: 2004

Permanent Link: https://doi.org/10.3929/ethz-a-004830844

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.

ETH Library

https://doi.org/10.3929/ethz-a-004830844

http://rightsstatements.org/page/InC-NC/1.0/

https://www.research-collection.ethz.ch

https://www.research-collection.ethz.ch/terms-of-use

Diss. ETHNo. 15532

Analysis and Modelling of the Energy Consumption

of Chemical Batch Plants

A dissertation submitted to the

SWISS FEDERAL INSTITUTE OF TECHNOLOGY

ZURICH

for the degree of

Doctor of Technical Sciences

Presented by

PATRIC S. BIELER

Dipl. Chem.-Ing. ETH

Dipl. NDS ETHZ in Betnebswissenschaften

born July 28, 1975

citizen of Luzern (LU) and Giswil (OW)

accepted on the recommendation of

Prof. Dr. K. Hungerbühler, examiner

Prof. Dr. D. T. Spreng, co-examiner

Prof. Dr. A. Wokaun, co-examiner

Dr. U. Fischer, co-examiner

Zurich 2004

ISBN 3-906734-39-0

Dedicated to my parents

i

AcknowledgementsThis thesis is based on research performed at the Safety and Environmental

Technology Group of the Swiss Federal Institute of Technology (ETH) in Zu¬

rich between June 2000 and April 2004. The funding rose by the Swiss Federal

Office of Energy (SFOE) (Project No. 39592) is gratefully acknowledged.First, I would like to thank my supervisor Prof. Dr. K. Hungerbühler for giv¬

ing me the possibility to perform my research in his group. Moreover, he kindlysupported my decision of conducting a postgraduate study in industrial science

during my dissertation.

Thanks should go to Prof. Dr. A. Wokaun and to Prof. Dr. D. T. Spreng for

their assistance as co-examiners and for the interesting discussions and chal¬

lenging questions concerning my work throughout the time of my thesis.

Dr. U. Fischer receives a special "thank you" for advising and managing myproject. I will never forget all the good discussions - not only in the field of my

thesis but also concerning topics from all over the world - and his marvellous

support during the whole time of my dissertation.

I want to thank the industrial partner company, that I can not name for con¬

fidentiality reasons, for enabling me to perform my dissertation in a challengingindustrial environment and supporting me in all the work I had to do, the ques¬

tions I had to pose and the measurements I had to perform. A great thank

should also go to the many people working at the company. I had a very goodand enriching time in the industrial environment. I will never forget all the in¬

teresting discussions with the site and the plant manager, the foremen, the op¬

erators, the people in the service teams, the production chemists, the engineersand all the people in my environment. Many people supported me with tips of

how to conduct the measurements and gather the required information - for this

I would like thank a lot everybody. A special thank should be addressed to the

people that conducted autonomously several measurements for supporting me in

my work.

A special thank is also going to Dr. Chr. Blickenstorfer, who started the pro¬

ject of energy modelling of chemical batch plants in 1996 within the Safety and

Environmental Technology Group and to Dipl. Chem.-Ing. ETHD. Dahinden,who carried out his diploma thesis within this project. He made major contribu¬

tions to the Excel®-programming resulting in the Excel®-model of the whole

plant.I do not want to miss to thank all the people from the Safety and Environ¬

mental Technology Group at ETH for the good time we had - first in the goodold CAB building in the centre of the beautiful little big-city of Zurich and af¬

terwards in the HCl building on top of Hönggerberg.

My greatest thank, nevertheless, goes to my parents, I. and P. S. Bieler-

Schmed who supported me during all the intense time of my studies in Chemi¬

cal Engineering and Industrial Management and during my dissertation. With¬

out their support and their care I would not have been able to perform this bigworkload and to succeed in my education.

Finally, yet importantly, I would like to thank my girlfriend M. Oeschger for

her support and her patience during the final phase of my thesis.

m

Engineering is the science of economy, of conserving the energy, kinetic and

potential, provided and stored up by nature for the use of man. It is the business

of engineering to utilize this energy to the best advantage, so that there may be

the least possible waste.

William A. Smith, 1908

v

Table of Contents

1 Introduction 1

1.1 Context and Motivation 1

1.2 State of the Art 2

1.3 Thesis Statement and Contribution 3

1.4 Thesis Organization 4

2 Structure of a Batch Plant 7

2.1 The General Structure of Batch Plants 7

2.2 The Differences between Batch Plants 10

2.2.1 The Monoproduct Batch Plant 10

2.2.2 The Multiproduct Batch Plant 10

2.2.3 The Multipurpose Batch Plant 11

3 Two Approaches for Energy Modelling 13

3.1 The Top-Down Approach 13

3.1.1 The Model for the Production Dependent Energy

Consumption 13

3.1.2 The Heating Steam Model 14

3.2 The Bottom-Up Approach 15

3.2.1 Equations for Heating and Cooling of Substances 15

3.2.2 Equations for Electric Equipment 16

3.2.3 Unified Equation for the Bottom-Up Modelling 17

4 Top-Down Modelling of Production Plants (TODOMO) 23

4.1 The Basic Equation for the Top-Down Modelling 23

4.2 The Characteristics of the Different Buildings Investigated 23

4.3 Analysis of the Different Energy Carriers 24

4.3.1 Steam 25

4.3.2 Electricity 31

4.3.3 Cooling Energy 34

4.4 Applicability of the Models 36

4.5 Conclusions 38

5 Modelling of Single Unit Operations 41

5.1 Reactors 41

5.1.1 Description of the Equipment 41

5.1.2 Measurements 42

5.1.3 Model and Conclusions 45

5.2 Nutsche Dryer 54




vii

5.3 Heat-Chamber 58




5.4 Vacuum Pumps 61

5.4.1 General Vacuum Pumps 61

5.4.2 Anti Pollution Vacuum Pumps (APOVAC) 62

5.4.3 Steam-Jet Vacuum Pumps 65

5.5 Stirrers and Motors 66




5.6 Continuous Equipment 72

5.6.1 Infrastructure 72

5.6.2 Short-Path Distillation Column 73

5.6.3 Falling-Film Evaporator 76

5.7 Horizontal Vacuum Rotary Dryer 78




5.8 Batch Distillation Column 82




5.9 Centrifuge 86




5.10 Conclusions 88

6 Bottom-Up Modelling of Multipurpose Batch Plants 91

6.1 Combining the Different Unit Operation Models to a Plant

Model (BOTUMO) 91

6.1.1 Description of the Program for Modelling MultipurposeBatch Plants 91

6.1.2 Modelling and Report Generation 93

6.2 Results of the BOTUMO 95

6.2.1 Modelling of Different Periods 95

6.2.2 Analysis of the Energy Consumption of the Building 100

6.2.3 Modelling of Different Aspects of the Reactors and

Nutsche Dryers 104

6.2.4 The Differences between the Products 110

6.2.5 The Differences between the Apparatus 116

6.3 Sensitivity Analysis of the BOTUMO 118

6.3.1 Time 119

6.3.2 Steam Heat Transfer 120

6.3.3 Brine Heat Transfer 121

6.3.4 Condensation Enthalpy of Steam 122

viii

6.3.5 Stirrer Input to the Reaction Vessels 123

6.3.6 Stirrer Electricity Consumption 124

6.3.7 Circulation Pump 125

6.3.8 Vacuum Pumps 126

6.3.9 APOVAC 127

6.3.10 Short Path Distillation 128

6.4 Conclusions 129

7 Conclusions and Outlook 131

7.1 Conclusions 131

7.2 Outlook 136

Appendix I

A The Model I

A.l The Assumptions for the BOTUMO I

A. 1.1 The Assumptions for the Single Unit Operation Models I

A.l.2 The Assumptions for the Plant Model II

A.2 The Excel® Model Ill

A.2.1 The Sheets of the Program Ill

A.2.2 Description of the Required Input Data VII

A.3 The Results of the Model X

A.3.1 Modelling Results X

A.3.2 Sensitivity Analysis XXI

B The Measuring Equipment XXIII

B.l Steam Measurements XXIII

B.l.l The Equipment XXIII

B.1.2 The Accuracy XXIV

B.2 Measurement of Brine XXIV

B.2.1 The Equipment XXIV

B.2.2 The Accuracy XXV

B.3 Measurement of Electricity Consumption XXVIII

B.3.1 The Equipment XXVIII

B.3.2 The Accuracy XXVIII

C Miscellaneous XXIX

C.l Distributions of the Times Given in the PSP XXIX

C.2 Reflux Conditions XXXI

C.3 Investigations on the Cleaning of Vessels XXXIII

D Measurements XXXVII

D.l Measurements for the TODOMO XXXVII

D.2 Measurements for the BOTUMO XLIII

E Improvement Potentials for the Investigated Plant LV

F Glossary LIX

IX

List of Figures

Figure 1-1: Structure of the thesis 5

Figure 2-1: Value chain in the chemical industry (shaded: typical batch

processes) 7

Figure 2-2: Engraving of a 16th century gold processing plant (Stitt 2002) 8

Figure 2-3: General structure of a batch plant 9

Figure 3-1: The basic concept of the BOTUMO 18

Figure 3-2: The principle of the BOTUMO 19

Figure 4-1: Consumption of production steam (5 and 15 bar) of the different

buildings as a function of amount of products per month

(according to Equation (3-1)) 26

Figure 4-2: Consumption of heating steam (5 bar) as a function of degree-days per month (according to Equation (3-2)) 29

Figure 4-3 : Normalized heating steam consumption (5 bar) as a function of

the number of degree-days per month (according to

Equation (3-3)) 30

Figure 4-4: Electricity consumption (excluding electricity for cooling

purposes) of the investigated buildings as a function of the

amount of chemicals produced per month (according to

Equation (3-1)) 32

Figure 4-5: Hourly electricity consumption of Building 1 (without electric

heating of a specific process) during an ordinary week in 2001... 33

Figure 4-6: Consumption of cooling energy of the different buildings as a

function of production output per month (according to

Equation (3-1)) 35

Figure 4-7: Modelled monthly electricity consumption as a function of

capacity usage for those buildings where the model accordingto Equation (3-1) was suitable 37

Figure 4-8: Flowchart for energy analysis in chemical batch production 38

Figure 5-1 : Scheme of a standard batch vessel with its heating/cooling-

system 42

Figure 5-2: Example of the steam measurements for a 10 m3, glass lined

reaction vessel heated with 5 bar steam 433

Figure 5-3: Example of brine measurements for a 10 m stainless steel vessel .44

Figure 5-4: Measurements of the electric heating of the 4 m3 stainless steel

high-temperature reaction vessel 45

Figure 5-5: Modelling and measurements of the steam consumption of

reaction vessels 47

Figure 5-6: Modelling results of the steam consumption of a 10 m3 stainless

steel reaction vessel (in comparison with measured steam

consumption and reaction time) 48

Figure 5-7: Measurements of the brine consumption of a 10 m3 stainless

steel vessel (regression according to Equation (3-16)) 50

Figure 5-8: Modelling of the brine consumption of a 10 m3 stainless steel

vessel (according to Equation (3-16); in comparison with

measured steam consumption) 51

XI

Figure 5-9: Modelling of the brine consumption (according to

Equation (3-16)) vs. measurements 52

Figure 5-10: Scheme of a nutsche dryer with its heating/cooling-system 54

Figure 5-11 : Modelling according to Equation (5-1) and measurements of

the nutsche dryer with simultaneous heating and cooling 56

Figure 5-12: Modelling according to Equation (5-1) of the drying of

product 1 in a 10 m2 nutsche dryer (in comparison with

measured steam consumption and drying time) 57

Figure 5-13: Scheme of a heat-chamber 58

Figure 5-14: Measured and modelled (according to Equation (5-2)) steam

consumption and experiment duration for the heat-chamber 59

Figure 5-15: Measured vs. modelled steam consumption of the heat-chamber

(according to Equation (5-2)) 60

Figure 5-16: Typical measurement of the electricity consumption of a

vacuum pump (here: P/v= 16.5 kW) 61

Figure 5-17: Measurements of the electricity consumption of the APOVAC

pumps (P/v = 27 kW) 63

Figure 5-18: Measurements of the cooling media consumption of the

APOVAC pumps (calculation of the cooling energy

consumption according to Equation (3-5)) 64

Figure 5-19: Principle of a steam-jet vacuum pump (El-Dessouky et al.

2002) 65

Figure 5-20: Power consumption (P) and temperature of the reaction mass

(IT) for a 6.3 m3 stainless steel vessel with an Intermig stirrer 68

Figure 5-21: Power consumption (P) and temperature of the reaction mass

(IT) of a 6.3 m3 glass lined vessel with an Intermig stirrer 69

Figure 5-22: Power consumption (P), rounds per minute, and temperature of

the reaction mass (IT) of a 6.3 m3 stainless steel vessel with a

Cross-Blade stirrer 69

Figure 5-23: Measurements of the relation of power consumption to nominal

power P/Pn of different stirrer types 70

Figure 5-24: Scheme of a short path distillation column 73

Figure 5-25: Measurements of the total electricity consumption of the short

path distillation column 74

Figure 5-26: Measured brine consumption of the short path distillation

column 75

Figure 5-27: Scheme of a falling-film evaporator 76

Figure 5-28: Modelled energy consumption of a one-day operation of the

falling-film evaporator (according to Equation (5-1);

parameters see Table 5-5) 77

Figure 5-29: Typical horizontal vacuum rotary dryer with agitator beinginstalled into shell (from (Mujumdar 1995)) 78

Figure 5-30: Steam measurements of a 4 m3 horizontal vacuum rotary dryer ..

79

Figure 5-31 : Measured and calculated steam consumption for two 4 m3

horizontal vacuum rotary dryers (according to Equation (3-16)). 80

xii

Figure 5-32: Modelled steam consumption for the 4 m3 Dryer 1 dryingProduct 2 (calculated according to Equation (3-16); in

comparison with measured steam consumption and drying

time) 81

Figure 5-33: Scheme of abatch distillation column 82

Figure 5-34: Measured and calculated steam consumption for the

investigated batch distillation column (according to

Equation (5-3)) 84

Figure 5-35: Modelling results of the batch distillation column (according to

Equation (5-3)) in comparison with measured steam

consumption and distillation time 85

Figure 5-36: Scheme of a centrifuge 86

Figure 6-1 : The four layers of the program for modelling the energy

consumption of chemical batch plants 91

Figure 6-2: The different layers and the structure of the BOTUMO program

and their contents 92

Figure 6-3: Modelling of the specific utility consumption (per t of product)of the whole building for one day of production according to

Equation (3-14) (in comparison with measured consumptionand modelled data according to CPM) 96

Figure 6-4: Modelling of the specific utility consumption (per t of product)of the investigated building for one month of production

according to Equation (3-14) (in comparison with measured

consumption and modelled data according to CPM) 98

Figure 6-5: Absolute modelled steam consumption of the building duringone month according to Equation (3-14) (PSP data) 101

Figure 6-6: Specific modelled steam consumption of the building during one

month according to Equation (3-14) (PSP data) 102

Figure 6-7: Absolute modelled electricity and brine consumption of the

building during one month according to Equation (3-14)

(PSP data) 103

Figure 6-8: Specific modelled electricity and brine consumption of the

building during one month according to Equation (3-14)(PSP data) 104

Figure 6-9: Modelled specific steam consumption of the reactors and

nutsche dryers according to Equation (3-25) (PSP data) 106

Figure 6-10: Modelled specific electricity consumption for the reactors and


Figure 6-11 : Modelled specific brine consumption for the reactors and


Figure 6-12: Specific modelled steam consumption of the different products(A, B,..,N, O) according to Equation (3-20) and number of

synthesis steps (PSP data; modelling period: W = one week,M = one month) Ill

Xlll

Figure 6-13: Specific modelled electricity consumption of the different

products (A, B,..,N, O) according to Equation (3-20) and

number of synthesis steps (PSP data; modelling period:W = one week, M = one month) 113

Figure 6-14: Specific modelled brine consumption of the different products

(A, B,..,G, I) according to Equation (3-20) and number of

synthesis steps (PSP data; modelling period: W = one week,M = one month) 114

Figure 6-15: Modelled specific steam consumption of the apparatus (1,2,..,26, 27) during one month according to Equation (3-16)(PSP data) 116

Figure 6-16: Modelled specific electricity consumption of the apparatus (1,2,..,26, 27) during one month according to Equation (3-16)

(PSP data) 117

Figure 6-17: Modelled specific brine consumption of the apparatus

(1, 2,..,26, 27) during one month according to Equation (3-16)(PSP data) 118

Figure 6-18: Sensitivity analysis of the batch time t with regard to the

specific utilities according to Equation (3-14) (one month;PSP data) 119

Figure 6-19: Sensitivity analysis of the steam loss coefficient Kst with

regard to the specific steam consumption according to

Equation (3-14) (one month; PSP data) 120

Figure 6-20: Sensitivity analysis of the brine loss coefficient Kq0 with

regard to brine consumption according to Equation (3-14)(one month; PSP data) 121

Figure 6-21: Sensitivity analysis of the steam condensation enthalpy AHvwith regard to steam consumption according to


Figure 6-22: Sensitivity analysis of the stirrer input tj with regard to utilityconsumption according to Equation (3-14) (one month;PSP data) 123

Figure 6-23: Sensitivity analysis of the stirrer electricity consumption y

with regard to utility consumption according to


Figure 6-24: Sensitivity analysis of the circulation pump efficiency y with

regard to electricity consumption according to Equation (3-14)(one month; PSP data) 125

Figure 6-25: Sensitivity analysis of the vacuum pump efficiency y with

regard to electricity consumption according to Equation (3-14)(one month; PSP data) 126

Figure 6-26: Sensitivity analysis of the APOVAC pumps efficiency y with

regard to electricity consumption according to Equation (3-14)

(one month; PSP data) 127

Figure 6-27: Sensitivity analysis of the short path distillation motors

efficiency y with regard to electricity consumption accordingto Equation (3-14) (one month; PSP data) 128

XIV

Figure 7-1: Analysis of the total modelled steam consumption of the

investigated plant (period: one month; PSP data; total

consumption: 1,354 MWh; heat of reaction: -80 MWh, stirrer

input: -23 MWh) 133

Figure 7-2: Analysis of the total modelled electricity consumption of the


consumption: 315 MWh) 134

Figure 7-3: Analysis of the total modelled brine consumption of the


consumption: 100 MWh) 135

Figure A-1 : Equations on the input sheet Daten in the Excel® model Ill

Figure A-2: Equations on the sheet Auswertung in the Excel® model IV

Figure A-3 : Equations on the sheet Berechnungen in the Excel® model V

Figure A-4: Equations on the sheet Auflieizen in the Excel® model VI

Figure A-5: Equations on the sheet Verdampfen in the Excel® model VI

Figure B-1 : Measuring principle for the steam measurements XXIII

Figure B-2: Scheme of the Portaflow X ultrasonic flow meter from

Fuji Electric43 XXIV

Figure B-3: Principle of the brine measurement XXV

Figure B-4: Picture of a LEM Memobox 800 XXVIII

Figure C-l: Detailed investigations on production time distribution for

the production process (Steps 1, 2,.., 50, 51) of Product A in

two 4 m3 and one 10 m3 glass lined reaction vessels XXIX

Figure C-2: Time measurements of Product J in a 6.3 m3 glass-linedreactor XXX

Figure C-3: Time measurements of Product G in a 4 m3 glass-linedreactor XXX

Figure C-4: Frequency of the measured times of reflux condition (i.e.,30 min of reflux) in a 10 m3 stainless steel vessel given in

Table C-2 XXXIII

Figure C-5: Measurements for a dirty 6.3 m3 glass-lined reactor XXXIV

Figure C-6: Measurements for the same clean 6.3 m3 glass-lined reactor XXXIV

Figure C-7: Modelling and measured values of the dirty and clean 6.3 m3

glass-lined reactor XXXV

Figure C-8: Modelling results of the dirty and clean 6.3 m3 glass-linedreactor (in comparison with measured values and

experiment duration) XXXVI

Figure D-l: Efficiency of standard motors at different levels of power

consumption (BBC 1976) XLIII

XV

List of Tables

Table 2-1: Sections of a generic batch plant 9

Table 4-1: Characteristics of the investigated buildings 24

Table 4-2: Summary of the different production energy consumption models

obtained for the different energy forms (m) in the different

buildings according to Equation (3-1) 28

Table 4-3: Summary of the models for heating steam consumption obtained

for the different buildings according to the normalised

Equation (3-3) 31

Table 5-1: Calculated loss coefficients for the steam consumption of the

reaction vessels and nutsche dryers investigated 49

Table 5-2: Loss coefficients for the brine measurements of the investigatedreaction vessels 52

Table 5-3: Different kind of stirrers used in the investigated building 67

Table 5-4: Base Consumption of the investigated building 73

Table 5-5 : Parameter values of the falling-film evaporator 77

Table 5-6: Standard sizes of centrifuges 87

Table 5-7: Summary of the Equations and Parameters for the SUOM 88

Table 6-1 : Example of a generic PSP and its translation for the data input to

the program 94

Table 6-2: Investigated periods 95

Table 6-3: Relative deviations of the different modelling methods for the

investigated utilities according to Equation (3-14) 98

Table 6-4: Comparison of Measurements, TODOMO results according to

Chapter 4 and BOTUMO results according to Chapter 6.1 for

one month of normal production 100

Table 6-5: Summary of the sensitivity analysis of Chapter 6.3 showing the

deviation of the objective functions Em according to

Equation (3-14) for changes in the parameter values of ±20%;

modelling period: one month 130

Table A-1 : Required input data for the Excel® models (sheet Daten) VII

Table A-2: Required input data for the Excel® models (sheet Parameter) VIII

Table A-3 : Required input data for the Excel® models (sheet WK) VIII

Table A-4: Required input data for the Excel® models (sheet VP-DS-Dest.

Kol.) VIII

Table A-5: Required input data for the Excel® models (sheet Reaktionen) IX

Table A-6: Required input data for the Excel® models (sheet Substanzen) IX

Table A-7: Required input data for the Excel® models (sheet Geräte) IX

Table A-8: Measurement and modelling of the utility consumption of the

investigated plant X

Table A-9: Modelling results for the total utility consumption of the

investigated building XI

Table A-10: Modelled steam consumption of one week of the reactors and

nutsche dryers of the investigated building XI

XVII

Table A-11 : Modelled specific steam consumption of one week of the

reactors and nutsche dryers of the investigated building XII

Table A-12: Modelled electricity consumption of one week of the reactors

and nutsche dryers of the investigated building XII

Table A-13: Modelled specific electricity consumption of one week of the

reactors and nutsche dryers of the investigated building XIII

Table A-14: Modelled brine consumption of one week of the reactors and

nutsche dryers of the investigated building XIII

Table A-15: Modelled specific brine consumption of one week of the

reactors and nutsche dryers of the investigated building XIV

Table A-16: Modelled steam consumption of one month of the reactors and

nutsche dryers of the investigated building XIV

Table A-17: Modelled specific steam consumption of one month of the

reactors and nutsche dryers of the investigated building XV

Table A-18: Modelled electricity consumption of one month of the reactors

and nutsche dryers of the investigated building XV

Table A-19: Modelled specific electricity consumption of one month of the

reactors and nutsche dryers of the investigated building XVI

Table A-20: Modelled brine consumption of one month of the reactors and

nutsche dryers of the investigated building XVI

Table A-21: Modelled specific brine consumption of one month of the

reactors and nutsche dryers of the investigated building XVII

Table A-22: Percentage of utility consumption of the produced chemicals.. XVII

Table A-23: Modelled specific steam consumptions of the apparatus [kg/t]for one month XVIII

Table A-24: Modelled specific electricity consumptions of the apparatus

[kWh/t] for one month XIX

Table A-25: Modelled specific brine consumptions of the apparatus [kWh/t]for one month XX

Table A-26: Results of the different sensitivity analysis for one month XXI

Table B-l: Test of the steam measurement device XXIV

Table B-2: Parameters for the flow measurements of the two kinds of

brine XXVI

Table B-3: Temperature comparison of the two temperature probes XXVII

Table C-1 : Time investigations of PSP and measurements XXXI

Table C-2: Time measurements for distillation of 300 1 of 1-butanol in a

10 m3 stainless steel vessel XXXII

Table C-3 : Steam measurements for the cleaning investigations for a

6.3 m3 glass lined reactor XXXV

Table D-l: Measurements of Building 1 XXXVII

Table D-2: Measurements of Building 2 XXXVIII

Table D-3: Measurements of Building 3 XXXIX

Table D-4: Measurements of Building 4 XL

Table D-5: Measurements of Building 5 XLI

Table D-6: Measurements of Building 6 XLII

Table D-7: Efficiencies of standard motors at different levels of power

consumption (BBC 1976) XLIV

xviii

Table D-8: Measurements of the steam consumption of the reaction

vessels XLV

Table D-9: Summary of the brine measurements XLVIII

Table D-10: Batch times for the electric heating in a 4 m3 stainless steel

reaction vessel (high temperature) XLIX

Table D-l 1 : Measurements with simultaneous heating and cooling in

10 m2 nutsche dryers XLIX

Table D-12: Measurements without simultaneous heating and cooling in

10 m2 nutsche dryers L

Table D-l3: Measured power consumption of different vacuum pumps L

Table D-14: Steam and cooling water consumption of different steam-jetvacuum pumps (four stages) according to (GEA.b ) LI

Table D-l 5: Summary of the Brine Measurements for the APOVAC pumps... LI

Table D-l 6: Infrastructure Measurements of the investigated building LII

Table D-17: Measurements of the steam consumption of a batch distillation

column LII

Table D-l8: Steam measurements (15bar) for the high temperature reactor

(4 m3 stainless steel reaction vessel) LIII

Table F-l: Definitions of the ISA-S88.01-1995 standard for batch

production (ISA 1995) LIX

XIX

List of Abbreviations, Symbols, and Indexes

APOVAC

BOTUMO

Anti Pollution VACuum

BOTtom-Up MOdel

CPM Company Proprietary Method

Fl

F2

M

Flow meter

Flow meter

Month

PI Pressure meter

PR Production Record

PSP Process Step Procedure

SUOM Single Unit Operation ModelTl Temperature meter

T2 Temperature meter

TAM Time Average Model

TODOMO TOp-DOwn MOdel

TSM Time Slice Model

W Week

SymbolsA Surface Area [m2]ACR Air Change Rate [h"1]B Base consumption of energy [MWh / period]C Constant [various]c Sound velocity [m/s]

Cp Heat capacity [kJ / kg / K]DD Degree-Days [°C • d]DSS Day-specific steam consumption [MWh / °C • d]E Energy consumption [kWh / s]F Energy defining factor (0 for electricity,

steam and brine)

1 for [-]

K Loss coefficient [kW / m2 / K]IT Temperature of reaction mass [°C]or[K]InT Inlet Temperature [°C]or[K]m Mass [kg]OT Temperature ofjacket [°C] or [K]OuT Outlet Temperature [°C] or [K]P Power [kW]PO Production Output [t / period]RR Reflux Ratio [-]S Specific energy consumption [MWh /1]SC Steam consumption [MWh / period]t Time [s / period]T Temperature [K]AH Enthalpy change [kJ/kg]

XXI

y Actual to nominal power consumption of a motor [%]

r/ Efficiency [%]

p Density [kg / m3]v Kinematic viscosity [m / s]

Indexes

1, 2 Start- & EndpointA ApparatusAir Air

Am Ambient

B Barrel

BC Batch Column

Br Brake

C CrystallisationCo CoolingES Evaporated Solvent

El ElectricityF Feed

FFE Falling Film EvaporatorHC Heat-Chamber

HJ Heating Jacket

/ Infrastructure

/ Chemicals type (different PSP)j Apparatus typek Number of different specifications of a chemical (PSP)L Loss

M Meltingm Energy form

N Nominal

n Number of different specifications of a apparatusND Nutsche Dryer0 OperationP Production

Pu Pump

q Indicator for different process steps / unit operations of one recipe (PSP)R Reaction

RD Rotary DryerRM Reaction Mass

RV Reaction Vessel

S Solvent

So Solid

SPD Short Path Distillation column

St Steam

Su SuspensionV VaporisationW Water

Z Centrifuge

XXll

Abstract

Two different approaches for energy analysis and modelling of chemical

batch plants (a top-down model and a bottom-up model) were conducted in this

thesis. Steam, electricity and brine were the investigated utilities. Steam was

used for heating the reactors and the building. Electricity was used by diverse

electric equipment in the building. Brine was used for low-temperature coolingof the reaction vessels and the nutsche dryers (i.e., cooling below a starting tem¬

perature of about 30 °C).A top-down model (TODOMO) consisting of a linear equation based on the

specific energy consumption per ton of production output and the base con¬

sumption of the plant was postulated. This TODOMO showed to be applicablefor batch plants of the following kind:

Monoproduct batch plantsMultiproduct batch plants with constant production mix

Multipurpose batch plants in which only similar chemicals are pro¬

duced

The results showed that the electricity consumption of infrastructure equip¬ment was significant and responsible for about 50% of total electricityconsumption. Base consumptions for the steam and the brine system were onlyminor. The specific energy consumption for the different buildings was related

to the degree of automation and the production processes performed.For the heating steam, a model only depending on air change rate and de¬

gree-days was applicable.For multipurpose batch plants with highly varying production processes and

changing production mix, the TODOMO was not applicable and produced inac¬

curate results. A bottom-up model (BOTUMO) was postulated for these plants.The model consists of a production dependent part and a production-

independent part accounting for the infrastructure consumption. The productiondependent part actually consists of a term related to the chemicals, another term

related to the equipment, and a time-dependent loss term.

With the help of numerous measurements, different apparatus and unit op¬

eration models were built. These models use only easily accessible substance

and apparatus information and account for the losses of the different apparatus.The models are therefore designed for being transferable to other batch plantsand products and not limited to one specific plant. The single apparatus models

showed that losses for steam and brine consumption are high. The losses were

characterised by a so-called loss coefficient. This loss coefficient represents the

heat transfer coefficient of the outside surface area of the equipment. For steam

consumption, a loss coefficient of about 4.2-10"2 kW / m2 / K was found while

for brine consumption a loss coefficient of about 1.7-10"2 kW/ m2 /K was

found. More than 50% of the losses of the steam are therefore due to the heat¬

ing/cooling-system design with its steam traps.With the help of the above-mentioned equations, an Excel® model was built

for the modelling of whole production plants according to the BOTUMO.

Modelling of the whole production plant was performed for one and two days,

XXlll

one week and one month. The production data were taken from either the pro¬

duction record (PR) or the process step procedure (PSP). The modelling re¬

sulted in a high accuracy for the longer periods (PSP data is used as input).

Analyses of the modelling results for one month showed that the apparatus

group reactors and nutsche dryers is the most important energy consumer in the

building (apart from infrastructure consumption in the case of electricity). More

detailed analyses of the energy consumption of this apparatus group showed,that about 30 to 40% of steam energy are lost and thus large optimisation poten¬tials are revealed. For the electricity consumption, it is shown that the small cir¬

culation pumps of the heating/cooling-system of the reactors and nutsche dryersrequire about 25% of the total electricity consumption of this apparatus group

(i.e., ca. 50%) of the consumption of the stirrers) if no electric heating is per¬

formed. Electric heating is used for one single high-temperature reactor. The

consumption of this heating circuit is larger than the consumption of all other

stirrers in the building (over 25 stirrers).A sensitivity analysis showed that the batch time has the largest influence on

the energy consumption. Variations of ±50% in the batch time resulted in

changes in energy consumption of about ±30%.

Different saving potentials, ranging from elimination of reflux conditions to

invention of a new heating/cooling-system for a generic batch reactor, were

identified.

The applicability of the BOTUMO is shown for short and long modelling

periods using different types of input data. Transferability and applicability to

other buildings and chemicals need to be investigated in further case studies.

XXIV

ZusammenfassungIn dieser Doktorarbeit wurden zwei verschiedene Arten der energetischen

Analyse und Modellierung (ein top-down Modell und ein bottom-up Modell)von chemischen Batch Produktionsanlagen entwickelt. Die untersuchten Ener¬

gien waren Dampf, Elektrizität und Sole. Dampf wurde sowohl zur Erhitzungder Reaktoren und Nutschentrockner (und ihres Inhaltes), als auch zur Gebäu¬

deheizung benutzt. Die Elektrizität wurde von den verschiedenen elektrischen

Apparaturen der Gebäude verbraucht. Die Sole wurde zu Tieftemperatur-Kühlzwecken verwendet (d.h. Kühlen unterhalb einer Starttemperatur von ca.

30 °C).Ein top-down Modell (TODOMO), bestehend aus einer linearen Gleichung,

basierend auf dem spezifischen Energieverbrauch pro Tonne Produktionsaus-

stoss und dem Grundverbrauch des Gebäudes, wurde vorgeschlagen. Dieses

TODOMO ermöglichte die energetische Modellierung von folgenden Typenvon Batch Produktionsanlagen:

Monoprodukt Batch Betriebe

Mehrprodukt Batch Betriebe mit konstantem Produktemix

Mehrzweck Batch Betriebe in denen ausschliesslich ähnliche Che¬

mikalien produziert werden

Die Resultate zeigten einen signifikanten Elektrizitätsverbrauch der Infra¬

strukturanlagen auf (ca. 50%) des totalen Stromverbrauches). Der Grundver¬

brauch für Dampf und Sole war nur gering. Der spezifische Energieverbrauchder untersuchten Gebäude zeigte einen klaren Zusammenhang mit dem Automa¬

tionsgrad der Produktionsgebäude und den produzierten Chemikalien.

Für den Heizdampfverbrauch des Gebäudes wurde ein Modell entwickelt,welches nur vom Luftwechsel innerhalb des Gebäudes und von den Heizgradta¬

gen abhängig ist.

Für Mehrzweck Batch Betriebe mit stark unterschiedlichen Produktionspro¬zessen und schwankendem Produktemix war das TODOMO nicht anwendbar

und ergab ungenaue Resultate. Für diese Betriebe wurde ein bottom-up Modell

(BOTUMO) postuliert und entwickelt. Das Modell besteht aus einem produkti¬

onsabhängigen Teil und einem batchzeitunabhängigen Grundverbrauchsteil. Der

produktionsabhängige Teil besteht aus einem von den Chemikalienspezifikatio¬nen abhängigen Term, einem von den Apparatespezifikationen abhängigenTerm und einem zeitabhängigen Verlutsterm.

Durch diverse Messungen konnten Einzelapparate- und Einzeloperations¬modelle entwickelt werden. Diese Modelle benötigen ausschliesslich einfach zu

bestimmende Substanz- und Apparatedaten und modellieren zudem die Verluste

der verschiedenen Apparate. Die Modelle wurden so entwickelt, dass sie sich

einfach auf andere Betriebe und Chemikalien übertragen lassen und nicht auf

einen spezifischen Betrieb beschränkt sind. Bereits aus den Einzelapparate¬modellen ging hervor, dass die Verluste für Dampf- und Soleverbrauch signifi¬kant waren. Die Verluste wurden durch einen Verlustkoeffizienten charakteri¬

siert. Dieser Verlustkoeffizient beschreibt den Wärmeübergangskoeffizienten an

der Aussenfläche eines Apparates. Für den Dampfverbrauch wurde ein Ver-

XXV

lustkoeffizient von 4.2-10"2 kW / m2 / K und für den Soleverbrauch ein solcher

von 1.7-10"2 kW / m2 / K gefunden. Hieraus kann geschlossen werden, dass über

50%) des Verlustes beim Dampf auf das Heiz/Kühlsystem mit seinen Kondensa-

tableitern zurückzuführen sind.

Zur Modellierung des Energieverbrauches ganzer Produktionsgebäude mit

Hilfe des BOTUMO wurden die oben erwähnten Gleichungen in ein Excel®

Modell integriert. Dieses Modell wurde zur Modellierung des Energieverbrau¬ches des ganzen Produktionsgebäudes für einen und zwei Tage, eine Woche,sowie einen Monat verwendet. Die Modellrechnungen zeigten sehr gute Ge¬

nauigkeiten für die Modellierung von längeren Perioden (mit Hilfe der PSP Da¬

ten).Analysen über die Periode von einem Monat zeigten, dass die Apparate¬

gruppe Reaktoren undNutschentrockner den wichtigsten Energieverbraucher im

untersuchten Gebäude darstellt (neben dem Infrastrukturverbrauch bei der Elek¬

trizität). Detailliertere Analysen dieser Apparategruppe zeigten, dass ca. 30-40%>

des Dampfverbrauches für Verluste aufgewendet werden musste. Dies weist auf

grosse Optimierungspotenziale hin. Beim Elektrizitätsverbrauch konnte gezeigtwerden, dass die kleinen Umwälzpumpen der Heiz-/Kühlsysteme der Reaktoren

und Nutschentrockner ca. 25% des gesamten Elektrizitätsverbrauches dieser

Apparategruppe benötigen (d.h. ca. 50% des Verbrauches der Rührwerke),wenn die elektrische Heizung nicht läuft (sonst entsprechend weniger). Die

elektrische Heizung wird für einen einzelnen Hochtemperaturreaktor benötigt.Der Verbrauch dieser Heizung ist grösser, als der Verbrauch aller im Betrieb

vorhandenen Rührwerke (über 25).Eine Sensitivitätsanalyse wurde durchgeführt und zeigte, dass von allen un¬

tersuchten Parametern, die Batchzeit den grössten Einfluss auf den gesamten

Gebäudeenergieverbrauch hat. Eine Variation der Batchzeit um ±50% resultier¬

te in einer Veränderung des Gesamtenergieverbrauches von ±30%.

Verschiedene Einsparpotenziale wurden gefunden. Diese reichen von der

Elimination von Rücklaufbedingungen bis zu einem völlig neuen Design für die

klassischen Heiz/Kühlsystème.Die Anwendbarkeit des BOTUMO wurde sowohl für kurze als auch für lan¬

ge Zeitabschnitte gezeigt. Die verschiedenen Zeitabschnitte wurden mittels un¬

terschiedlicher Genauigkeiten der Eingabedaten modelliert. Die Übertragbarkeitauf andere Produktionsgebäude und Chemikalien muss in zusätzlichen Fallstu¬

dien untersucht werden.

XXVI

Introduction

1 Introduction

1.1 Context and Motivation

About 50%) of industrial processes (Stoltze et al. 1995) and chemical

production (Phillips et al. 1997) worldwide are batch processes.

Energy consumption of production processes contributes significantly to

overall resource use. The fewer resources the production of a substance (orfunctional unit) uses, the more environmentally friendly the process is (assum¬ing that all other parameters remain constant). Moreover, about 75% of man-

made air pollution is caused by energy use (Wang and Feng 2000). Therefore,minimization of energy consumption is listed as the sixth principle of green

chemistry (Anastas and Warner 1998).The chemical industry is a large, and in certain sectors, intensive user of en¬

ergy. For example, U.S. chemical industry accounted for about 20% of total

manufacturing primary energy consumption in 1994 (i.e., about 5.4 EJ) as stated

in (DOE 2000; Worrell et al. 2000). This value is even greater, if oil and gas

feedstock were included. The US chemical industry sets in their "Vision 2020"

the clear target to reduce energy consumption of chemical production and to

improve energy efficiency (ACS 1996; Eissen et al. 2002).

Energy consumption of plants engaged in continuous chemical productionshas been investigated extensively in the past by pinch technology (Linnhoff

1993). Similar methods for batch production are not yet well established. Fur¬

thermore, such studies are usually limited to heat-integration (Bouhenchir et al.

2001; Kemp and Macdonald 1988) and therefore rely on available storage ca¬

pacity or constant production schedules. Other studies account for time-varying

temperatures (Vaselenak et al. 1986) and rescheduling (Vaselenak et al. 1987).The use of these methods in batch production is limited because most of them

are considered as too complicated, lengthy, demanding and complex to be of

practical interest for most of the cases encountered (Stoltze et al. 1995). The

fact that energy costs amount to about 5 to 10%> of total production costs for

common chemicals produced in batch operation (Vaklieva-Bancheva et al.

1996) limits the efforts undertaken in achieving high energy efficiency. A help¬ful overview of energy consumption and management in batch production is

provided in (Grant 1996).Reliable statements on energy efficiency and improvement potentials of

production processes need standardized parameters characterizing energy con¬

sumption. It is only reasonable to set energy targets if the relation between the

actual and the minimal practical energy consumption is known. In multiproductand in multipurpose batch plants, this energy consumption has to be allocated to

different products and unit operations. Focus may then be put on the greatest

saving potential of the largest energy consumers. This prevents a wasting of the

limited resources for re-engineering by using them for the most effective savingpotentials.

A survey on the chemical industry in the U.K. showed that, on average for

different chemical branches, the most energy is used for process heating (40%>),

1

State of the Art

with distillation (13%>), drying (10%>) and compression (10%>) being the other

major energy-consuming unit operations (Anonymous 1986b).

Energy models for multiproduct and multipurpose batch plants are lackingin industry. It is known that energy consumption is, to some extent, related to

production output, but exactly where energy is used is not known. Whether the

dependence on production output is strong or whether the base load consump¬

tion of a building is dominating is not known. Energy consumption models on

building level are needed for providing consumption forecasts to the energy

supplier and for calculating total production costs.

1.2 State of the Art

Many papers, models and theories of the past and present research have

dealt with energy modelling of continuous processes as stated in (Linnhoff1993; Worrell et al. 2001; Zalba et al. 2003) or heat exchanger networks

(Furman and Sahinidis 2002; Gundersen and Naess 1988; Jezowski 1994a; Je-

zowski 1994b; Zhao et al. 1998). Batch production is hereby most of the time

neglected or the models are considered as too complex for industrial use (Stoltzeet al. 1995). Nevertheless, much literature is available on scheduling of batch

plants, which allows a more efficient use of energy by reducing waiting and

changeover times (see e.g., (Calderôn et al. 2000; Papageorgiou et al. 1994;Reklaitis et al. 1997; Sahinidis et al. 1989; Suhami and Mah 1982; Verwater-

Lukszo 1996; Vin and Ierapetritou 2000)).A novel approach named as Time Average Model (TAM) or Time Slice

Model (TSM) is introduced by (Linnhoff et al. 1988) and further used by several

authors (e.g., (Krummenacher 1997; Stoltze et al. 1995; Zhao et al. 1998)).Both the TAM and the TSM adapt the concept of pinch analysis introduced by(Linnhoff et al. 1982) to batch processes. The TAM assumes that all batch op¬

erations can be performed at any time and in any order, so that no account is

given to scheduling or time availability of energy flows. The time dependent

consumption of a batch reactor is averaged over the whole batch time for one

process resulting in a mean consumption for the whole process. In other words,time is completely ignored as a constraint and the energy source and sink values

become averaged over a chosen period. This results in a model similar to con¬

tinuous processes that can be handled by pinch analysis. This model is easy-to-

use but has, nevertheless, not much in common with the real behaviour of batch

production and is therefore of no significant practical use. The TSM, on the

other hand, does incorporate assumptions about time, e.g., cycle times and time

availability. Time is then 'sliced' into periods during which process energy

flows can be analysed and a separate model is calculated for each slice. For

each of these slices, energy consumption is again analysed as an average con¬

sumption over the whole time of the slice. Both the TAM and the TSM, never¬

theless, have no wide acceptability in industry. Furthermore, they have not been

applied to different energy carriers (only examples for steam are available) and

different products and processes in one unified model.

Some authors mention that significant savings of energy cost (and consump¬

tion) in batch plants of up to 25%> are possible (Allen and Shonnard 2002;Ashton 1993; Benz 2003; Krummenacher et al. 2002; Phillips et al. 1997;

2

Introduction

Rumazo et al. 2000). (Jiménez-Gonzâlez and Overcash 2000) state, that espe¬

cially energy challenging in early process phases reduces the level of emissions

during the whole lifecycle of the product. In this paper, energy lifecycle infor¬

mation is developed to support the decision-making process.

Besides these detailed papers mentioned above the basic concept of energy

audit is essential for performing an energy analysis of a whole production plant.The concept of energy analysis is widely discussed in literature; some examples

may be found in (Bhatt 2000a; Bhatt 2000b; Ganji 1999; Haman 2000; Hoshide

1995; Robert and Markus 1994). (Blickenstorfer 1999) provides a good over¬

view of literature dealing with energy analysis.No models are available in the literature to compute the energy consumption

of batch processes, accounting for the consumption caused by the chemical

process itself, the consumption due to the equipment and especially the losses of

the different systems. This will be investigated and analysed in this thesis (seethe next chapter).

1.3 Thesis Statement and Contribution

Energy consumption plays an important role in today's business since most

of the processes are not possible without an appropriate energy source (Kürsten

1996). Allocation to different processes and products is, nevertheless, often not

possible for batch production. As stated in the preceding chapter, energy con¬

sumption contributes quite significant to production costs and to environmental

hazard in the producing industry. Nevertheless, accurate and ready-to-use tools

for predicting or modelling the energy consumption of chemical batch plants are

missing. Goals for energy savings or targets for focusing on improvement po¬

tentials are most of the time set according to common (engineering) sense or

political targets. This is, contrary to continuous production processes, where

detailed models for energy consumption and integration methods are available,an unsatisfying situation. Moreover, legislation needs tools to predict the en¬

ergy saving potentials of plants to meet the goals set (see e.g.,

(Eidgenossenschaft 1999)) and the Kyoto protocol (see (http://unfccc.int/

resource/docs/convkp/kpeng.html ) for the text of the protocol and (Râsonyi2002; Thöne and Fahl 1998; Würsten 2003) for some comments). The goals set

in C02-legislation as mentioned in (Eidgenossenschaft 1999), lead to voluntarysavings and agreements of objectives with industry as mentioned in (BFE2001a; BFE 2001b) and in (BFE 2002). To succeed in these agreements of po¬

tential savings, detailed models for energy consumption are required. Without

such models, it would not be possible to control whether or not the goals are

achieved.

For all these reasons, easy to use tools should be available for energy model¬

ling of chemical batch production plants. A former thesis by (Blickenstorfer1999) showed the possibility of energy modelling on building level for a spe¬

cific kind of batch production (top-down approach for one kind of batch produc¬tion plant as discussed in Chapters 3.1 and 4 below). Applicability of this ap¬

proach to other buildings will be investigated in this thesis.

In this thesis, easy-to-use and adaptable single unit operation models

(SUOM) on apparatus level are developed. The new approach of the thesis of-

3

Thesis Organization

fers the possibility to model the energy consumption of a complete production

plant with a detailed bottom-up model based on the SUOM with the help of eas¬

ily accessible data. The required data consists of apparatus specifications,

building infrastructure consumption, specifications of the chemicals and the

production processes as well as operation times from the process step procedure

(PSP). With the help of this model, it is possible to gather information on the

energy consumption of a specific batch plant with a minimal of surplus meas¬

urements and data requirements. The data may be aggregated for different lev¬

els of analysis, as the user likes.

The applicability, usability, and accuracy of such models will be investi¬

gated in this thesis. The models should be simple enough to be useable in daily

production and accurate enough to analyse the energy consumption of a produc¬tion plant in detail. Such models would help legislation and particularly the

production chemists and plant management to analyse and in a second step op¬

timise the energy consumption of their production plants.

1.4 Thesis Organization

Figure 1-1 shows the structure of the thesis with its different chapters. The

chapters are organized to provide first the theoretical background and show the

postulated models for modelling of the energy consumption of chemical batch

plants and show afterwards the measurements, modelling results and outcomes

of the investigations of this thesis.

Chapter 1 summarises the literature in the field of the thesis, gives the the¬

sis statement and shows how the thesis results integrate in the literature known

so far. In Chapter 2, the structure of batch plants is described. The different

types of batch plants and their characteristics relevant for the investigations of

this thesis are discussed. The two models (i.e., top-down and bottom-up model)are introduced in Chapter 3. The generic modelling equations for the two ap¬

proaches are given and shortly discussed. The general principles, definitions,and usage of the single unit operation models are presented as well as the differ¬

ent possibilities and levels of energy modelling and analysis of whole produc¬tion plants according to the bottom-up approach. The measurements, modellingresults, and model applicability for the top-down model are presented in Chap¬ter 4. The characteristics of the different investigated buildings are discussed in

terms of their influence on the applicability of a top-down model to these plants.In Chapters 5 and 6, the bottom-up model is discussed in detail. Chapter 5

introduces the specific single unit operation models and shows the measure¬

ments performed for investigating the required parameters for the models pre¬

sented in Chapter 3. Chapter 6 combines the specific single unit operationmodels developed in Chapter 5 and the infrastructure consumption to a model of

the whole plant according to the general equations presented in Chapter 3. The

model is discussed and the different levels of analysis are presented according to

the equations given in Chapter 3. Sensitivity analyses of the different model-

parameters are given and discussed as well. In Chapter 7, the different results

of the thesis are summarised and a short outlook on future work and open ques¬

tions is given.

4

Introduction

1 Introduction

*

2 Description of the Structure of Batch Plants

*

3 Description of the Top-Down and the Bottom-Up Model

Bottom-Up Modelling

5 Single Unit Operations |

6 Whole Plants ]

7 Conclusions & Outlook

Appendix

Figure 1-1: Structure of the thesis

i®The Appendix presents the model assumptions, the actual Excel model of

the whole plant and the modelling results in Chapter A. Chapter B in the

Appendix presents the measuring equipment and discusses the accuracy of the

different measurements performed. In Chapter C in the Appendix, the results

and measurements for different special investigations performed during the the¬

sis are given. These investigations include the distribution of the batch times,the investigations concerning the reflux conditions, and investigations on the

influence on the cleaning of a vessel on its energy consumption are presented.

Chapter D in the Appendix presents the results of the measurements performedon apparatus and building level during the thesis. Chapter E in the Appendixsummarises the main improvement potentials for a batch plant and gives a

checklist for daily use where energy may be saved in daily operation. In Chap¬ter F in the Appendix, a small glossary of the terminology of batch operations is

provided.

5

Structure of a Batch Plant

2 Structure of a Batch Plant

2.1 The General Structure of Batch Plants

An introduction to the terminology of batch production is provided in (ISA1995) and in (Blickenstorfer 1999). The most important definitions for the pur¬

pose of this thesis and of batch production in general are given in Table F-l.

A batch plant cannot be operated by itself. Many different processes, plantsand operations have to be performed before a raw material enters the plant and

after a substance (product) leaves the plant. A generic value chain of a chemical

production is depicted in Figure 2-1. Basic chemicals like crude oil are ex¬

tracted from nature, transformed and upgraded to intermediate chemicals that

are the required raw materials for the pharmaceutical and fine chemical indus¬

try. These intermediate chemicals are most of the time produced with continu¬

ous processes in large amounts. Fine chemicals on the other hand, are high-value, low-tonnage products. These products are often produced in batch proc¬

esses to maintain the flexibility and efficiency of low production amounts. For

a general overview of the chemicals produced in batch production, see

(Parakrama 1985) or (Anonymous 2001). The same is true for the upgrading(i.e., formulating and mixing) of the fine chemicals. This is often done with the

help of batch processes as well. The final industrial application and the end us¬

ers often use batch processes too for their purposes. Therefore, batch processes

are of high interest. Because of the difficulties related to the modelling of batch

processes and the high prices often achieved on the market (compared to the

total production costs), energy optimisation was only a minor issue so far. To¬

day, prices of the fine chemicals are decreasing, production and raw material

costs are increasing (i.e., decreasing margins). Moreover, environmental legis¬lation gets stricter and energy consumption is sanctioned (see e.g., (Burkhardt2002; Eidgenossenschaft 1999; Ewers 2000; Gundersen 1991; Râsonyi 2002;Würsten 2003)). Therefore, the importance of minimising energy use is increas¬

ing. Moreover, modelling is required to declare and check the voluntary agree¬

ments of objectives for energy-savings in industry as mentioned in (BFE 2001a;BFE 2001b).

Natural

Products

Basic

Chemicals

Intermediate

Chemicals

Industrial

Applications

End

Users

Figure 2-1: Value chain in the chemical industry (shaded: typical batch proc¬

esses)

The shape of a batch reactor has little changed for the last 500 years. The

stirred tank has remained the same from the alchemist's time until today, al¬

though new concepts are available and propagated today (e.g., micro-reactors

etc. as described in (Höller and Renken 2000; Stitt 2002)). The uncanny resem¬

blance between a 16th century gold plant depicted in Figure 2-2 and a modern

fine chemicals plant, with both being dominated by the stirred tank reactor, has

been noted by (Stankeiwicz and Moulijn 2000).

7

The General Structure of Batch Plants

Figure 2-2: Engraving of a 16 cen¬

tury gold processing plant (Stitt 2002)

A batch plant (i.e., area, see Table F-l) usually consists of several parts, as

depicted in Figure 2-3 and Table 2-1. The batch production equipment repre¬

sents the heart of the batch plant (i.e., batch reactors, batch dryers, nutsche fil¬

ters, etc.). In this equipment, the process input is transformed to the process

output (i.e., the actual value is added to the product).Another part of a batch plant consists of so called special equipment. This is

equipment with special features, not common to the usual batch reactor like

high-temperature devices, continuous equipment such as distillation columns for

solvent recovery or continuous drying equipment, or equipment for filling and

packaging. This equipment is, in contrast to the batch production equipment,

very different from plant to plant depending on the kind of process output of the

plant.The production infrastructure is required for specific processes. Equipment

like circulation pumps, vacuum pumps, etc. could fall in this category. These

apparatus are not operated continuously for the whole building but specific for

one or the other process.

The final part of a batch plant is represented by the building infrastructure.This infrastructure consists of heating and ventilation systems, general vacuum

systems, waste-air treatment, etc. All equipment units that cannot be allocated

to one specific process and that are therefore operated continuously or stepwiseare considered as building infrastructure for the purpose of this thesis.

8


Batch Plant

Special Equipment

Building-Infrastructure

Production-Infrastructure

3] 3Contractors

Storage and Recovery of Solvents

£ £Contractors

Utilities

(Electricity, Steam,

Brine, etc.)

Figure 2-3: General structure of a batch plant

In general, allocation of the total energy consumption in such a building to

its different parts is unknown, i.e., it is not known which part of a batch plant is

the largest energy consumer and where savings would be most effective. There¬

fore, this thesis should provide industry with a tool for a fast and easy allocation

of energy consumption in batch production plants.

Table 2-1: Sections of a generic batch plant

Section Description Equipment ExamplesBatch- Standard unit-operations Reactors;Production Distillation columns;

Crystallisers; Nutsche

dryers

Special Dedicated equipment used for special High-temperature

Equipment purposes or less common equipment equipment;Continuous distillation

columns

Production Infrastructure needed for production but Vacuum systems;Infrastructure not related to one specific production Waste air treatment

process (absorbers, ventilation

system)

Building Infrastructure not necessarily needed for Space heaters;Infrastructure production but required to improve

workplace conditions

Lights

The Differences between Batch Plants

The different utilities (e.g., steam, electricity) required in a production build¬

ing are most of the time produced externally in a central facility for a completesite. Typically, cooling media production is an exception from this rule of cen¬

tralized production. The term cooling medium, as used in this thesis, stands for

ice or low-temperature fluids like brine (i.e., cooling water is not investigatedbecause of the lack of measurements). Cooling media production is mostlydone directly in the specific plant because decentralized production of coolingmedia is efficient and transportation losses would be significant in centralized

production.The recovery and storage of spent solvents is done either within the batch

plant or by a contractor. Large equipment is required for this purpose. This

equipment is considered independently (decoupled from batch-production). Op¬timisation of the regeneration operation can thus be done independently as well.

2.2 The Differences between Batch Plants

In batch production, different kinds of batch plants can be differentiated:

The monoproduct batch plantThe multiproduct batch plantThe multipurpose batch plant

The characteristics of these different plants will be discussed shortly in the

following sections.

2.2.1 The Monoproduct Batch Plant

A monoproduct batch plant is a plant that is designed especially for the pro¬

duction of one specific chemical. It is a dedicated plant with fix installation.

The path of an amount of raw material through the plant is clearly defined. No

or minimal manual operation is usually required since automation is elaborated

and recipes are seldom changing (if cheap labour is available, degree of automa¬

tion may be low as well). In an automated plant, data availability is most of the

time good. Because of the constant production steps, focus is given to optimisa¬tion of the production process (e.g., energy savings by heat integration(Krummenacher 1997; Krummenacher and Favrat 2001)).

2.2.2 The Multiproduct Batch Plant

A multiproduct batch plant is a plant where different chemicals are produced

throughout the year, but the same production steps are mostly performed in the

same equipment (see (Rippin 1992) as well). Fixed lines of batch reactors are

producing different products (probably different products on one line at differ¬

ent times of a year). The amounts of the different products may vary with sales

requirement. Therefore, production mix may not stay constant and may have an

influence on scheduling and utility requirements. Each line in a multiproductbatch plant may be considered as a (small) monoproduct batch plant (i.e., fixed

material pathways, potential for specific optimisation) for each production pe¬

riod.

10


2.2.3 The Multipurpose Batch Plant

A multipurpose batch plant, on the other hand, produces different chemicals

like the multiproduct batch plant, but in each equipment unit, different produc¬tion steps might be performed (i.e., such plants are characterized by high flexi¬

bility; (Rippin 1992)). The units are most of the time independent from each

other and connected via (flexible) pipes. This allows a construction of a produc¬tion path for the purposes of one specific chemical, each time this chemical may

be produced in the plant in a different way (i.e., in different reaction vessels).The pathway of a chemical in the monoproduct and the multiproduct batch plantis most of the time from top to bottom for reasons of ease of transportation (i.e.,

gravitation is helping to transport the chemicals). In multiproduct batch plants,this is probably considered as well, but not necessarily, because this would re¬

strict the flexibility of the plant.No or few dedicated equipment can be found in a multipurpose chemical

batch plant. This implies that all the equipment items are capable to perform all

possible unit operations and limits the optimisation potential.The infrastructure part of the multipurpose batch plant may also differ from

the infrastructure of the other two kinds of batch plants. Because of the multi¬

purpose characteristic of these plants, the infrastructure is not optimised for one

specific use. It is tried to operate as few infrastructure equipments as possible

(cost savings) but to install the equipment as flexible as possible. This allows

producing many different products. If a new product with new infrastructure

requirements is introduced to the plant, the new infrastructure equipment has to

be integrated in the former concept. This opens doors for oversizing and ineffi¬

ciencies in a specific production campaign but ensures the flexibility of the

plant.

11

Two Approaches for Energy Modelling

3 Two Approaches for Energy Modelling

For the modelling of energy consumption, two basic models are traditionallyfound in literature (see e.g., (Aebischer et al. 1988; Kubier 1986)): a "Top-Down"-model (TODOMO) based on measurements of a complete system and a

"Bottom-Up"-model (BOTUMO) based on diverse measurements of single

parts of a system and summation of the single energy consumers (as stated e.g.,

in (Werbos 1990)). Both models were adapted, elaborated, and investigated in

this thesis for the specific purposes of the chemical industry.The purpose of the two models in this thesis is to model and allocate energy

consumption of (chemical) batch plants. The time horizon will be no shorter

than one day. This limitation was set, because the short-term modelling would

require clumsy integral equations that would need many input parameters usu¬

ally unavailable in production business. Moreover, the important period for a

production plant is one week or even one month. For those periods, accountingof the production output is available and contractors bill the energy consump¬

tion based on the consumption during such periods.The following two subchapters introduce the modelling concepts of the

TODOMO and the BOTUMO together with the equations of these models.

3.1 The Top-Down Approach

3.1.1 The Model for the Production Dependent Energy Consump¬tion

For each utility, a model that computes the energy consumption of a build¬

ing as a function of the specific consumption per ton of product output and the

base consumption was postulated. The equation for the TODOMO is repre¬

sented by Equation (3-1).

Em = Sm-PO+ Bm (3-1)

where Em is the overall consumption of a specific energy form in a specified

period (i.e., longer than one day, mostly one month) in kWh per period, Sm is

the specific consumption of one energy form per ton of products in kWh /1, PO

is the production output on a weight basis during the period specified (includingall products and intermediates leaving the plant, excluding solvents and aggre¬

gates) in t per period, and Bm is the so-called base consumption of the buildingof a specific energy form in kWh per period. The base consumption is the con¬

sumption of a warm production building that is ready to start production but in

which no production process is actually running (i.e., base consumption meas¬

ures infrastructure consumption and infrastructure losses).Two different possibilities exist for the determination of the base consump¬

tion. Each building undergoes a period of revisions at least once a year. Duringthis period, maintenance activities are undertaken and production is shut down.

Therefore, it is possible to measure the consumption of the warm (ready to pro¬

duce but not yet producing) and the cold (only safety equipment is running)

13

The Top-Down Approach

building. Losses of the whole system have to be analysed in this way. A sec¬

ond possibility is the direct measurement of the consumption of the specific in¬

frastructure equipment itself since it is known which apparatus is on stream dur¬

ing shutdown or production.Such linear models were also postulated by (Blickenstorfer 1999). Models

of this kind are only applicable to monoproduct or multiproduct batch plants or

multipurpose batch plants with similar products as will be discussed in Chap¬ter 4.

For multipurpose batch plants with large differences between their productsand changing production mix, linear TODOMO are not applicable as will be

shown in Chapter 4. For these buildings, that are the main research topic of this

thesis, a new BOTUMO is postulated and discussed in the Chapter 3.2.

3.1.2 The Heating Steam Model

Production plants are heated by heating the fresh air entering the building.This is (unlike to residential buildings, where radiators are used most of the

time) done by heat exchangers with condensing steam. This (comfort) heatingsteam is measured separately. A linear model only depending on degree-days(see (http://www.eia.doe.gov/neic/infosheets/degreedays.htm )) was first postu¬lated according to Equation (3-2) but found to be not applicable as discussed in

Chapter 4.3.1.

SC = DSS DD + B (3-2)

where SC is the steam consumption in MWh / month, DSS is the degree-

day specific steam consumption in MWh / °C / d, DD is the number of degree-

days in °C • d / month and B is the base consumption of heating steam in

MWh / month, which is unique for each building.Since the air change rate of production buildings is significantly higher than

for residential buildings (safety reasons), the model was adapted to account for

the air change rate. The adapted model was found to be applicable for the heat¬

ing steam consumption of batch plants (see Chapter 4.3.1) and is depicted in the

following equation:

SC = 0.32 • ACR -DD + B (3-3)

where ACR is the air change rate of a building in h"1.

If no production infrastructure uses heating steam and if the main pipe of

heating steam is closed during summer, the base consumption is almost equal to

zero. Otherwise, the base consumption has to be measured or estimated before

predictions of heating steam consumption can be made, as discussed in Chap¬ter 4.3.1.

14


3.2 The Bottom-Up Approach

The basic equations for the BOTUMO, describing the concepts of calculat¬

ing the energy consumptions for heating and cooling procedures (Chapter 3.2.1)and calculating the energy consumption of the electric equipment (Chap¬ter 3.2.2) are presented here. These basic equations are combined in different

ways for the different unit operation models on single apparatus level presentedin Chapter 3.2.3 and 5. The single unit operation models are summarised to re¬

sult in a model of a whole plant (see Chapter 3.2.3 as well).

3.2.1 Equations for Heating and Cooling of Substances

In any book dealing with heat transfer and physical chemistry (e.g., (Atkins

1990) or (Wedler 1987) or (Perry et al. 1997)), the basic equations for the heat¬

ing and cooling of substances can be found. The heating or cooling of a sub¬

stance without phase change can be calculated by Equation (3-4).

AH = \m-cPdT (3-4)

where AH is the enthalpy change in kJ, Ti and 7~2 are the temperatures at

the beginning and the end of the heating process in K, Cp is the heat capacity of

the product in kJ / kg / K and m is the mass of the heated substance in kg.With the help of the assumption, that m as well as Cp stay constant in the in¬

vestigated temperature range as investigated by (Dahinden 2003), Equa¬tion (3-4) can be simplified, resulting in Equation (3-5).

AH =E = m-cP-(T2-T^ (3-5)

The generic equation for the energy consumption of a substance undergoinga phase change (i.e., crystallisation, freezing or evaporation) or performing a

chemical reaction is presented in Equation (3-6).

AH = E = m-AH, (3-6)

where AH, in kJ / kg signifies the heat of reaction (r), evaporation (or con¬

densation) (v), freezing (m), or crystallisation (or melting) (c), respectively.As stated in (Perry et al. 1997), heat losses through a solid wall are propor¬

tional to the temperature difference, the surface, and the time of operation and

are insulation-specific coefficients as shown in Equation (3-7).

AH=E = K-A.(THJ-TAm).At (3-7)

here, K is the heat transfer coefficient in kW / m2 / K, A is the total surface

area of the apparatus in m2, Thj and T/\m are the temperature in the heating

15

The Bottom-Up Approach

jacket of the apparatus and of the ambience in K, respectively, and At is the op¬

eration time in s.

3.2.2 Equations for Electric Equipment

The energy consumption of electric equipment is strongly related to its

nominal power. The nominal power is a physical property describing electric

equipment. Measurements of the actual power consumptions lead to Equa¬tion (3-8).

EEI=yPN-At (3-8)

here, y is the fraction of nominal power consumed by the equipment, ex¬

pressed in percent, P/v is the nominal power of the equipment in kW and At is

the time of operation of the equipment in s.

As stated in (BBC 1976)1 the efficiency of an electric motor decreases when

not operated at nominal power. For standard motors, the values are given in

Table D-7 and depicted in Figure D-l. Moreover, shaft power is lost2 in the

transmission (about 5%) and by the use of frequency converters (about 10% be¬

cause of imperfect sinusoid-curves of the electric current after the frequency

converter). Shaft power of a stirrer is considered to directly contribute to heat¬

ing of the vessels according to Equation (3-9).

E=r/-y-PN-At (3-9)

where r/ is the efficiency of the motor given in Table D-7 in %.

According to (Perry et al. 1997), the power consumption of a vacuum pump

can be calculated as follows:

P = pV (3-10)

where P is the power consumption in J (or kWh), p is the pressure at which

the pump is operating in Pa, and V is the volume the pump is extracting from

the vessel in m3.

1

Although this source is rather old, its findings are still valid today according to industry ex¬

perts2

According to discussions with industry experts

16


Power consumption of electric equipment may be calculated in general ac¬

cording to Equation (3-11) (see e.g., (Kneubühl 1994) for detailed explanationof the equation).

p =

S-l-U-COSÇ1000

here, P is the electricity consumption in kW, / is the current in A, U is the

total voltage in V (i.e., 500 V), and COScp is the power factor, specific to each

motor.

The general equation for the mixing of a fluid inside a stirred vessel is of the

form of Equation (3-12) (see e.g., (Mersmann et al. 1975)):

P = Ne -p-n3 -d5 (3-12)

where P is the power needed for mixing in kW, Ne is Newton's number, pis the density of the fluid in kg / m3, n is the number of revolutions per minute

in min"1, and d is the diameter of the stirrer in m.

The general model for constant consumption is postulated according to the

following equation:

Em = C-t (3-13)

here, Em is the consumption of the specific energy form m (steam, electric¬

ity, brine) in kWh, C is a constant consumption per time of the specific energy

form, and t is the operation time in s.

3.2.3 Unified Equation for the Bottom-Up Modelling

The concept of the BOTUMO is given in Figure 3-1. The energy consump¬

tion of a production plant is split into infrastructure consumption and a produc¬tion dependent consumption part. So far, this is a similar concept as the

TODOMO discussed in Chapter 3.1. In addition to the TODOMO, the produc¬tion dependent part is analysed by the BOTUMO in more detail. This will be

discussed in the following part of this chapter.

17


Number of

EnergyForms m

er of

Products

(PSP) i

/Number of

Equipments j

Figure 3-1: The basic concept of the BOTUMO

The model of a whole plant will be built similar to the TODOMO and ac¬

cording to the following basic equation:

Ep+E'm m

t (3-14)

here, Em is the energy consumption of one energy carrier in the whole build¬

ing in kWh per period, EPm is the production dependent consumption of one en¬

ergy carrier in kWh per period (including the production infrastructure depend¬

ent energy consumption), E'm is the energy consumption of one energy carrier

for the building infrastructure3 in kWh per s, and t is the length of the period in s

per period. The building infrastructure energy consumption is specific for each

plant and measured or calculated on building level. The production dependent

energy consumption on the contrary is related to the actual production and uni¬

fies the equations given in the Chapters 3.2.1 and 3.2.2 (see Chapter 5 as well).The production dependent energy consumption is divided in a term that is

related to the reaction mass, another term is associated with the apparatus and a

last term that relates to the losses.

Equation (3-15) gives the basic concept of the production dependent energy

consumption calculation.

P= RM A L

(3-15)

where Ep is the total production dependent energy consumption of one spe¬

cific energy carrier of the whole building in kWh, E*M is the ideal consumption

Including ventilation systems, heating, lighting, personal computers, warm tab water prepara¬

tion, general vacuum pumps, etc.

18


of one energy carrier related to the reaction media in kWh, E^ is the ideal con¬

sumption of one energy carrier related to the apparatus in kWh, ELm is the time

dependent loss and motor term of the consumption of one specific energy carrier

in kWh.

Each of these energy consumption terms consists of different parts: different

forms of energy (m, as mentioned above), production in different apparatus (J)and the production of different chemicals (/). This leads to a split of the energy

cube depicted in Figure 3-1 as shown in Figure 3-2.

Figure 3-2: The principle of the BOTUMO

In Figure 3-2, the small cubicle EPm represents the energy consumption of

one specific product / (probably only one step of its production recipe), pro¬

duced in one specific apparatus ;, requiring the energy form m. The generalequation for this calculation is shown in Equation (3-16).

ijm~

ijm ijm ijm VJ"^°/

here, Epm is the above-mentioned energy consumption of one specific (in¬

termediate) product produced in one specific apparatus with the help of one spe¬

cific energy form in kWh per batch, E is the energy consumption of one spe¬

cific energy carrier in kWh per batch of the reaction mass of one specificchemical in a specific apparatus, E*m is the energy consumption of one specific

energy carrier in kWh per batch in an apparatus of one specific production rec¬

ipe, and Ejm is the time dependent loss and motor term of a specific apparatus,

performing a specific recipe for one specific energy in kWh per batch.

The energy consumption of the reaction mass E is defined by the follow¬

ing equation:

19


E = Fm -SZ(C -m* AT>q +mm AH*) (3"1?)

where Fm is a dimensionless factor defining the kind of energy used (i.e., 1

for brine and steam and 0 for electricity if electric heating is not considered), Cp

is the heat capacity in kJ / kg / K, m,jk is the mass in kg, AT is the temperature

difference of the reaction mass in K and AH is the enthalpy (of vaporisation,reaction, melting, etc.) in kJ / kg, the index k is an indicator for the different

chemicals used in the step, and the indexq

is an indicator for the different proc¬

ess steps (e.g., temperature levels, unit operations) of the specific recipe.The energy consumption dependent on the specifications of the reactor Ei

A

ijm

may be explained by:

Ein =Fm-YXCPn- mjn ' ^ (3-18)q n

where the index „ is an indicator for the different aspects (i.e., materials,

cooling liquid) of an apparatus. The temperature difference AT represents the

difference between the starting temperature of the jacket and the final tempera¬ture of an operation. It is assumed (see Chapter A.l in the Appendix), that the

whole vessel reaches the jacket temperature. Heating of the insulation is ne¬

glected.

The losses and the consumption of electric motors E\jm will be expressed by

time dependent equations of the following type:

iKm-vA^-prv^)-u£in =

-/ C3-19)

3600%

where K is the loss coefficient (i.e., heat transfer coefficient) of the appara¬

tus in kW / m2 / K, A is the total surface of the apparatus in m2, AT is the tem¬

perature difference between the ambient temperature and the heating jacket in

K, r is the nominal power of the motor in kW, fis the relation of the actual

power consumption of the motor to the nominal power consumption in %, rj is

the efficiency of the motor in % (i.e., correlation of the power the electric

equipment is actually delivering to the actual power consumption of the corre¬

sponding apparatus), and fis the operating time of one specific process step qin

one specific equipment;, producing one specific product /, requiring one specific

energy formm

in s per batch. The factor 3,600 s / h converts kWs in kWh.

Equations (3-17) to (3-19) are inserted into Equation (3-16). This is the base

equation for the BOTUMO depicted in Figure 3-2. Now, each apparatus repre¬

sents for each chemical produced and for every energy form a single cubicle in

the production dependent cube of Figure 3-2. With the help of the number of

batches (n,) of one chemical / produced in a certain period in apparatus; utilising

energy formm, a summation along the three axes of the energy consumption

20


cube depicted in Figure 3-2 is possible. This leads to different results (e.g., the

total energy consumption of one chemical in all the apparatus required for its

production Ep shown in Equation (3-23)) and finally to the total production

dependent energy consumption E (Equation (3-26)). This is shown in the fol¬

lowing equations without further explanation.

EPn=nl-^E;m (3-20)i

Epm=Yjn,-E;m (3-21)

EP=ni.^E;m (3-22)m

Er=",-IIC (3-23)m ]

Ef=ZZ^-C (3-24)m /

£=S&-£; (3-25)J I

EP=ZZZ^-C (3-26)m ] i

According to these summations, different statements like the energy con¬

sumption of one specific energy form for the production of all the chemicals in

all the apparatus (i.e., Equation (3-25)) are possible.The production dependent energy consumption of the whole plant E given

in Equation (3-26) is equivalent to the one given in Equation (3-15). This

production dependent energy consumption may then be inserted in

Equation (3-14) to result in the total energy consumption of a production plant.These generic equations will be used to model the different kinds of unit op¬

erations. Therefore, the parameters (especially the loss coefficients of the dif¬

ferent apparatus) have to be evaluated with the help of measurements as will be

discussed in Chapter 5.

21

Top-Down Modelling of Production Plants (TODOMO)

4 Top-Down Modelling of Production Plants

(TODOMO)

4.1 The Basic Equation for the Top-Down Modelling

Consumption of the different utilities was measured at the defined system

boundary (see Figure 2-3). These data were collected on a monthly basis. In

addition to these energy consumption data, the production output (tons of prod¬ucts) of the different buildings was determined on a monthly basis as well.

4.2 The Characteristics of the Different Buildings Investigated

Table 4-1 summarizes the characteristics of the investigated buildings.These buildings are typical for production in the specialty chemicals industry.

Buildings 1 to 3 are multipurpose batch plants, conducting chemical reactions

that use either organic compounds (Buildings 1 and 2) or water (Building 3) as

the main solvent. A drying plant (Building 4; multiproduct batch plant), a mul¬

tiproduct batch plant (Building 5) and a monoproduct batch plant (Building 6)

complete the investigation. The buildings are of different sizes, and their pro¬

duction processes vary significantly, as shown in Table 4-1. The analysis of

such a variety of different buildings permits the investigation of the applicabilityof general models for depicting the energy consumption of production build¬

ings. The drying plant (Building 4) consists of several different dryers; mainlyhorizontal vacuum rotary dryers (see Chapter 5.7) and filter presses. Further¬

more, grinding and mixing equipment is available in the plant to shape the dried

products and pack them for the customers.

23

Analysis of the Different Energy Carriers

Table 4-1: Characteristics of the investigated buildings

Building Description Number Main Sol¬ Variability Change of Range of

No. of major vent of Production Reaction

equipment Products Mix Temperatures

pieces

I4 Multipurposebatch plant

29 Organic High High< -10 °C to

> +200 °C

2Multipurposebatch plant

55 Organic High High<-10°Cto

> +100 °C

3Multipurposebatch plant5

180 Water Medium Medium0°Cto

~ +30 °C

4Multiproduct

Drying Plant55

Organic and

WaterHigh High

+60 °C to

> +100 °C

5MultiproductBatch Plant

746 Organic Medium Medium< -20 °C to

> +250 °C

6MonoproductBatch Plant

86 Organic Low None<-10°Cto

> +200 °C

4.3 Analysis of the Different Energy Carriers

In the discussion and examination of the results obtained for the energy

analysis of the different buildings, it must be kept in mind that the model is

based on the total amount of chemicals produced (products and intermediates

that leave the production plant). This is not equal to the degree of usage of the

equipment, as different products have different production processes (especiallyin multipurpose batch plants). If, for example, the amount of chemicals pro¬

duced in one month is only half the amount of chemicals produced in another

month, this could mean that the utilization of the equipment is only 50%. How¬

ever, it could also mean that the chemicals produced in the former month have

more complicated production paths that use more equipment, resulting in almost

100% usage of the equipment. Aside from this shortcoming, the monthly pro¬

duction provides a good picture of the average productivity of the different

buildings and is easily accessible. In the following, the correlation of different

energy consumptions with production output is studied.

Note that for clarity reasons and effective communication possibilities with

industry, energy is accounted in MWh in this thesis. MWh can be converted to

MJ by multiplying the value given in MWh by 3,600.All the different measurements performed for the investigations on building

level at the different buildings are presented in the tables found in Chapter D.l

in the Appendix.

4

Building 1 is discussed and analysed in more detail in Chapters 5 and 6 with the help of a bot¬

tom-up approach5See Blickenstorfer, C. (1999). "Analyse des Energieverbrauchs eines Mehrprodukte-Batch-

Betriebes," Ph.D. dissertation, No. 13411, Zurich, ETH, http://e-collection.ethbib.ethz.ch/cgi-

bin/show.pl?type=diss&nr=13411.6Number excludes cooling machines

24


4.3.1 Steam

Steam is provided at two pressure levels and for two different purposes in

most of the buildings. One purpose is the heating of the building for comfort

reasons and the other is the usage for the production processes. Production

steam is provided at two different pressure levels (i.e., 5 and 15 bar above ambi¬

ent pressure) for providing heating utility at two different temperature levels.

This steam is usually produced in a boiler house for a whole site at a pressure

level of about 40 bar. This high-pressure steam is relaxed to the abovemen-

tioned pressure levels with the help of steam turbines producing electricity as a

by-product. Production steam and heating steam are investigated separately in

the following subchapters.Steam is a very important energy provider for industry, since about 45% of

all fuel burned by manufacturers is expended on steam generation (according to

US department of energy, see (Aggarwal 2002)).

Production Steam

For a comparison of steam, cooling media and electricity consumption, the

energy content of one ton of steam was considered as 0.8075 MWh according to

discussions with industrial representatives and values given in (Lide 1995).The consumption of production steam as a function of the production output

of the different buildings is shown in Figure 4-1 and summarized in Table 4-2.

The two multipurpose batch plants with largely varying production (Build¬ings 1 and 2) show the lowest correlation. The processes conducted in these

plants vary significantly in terms of process temperature and time. Some of the

chemicals have to be produced at more than 200 °C and some at room tempera¬

ture or even below. Some of them show high heat of reactions and some low.

Therefore, a correlation between steam consumption and total amount of chemi¬

cals produced does not exist at all for these buildings. The differences between

the products are too large to obtain an accurate model with this simple approach.

25


2500

c

£ UÏ ~ 2000°-

cS

leg i5oo i-ai S E

ai o

5)""

500

0

^, = 0 96 FO + 759E

R2 = 011

Building 1

fo° 9p o

cftc

200 400 6

Amount of Products PO [t/month]

5 UJ

E û.

ai o

2500:

2000 :

1500 :

1000 :

500 :

0 :

: 1 36 FO + 773 5

R2 = 019

Building 2

200 400 600


Steam

Consumptionfor

ProductionEst

[MWh/

month]en

o

en

o

en

o

o

o

o

o

Building 3

-

-

-

0

^F 8

0 500 1000 1500 2000

Ea = 0 60 FO + 706 3

R2 = 0 26


Building 4

s "J

o gE 0-

^, = 0 69 FO

R2 = 0 32

0 100 200 300 400

Am ou nt of Products PO [t/month]

Building 5

c o tSr,

S<SE 5000 -

° °J8W°

nsum uctior/

mon4000 -

3000 -

S<T°o o 2000 - S°

E û. | ^

1000 -

o

8S-

0 1000 2000 3000 4000

Amount of Products PO [t/month]Ej, = 1 54 FO

R2 = 0 82

s W'

Building 6

rooo -

800 -

600 -

400 -

n q.B^o~

^^bo"SD

200j

0 -

500 r000 V500 2000 2'500

= 033 FO+161 5

R2 = 0 77


Figure 4-1: Consumption of production steam (5 and 15 bar) of the different

buildings as a function of amount of products per month (according to Equa¬tion (3-1))

Building 3 shows a slightly better correlation. Here, the variations between

the different products are smaller than in Buildings 1 and 2, since exclusivelyone family of chemicals is produced. These chemicals, although different, all

have similar production processes and, most importantly, similar production

temperatures.The regression of the data for Building 4 shows almost no base consumption

(no measurements at zero production were done for this building). The lack of a

base consumption found by the regression is reasonable, given that steam con¬

sumption in a drying plant is shut down if nothing is dried and infrastructure

losses are generally small, as shown by the measurements discussed below. The

drying processes for different chemicals in horizontal vacuum rotary dryers are

quite similar7. The heat capacity of most organic solvents is about 2 kJ / kg K

and the heat of vaporization lies in the region of 1 MJ / kg (Lide 1995). There-

7For a description of horizontal vacuum rotary dryers, see Mujumdar, A. S. (1995). Handbook

ofIndustrial Drying, Marcel Dekker, Inc., New York.

26


fore, the main variation between the different drying processes lies in the dryingtime. This time depends on the amount and the relative moistness of the chemi¬

cals to be dried, as well as the adhesion of the solvent to the surface. Most

products do not vary as greatly in these attributes than they do in the attributes

of synthesis steps, and therefore, a better correlation is obtained for the steam

consumption of this drying plant than for multipurpose batch plants.Of the investigated plants, Buildings 5 and 6 show the best correlations be¬

tween production steam consumption and production output. In these buildings,the production mix stays relatively constant (for Building 6, it is completelyconstant because only one specific chemical is produced). This explains the

rather good correlations obtained for steam consumption of these buildings as

compared to the other buildings.The base consumption of production steam was measured only for Build¬

ings 3 and 6. The base consumption accounts for a significant fraction of the

production steam consumption. Nevertheless, the relative amount is smaller

than for electricity. Here, base consumption means losses from steam pipes and

equipment and steam for tab water boilers, for example (if not included in the

heating steam consumption). In contrast to electricity, only minor infrastructure

equipment uses steam.

The measurement of the base consumption of Building 1 was conducted

during a shutdown period (i.e., consumption of the cold building). These meas¬

urements showed that the consumption of production steam of this cold produc¬tion building is significantly smaller than about 70 MWh / month (precision of

the measuring device). The relative base consumption of the cold building is

therefore much smaller than the average steam consumption for this buildingand less than 10% of the regressed base consumption of the warm building(about 760 MWh / month for Building 1 as shown in Figure 4-1).

27


Table 4-2: Summary of the different production energy consumption models

obtained for the different energy forms (m) in the different buildings accordingto Equation (3-1)8

Utility Building O/TJ

[MWh/t]

Bm

[MWh/month]

r2 Base Load [%]50% PO 100% PO

Electricity

1

2

3

4

5

6

0.28

0.23

0.41

0.16

0.12

0.06

1309

1379

277.59

48.610

43.59

1079

0.16

0.41

0.89

0.55

0.86

0.9

64

66

49

65

17

60

47

49

32

48

9

43

Production

Steam11

1

2

3

4

5

6

0.96

0.73

0.6

0.69

1.54

0.33

759.810

773.510

706.39

0io

0io

161.59

0.11

0.19

0.26

0.32

0.82

0.77

75

77

62

0

0

29

60

63

45

0

0

17

CoolingMedia12

1

2

3

5

6

0.1

0.16

0.23

0.04

0.01

99

7.39

09

09

o9

0.03

0.30

0.86

0.9

0.93

26

13

0

0

0

15

7

0

0

0

Heating Steam

The heating steam consumption was set in correlation to the number of de¬

gree-days per month. The heating frontier was set to 12 °C. The daily mean

temperatures were received from a meteorological institution close to the pro¬

duction plants. Explanations on degree-days can be found in the literature (see,

e.g., (http://www.eia.doe.gov/neic/infosheets/degreedays.htm )).The correlations between heating steam consumption and degree-days ac¬

cording to Equation (3-2) can be seen in Figure 4-2. Buildings 5 and 6 consume

no heating steam, as these buildings are heated with condensate (i.e., hot water)originating from production heating, which is not measured separately. There¬

fore, only Buildings 1 to 4 are analysed.The two multipurpose batch plants with significantly changing production

(Buildings 1 and 2) show similar specific heating behaviours (slopes of the re¬

gression lines). The high base consumption (intercept at zero degree-days) of

8

Applicable energy models are printed in bold face9Measured data at zero production

10Result of the regression for zero production

11

Energy content of steam is set to 0.8075 MWh/t according to discussions with industrial rep¬

resentatives and values given in Lide, D. R. (1995). "Handbook of Chemistry and Physics.",CRC Press, London.12

Energy (electricity) used for the production of the cooling media and not the energy content of

the cooling media

28


Building 2 appears extraordinary. Nevertheless, it can be explained by infra¬

structure equipment (such as heating of storage tanks) running on heating steam

over weekend shutdown periods for safety reasons. This explains the inferior

correlation (varying production infrastructure consumption) as well as the highbase consumption.

5) 200

O1 O Building 2 i

SC = 060-DD + 179l!

D Building 4

SC = 0 64 DD + 48 0

R2 = 0 94

A Building 1

SC = 0 58 DD + 1

R2 = 0 96

.. T.-'.O Building 3 ;

SC = 0 30 DD + 42 1|R2 = 0 89 j

200 300 400 500 600

Number of Degree-Days per Month DD [°C • d / month]

Figure 4-2: Consumption of heating steam (5 bar) as a function of degree-days

per month (according to Equation (3-2))

Despite the different sizes of the two Buildings 1 and 2 (see Table 4-1), the

specific energy consumption per degree-day (slopes of the regression lines in

Figure 4-2) is similar for the two buildings. Building 3, which is the largest of

the investigated plants, shows the smallest specific energy consumption. This is

even more astonishing as the processes in this building produce only minor

amounts of heat and are conducted at moderate temperatures (i.e., low irradia¬

tion of the equipment). This should result in a higher heating requirement for

the plant. An explanation of this behaviour could be the heating regime of pro¬

duction buildings: Unlike apartment or office buildings, radiators are seldom

found in production plants. Heating is performed mostly with the help of heat¬

ing ventilators working with steam. Moreover, the air change rate in productionplants is higher than in apartment buildings for safety reasons (about 2 h"1 ver¬

sus about 0.5 h"1). The fresh air has to be heated before entering the building.The number of times, the air volume of a building is exchanged is called air

change rate ACR and is given in h"1. It will now be considered, how this air

change rate influences the heating steam consumption levels of the different

buildings.The result of this investigation is shown in Figure 4-3, where the normalized

consumption of heating steam is plotted as a function of degree-days. The spe¬

cific normalized heating steam consumption of the investigated buildings is

about 0.32 (MWh • h) / (°C • d). The normalized base consumption of the inves-

29


tigated buildings still varies largely. As mentioned above, this can be explained

by the varying use of the heating steam for production infrastructure (i.e., inde¬

pendent of ambient temperature). The good correlation shows that the normal¬

ised model according to Equation (3-3) is adequate for the heating steam con¬

sumption of production buildingsIt may be seen from the investigations, that if no production infrastructure

uses heating steam and if the main pipe of heating steam is closed during sum¬

mer, the base consumption is almost equal to zero. Otherwise, the base con¬

sumption has to be measured or estimated before predictions of heating steam

consumption can be made.

S O

Air change rate Building 1 = 1 8 h1

Building 2 = 1 9 h1

Building 3 = 1 0 h1

Building 4 = 2 0 h1

1O Building 2

,SC/ACR = 0 33 DD + 90 5'

O Building 3

SC/ACR = 0 32 DD + 33 5

R2 = 0 90

D Building 4

SC/ACR = 0 32 DD + 24 8

R2 = 0 94L J

A Building 1

SC/ACR = 0 32 DD

R2 = 0 95

300 400 500

Number of Degree-Days DD [°C d / month]

Figure 4-3: Normalized heating steam consumption (5 bar) as a function of the

number of degree-days per month (according to Equation (3-3))

The main reason for the lack of a base consumption for most buildings is

that the heating steam pipe is closed during summer times and therefore no

losses occur when no heating is required. Some minor errors arise because the

measuring location for the mean day temperatures (required for the calculation

of the degree-days) was not located directly at the plants location, which mightcause an error of ±1 °C.

The results are summarized in Table 4-3.

30


Table 4-3: Summary of the models for heating steam consumption obtained for

the different buildings according to the normalised Equation (3-3)8

Building ACR -0.32

[MWh/°C-d]

B

[MWh/month]

r2 SC13

[(MWh-h)/(°C-d)]

ACR

[h"1]

Base Load

[%]350 700

DD DD

1

2

3

4

0.58

0.60

0.30

0.64

0.99

179.1

42.1

48.0

0.96

0.79

0.89

0.94

0.32

0.33

0.32

0.32

1.8

1.9

1.0

2.0

0

46

29

18

0

30

17

10

4.3.2 Electricity

The electricity consumption data of the different buildings investigated are

shown in Figure 4-4. The investigations exclude the electricity consumption for

cooling purposes since this consumption is discussed separately in Chap¬ter 4.3.3.

For Buildings 1 and 2, the electricity consumption is barely correlated with

the monthly production. The energy consumption of these buildings varies

greatly between the different production processes, mainly due to differences in

reaction time and in batch size, which are not always correlated with equipmentsize (standardized reactors with large motors).

Buildings 3, 5 and 6 show a different behaviour as shown in Figure 4-4. For

these buildings good correlations between electricity consumption and amount

of chemicals produced are obtained. Building 3 shows high specific electricity

consumption (slope of the regression line). This can be explained by the largestirring equipment used in this building. The large stirring equipment is in¬

stalled because high stirring powers are sometimes required for the processes

(e.g., for dissolving a solid). Although about 160 different chemicals are pro¬

duced in this plant, the production processes are quite similar. This results in a

good correlation between electricity consumption and production output.

Building 5 shows lower specific electricity consumption than Building 3. In

Building 5, each production line is exclusively constructed for one product. The

products originate from one family. The product mix stays constant over the

year. The dedicated facilities and the uniformity of the production processes for

the different products explain the good correlation obtained for this plant.Building 6 shows a good correlation between electricity consumption and

production output. Because only one product is produced in this building and

the production process is highly automated, the differences between different

batches are minimal. Therefore, this set of data shows the highest correlation

coefficient.

Building 4 is completely different from the other buildings. The drying of

different chemical products shows differences in drying time and initial moist-

ness. The equipment sizes in this building do not vary largely. Motors for vac-

13

Specific normalized heating steam consumption

31


uum pumps and stirring dominate the electricity consumption. This results in

only minor differences between different products. Therefore, the correlation of

electricity consumption and dried products lies between those of Buildings 1

and 2 and of Buildings 3, 5 and 6. The obtained results are summarized in

Table 4-2.

Uj

o

o

o

Üo

UJ

350

300 --

250 :-

200

150 ±

100

50

0

O Building 1 ;

EB = 0 28.PO + 130;

FT = 0 16

'-^^ xS> O

DD

D D

D

q>i D Building 2

j EB = 0 23.PO + 137i

! R2 = 0 41 !

O Building 4

EB = 0 16.PO + 48 56

R2 = 0 55

0 100 200 300 400 500

Amount of Products PO [t / month]

600 700

„ 900

/AABu Iding 3

EB = 0 41 PO + 277 47

R2 = 0 89

1000 1500 2000 2500 3000

Amount of Products PO [t / month]

3500 4000

Figure 4-4: Electricity consumption (excluding electricity for cooling purposes)of the investigated buildings as a function of the amount of chemicals produced

per month (according to Equation (3-1))

32


For Buildings 1, 2, 3, 5 and 6, measurements of the base consumption of

electricity were conducted (measured values at zero production in Figure 4-4) as

indicated in Table 4-2 For Building 4 no measurements of the base consump¬

tion were made

In terms of electricity, base consumption means mainly infrastructure con¬

sumption, as losses are minimal The proceeding of the measurements was as

follows If possible, the consumption of electricity of the infrastructure was

measured continuously, such as the 400 V electricity consumption of Building 1

(see Figure 4-5) Where this was not possible, measurements of the consump¬

tion were done manually at different times The different measurements showed

only minor fluctuations Therefore, an average value was calculated and as¬

sumed constant over the whole investigation Among others, ventilation, vac¬

uum systems, and waste air treatment were considered as infrastructure equip¬ment and use 500 V electricity while lighting consumes 400 V electricity

500 T-

o

Ü

LU

-r-20

400 V Infrastructure Consumption (approx 30 kW)

500 V Infrastructure Consumption (approx 150 kW)

~i—

40

-1—

60 80

Time [h]

—i—

100 120 140 160

Figure 4-5: Hourly electricity consumption of Building 1 (without electric heat¬

ing of a specific process) during an ordinary week in 2001

All buildings show high base levels of electricity consumption Figure 4-4

shows that the infrastructure consumption of electricity contributes significantlyto the total electricity consumption The highly automated Building 6 shows a

higher percentage of infrastructure consumption compared to the less automated

Building 5 In automated buildings, much equipment is running independentwhether it is in use Automation shows only small consumption as stated in

(Schalcher et al 2003 a, Schalcher et al 2003b) Major consumers of electricityin this building are high temperature equipments requiring electricity Duringshutdown periods, trace heating of these equipments is left active since shut¬

down would cause the product to crystallise in the pipes Dedicated plants with

manual operation like Building 5 show smaller levels of infrastructure consump-

33


tion, if the processes do not need large and dedicated equipment, which would

consume higher amounts of base load. In this specific building, fewer scrubbers

are installed compared to the others. This results in a lower electricity con¬

sumption of the infrastructure. In addition, the electricity consumption in Build¬

ing 5 is more dependent on production output than the consumption of Build¬

ing 6 as shown by the higher slope of the Building 5 regression line in Figure4-4.

The high flexibility of a multipurpose batch plant implies a high flexibilityof the infrastructure equipment. The equipment is therefore built to handle the

highest possible requirement of the plant. This explains the high base consump¬

tion of Buildings 1 and 2 as shown in Figure 4-4. The lower percentage of the

base consumption of Building 3 as compared to Buildings 1, 2 and 4 (see Table

4-2) can be explained by the limited variability of the chemicals produced. This

fact makes it easier to size the utility equipment. Finally, no correlation be¬

tween energy consumption and the number of active equipment units was found

as shown in Figure 4-5 for Building 1.

4.3.3 Cooling Energy

The investigated buildings used two types of cooling media. For coolingabove about ambient temperature, cooling water (taken from a river) was used.

Because this water was not measured separately from the other water used for

the production and no cooling towers were in use, the cooling water consump¬

tion was not investigated. For low-temperature cooling, three different cooling

systems were in use. Buildings 1 and 2 use brine that was produced (i.e.,cooled) externally. In Building 1, the internally used brine is cooled down with

external brine using a heat exchanger for safety reasons, whereas in Building 2

the external brine was used directly. Building 3 used no brine at all; rather, the

processes in this building used ice for direct cooling. This ice was producedinternally with two ice machines. Building 4, as a drying plant used only water

for cooling purposes. Therefore, this building was not considered in this inves¬

tigation. Buildings 5 and 6 used brine that was produced internally. Here, the

energy content of the cooling media was not measured. The investigated energy

consumption is the energy consumption (electricity) required to produce the

cooling media. Assuming a reasonable efficiency of the cooling machines

(about 200%) in reference to electricity input as stated in (http://www.aie.org.au/

melb/material/resource/cop.htm ; Wang 2000)), the effective cooling duty could

be estimated, but this was not done here for better comparison with the other

utilities investigated.The consumption of cooling media for the different buildings can be seen in

Figure 4-6.

34


140

120 +

100

80 +

60

40 +

20

DD

D

| D Building 2 j

!eCo = 0 16.PO + 7 3o!

! R2 = 0 30 I

D

o <2}^er o o

D

o o

Building 1

ECo = 0 10.PO + 9

R2 = 0 03

100 200 300 400 500


600 700

350

A Building 3

Eoo = 0 23.PO

[ D Building 5

|Eoo = 0 04.PO

Rz = 0 90

1000 1500 2000 2500 3000


4000

Figure 4-6: Consumption of cooling energy of the different buildings as a func¬

tion of production output per month (according to Equation (3-1))

Buildings 3, 5 and 6 exhibit good correlations between cooling media con¬

sumption and production output. This is because of not only the internal pro¬

duction and the controlling mechanism, but also the uniformity of the produc¬tion processes in each of these buildings.

The multipurpose batch plants with varying production (i.e., Buildings 1 and

2) show, once again, a different behaviour. The large variety of the productsresults in a poor correlation between cooling media consumption and amount of

products. Models that are more detailed have to be built to model the energy

consumption of such facilities.

35

Applicability of the Models

The buildings that produce their cooling media internally (Buildings 3, 5

and 6) show no base consumption as shown in Figure 4-6. This can be ex¬

plained by the fact that the cooling machines are shut down if not in use. The

machines are controlled by measuring the cooling media consumption of the

plant (i.e., the temperature of the backflow). The cooling power of the ma¬

chines is adapted accordingly with frequency converters.

For Buildings 1 and 2, base consumption levels of about 10%> of the average

production consumption results. The infrastructure has to provide a base load

even if no production occurs. Moreover, losses of the system are higher be¬

cause of the longer piping systems (piping between the cooling machines and

the different production plants). The higher base consumption of Building 2

might be due to the two brine systems (external and internal), which are joinedwith a heat exchanger. This heat exchanger has a specific heat loss (i.e., about

1 °C temperature difference between the supply temperature of the external

brine and the heat exchanger outlet stream temperature of the internal brine) that

results in a higher base consumption of the building. The results are summa¬

rised in Table 4-2.

4.4 Applicability of the Models

Generally, it can be stated that for multipurpose batch plants with highlyvarying production processes and changing production mixes (i.e., Buildings 1

and 2), energy consumption models according to Equation (3-1) are not suitable.

The variations between the different products are too large to be modelled with

highly aggregated energy models on a building level.

The results in the preceding section show that modelling the energy con¬

sumption on the building level according to Equation (3-1) is suitable for some

production plants and not suitable for others. The postulated model for energy

consumption on the building level is suitable for dedicated monoproduct batch

plants (Building 6) or for multiproduct or multipurpose batch plants in which

similar chemicals are produced or the product mix stays constant over time

(Building 3 and 5). The buildings where an energy model according to Equa¬tion (3-1) can be applied are printed in bold face in Table 4-2.

For electricity consumption, the model expressed by Equation (3-1) was

suitable for Buildings 3, 5 and 6. Figure 4-7 shows an example of the modellingof the electricity consumption of these buildings according to Equation (3-1)with the parameters given in Table 4-2. As mentioned above, the maximum

production capacity (i.e., 100%>) was taken as the highest observed productionduring the investigated period. The percentage contribution of base consump¬

tion to energy consumption is specific to each building. As can be seen in

Figure 4-7, the total amount of energy consumed per unit of chemical produceddecreases with increasing plant usage since the base consumption of the build¬

ing stays constant. At higher plant usage, the base load can be distributed to a

higher number of products. From an energetic point of view, it is therefore bet¬

ter to run a plant half a year at full capacity and shut it down for the rest of the

year than producing at half capacity for the whole year. Considering only en¬

ergy costs, higher plant usage results in lower production costs.

36


The modelling of the electricity consumption of Buildings 3 to 6 and the

production steam consumption for Buildings 5 and 6 according to Equa¬tion (3-1) showed that a significant part of the energy consumption of a batch

plant is independent of production (i.e., base load).

o

50% | 75%

Building 6

Production Output PO (as percentage of highest)

Figure 4-7: Modelled monthly electricity consumption as a function of capacity

usage for those buildings where the model according to Equation (3-1) was suit¬

able

For heating steam, a model according to Equation (3-3) was proposed. This

model was suitable for all the buildings investigated (i.e., Buildings 1 to 4). The

investigations showed that the heating steam consumption depends only on the

number of degree-days and air change rate of the building. The correspondingbase load depends on the infrastructure (both production and building) that is

running with heating steam and therefore varies significantly between the dif¬

ferent buildings. These results are summarized in Table 4-3.

The discussions above and in the preceding sections lead to the flowchart for

energy modelling in batch plants depicted in Figure 4-8.

37

Conclusions

Energy Modellingof

Batch Plants

Heating Steam Model:

SC = 032- ACR -DD + B

(See Equation (3-2) and

Table 4-3)

Yes-

YesH

Production Energy Models

Bottom-Up:

Single Unit Operation Models

(see Chapters 5 & 6)

Production Energy Models

Top-Down:

Em = Sm-PO + Bm(see Equation (3-1) and

Table 4-2)

Figure 4-8: Flowchart for energy analysis in chemical batch production

4.5 Conclusions

For plants with only minor changes in production mix, it is possible to ob¬

tain a good description of energy consumption by use of Equation (3-1). For

these buildings, one can allocate energy use per mass of chemicals producedafter determining (measurement or estimation) the required parameters (i.e.,base consumption, specific energy consumption per ton of chemicals produced).The energy consumption per ton of product depends significantly on the plant

usage. The higher the plant usage, the smaller the ton-specific energy consump¬

tion because of the constant base-consumption of the building, thus providingpossibilities to optimise the production plans of such buildings.

For the heating steam consumption of chemical batch plants, a model ac¬

cording to Equation (3-3) is suitable. The model depends only on the amount of

degree-days, the air change rate, and the empirical base load and is therefore of

general use for production buildings. Optimisation could be performed in terms

38


of minimizing base consumption and optimising air change rate and room tem¬

perature (changing the heating frontier).In cases where these equations are suitable, the allocation of energy con¬

sumption to produced amounts of chemicals is possible as is the forecasting of

energy consumption or adequate costing. It is possible to distinguish the base

load from production-dependent energy usage. This shows whether the con¬

sumption of the processes or of the infrastructure is most promising for optimi¬sation. A detailed allocation of the energy consumption to single unit opera¬

tions or products is nevertheless not possible with this top-down approach of

energy investigations. This is a main drawback of the top-down approach: it

precludes detailed optimisation. A better, although more intricate, possibility is

therefore a bottom-up energy model. This model consists of a sum of detailed

energy consumption models for single unit operations as shown in Chapter 5.

These unit operation models, together with the production recipes, reflect the

energy requirements of different products and, thus, allow an allocation of en¬

ergy costs to single products. Furthermore, these detailed models reveal the

amount of energy consumed for each production step of each product and how

large the losses are. A model of a complete production building is possible bysummarising the single apparatus and unit operation models and the infrastruc¬

ture consumption as shown in Chapter 6. Therefore, the application of such a

model leads to the identification of detailed improvement potentials in singleunit operations and production steps (e.g., optimal choice of a solvent used in a

certain operation, optimised insulation of an equipment unit).The modelling of a multipurpose batch plant with varying production (i.e.,

Buildings 1 and 2) has to be done using this more detailed type of energy mod¬

els. These bottom-up models will be investigated in Chapters 5 and 6.

39

Modelling of Single Unit Operations

5 Modelling of Single Unit Operations

Different measuring equipment was used for the measurement of the elec¬

tricity, the cooling energy and the heating energy consumption. The measuring

equipment and its accuracy is discussed in Chapter B in the Appendix.Only one of the six production plants discussed in the preceding chapter is

investigated further on single unit operation level (i.e., Building 1). Therefore,this building will now not be called Building 1 anymore (characteristics are pre¬

sented in Table 4-1), but just (investigated) building for simplicity reasons. The

apparatus in the investigated building are mostly built according to DIN stan¬

dard (see (http://www2.din.de/index.php?lang=en )). This standard is widelyused in industry. The results of the different measurements should therefore be

transferable to other industrial buildings. As discussed in Chapter 7.2, this has

to be investigated in further studies.

The base equations for the BOTUMO are described in Equations (3-14) and

(3-15). The different equations building the BOTUMO may be found in Chap¬ter 3.2. In each of the following subchapters, the equipment will first be de¬

scribed shortly, then the measurements will be presented, finally the generalmodel for the unit operation will be presented, and the equations will be de¬

scribed.

For the conversion of kg of steam to kWh of energy consumption (and vice

versa), values may be found in (Lide 1995). This data source and discussions

with industry experts led to the conclusion, that a value for the energy content of

about 0.65 kWh / kg of steam (including cooling down of the condensed steam

to the temperature of the water in the jacket) is reasonable. This value was

taken for 15 bar as well as for 5 bar steam. The same value was taken for both

pressure levels of the steam since heat of vaporisation is not changing greatlywith changing temperature (according to the accuracy of this investigation).

5.1 Reactors

5.1.1 Description of the Equipment

A scheme of a standard batch reactor as it is operated in the investigated

building is shown in Figure 5-1 together with its heating/cooling-system. The

reactor consists of a vessel with its stirring equipment (for description of the

stirring equipment see Chapter 5.5).The heating/cooling-system consists of a heating jacket (either a double-

jacket for most of the glass lined vessels or a construction with half-pipes for

most of the stainless steel vessels) in which the heating and cooling fluids circu¬

late. Heating up the water in the jacket under pressure performs the heating of

the vessels.

41

Reactors

Steam Water

Brine

Condensate &

Wastewater

Brine

M

Figure 5-1: Scheme of a standard batch vessel with its heating/cooling-system

While heating, the system is filled with water (circulated by a pump). Steam

(either 5 or 15 bar) is injected in the circulating water. This heats the water up

to the desired temperature. The system is open to the wastewater system via an

expansion vessel and a steam trap to prevent cavitation within the pump.

When cooling with water, the system operates similarly, except that the

steam entrance is closed.

If the vessel is cooled with brine, water and steam entrances are closed and

the outlet to the brine system is opened. The circulation pump has to be oper¬

ated as well to allow free flow in the system. For brine, the system is a clear

input-output-system, since the brine enters the system, flows through the jacketand leaves the system through the brine outlet immediately.

5.1.2 Measurements

Different measurements for the brine and the steam consumption of the

reaction vessel are conducted. For all the different types of reaction vessels,different measurements were taken if possible. For brine, this was not possible,since only few reactors needed brine for their operation. Moreover, only some

reactors were connected to the brine system.Care had to be taken not to interfere with daily production of the investi¬

gated building. For this reason, all measurements were taken during normal

production (with exception of the investigations of the cleaning of a vessel de¬

scribed in Chapter C.3 in the Appendix) with only minor disturbance of the pro¬

duction processes.

Steam

The measurements of the steam consumption were conducted as discussed

in Chapter B. 1 in the Appendix. An example of the measurements performed is

shown in Figure 5-2. It can be seen that at the beginning of a batch, the most

steam is consumed (fast heating up of the reaction mass) and that a smaller

amount of steam is used for holding the temperature at a constant value duringthe operation. The results of all the measurements performed are presented in

Table D-8.

42


Lj \l16 09 200312 00 17 09 2003 12 00 16 ) 2003 12 00 19 09 2003 12 00 20 09 2003 12 00 211

Date & Time

-Steam Consumption

Figure 5-2: Example of the steam measurements for a 10 m, glass lined reac¬

tion vessel heated with 5 bar steam

For the high temperature reactor discussed in the Subchapter Electricity,measurements of the steam consumption were also taken. This vessel is onlyheated with steam up to a certain temperature during the starting period of each

batch. Above this temperature, the electric heating is introduced for surplus

heating power and above a significantly higher temperature, no steam is used at

all and only electric heating is provided to the heating system. The special na¬

ture of this equipment is considered in the modelling (see below). The meas¬

ured values may be found in Table D-l8.

Brine

The measurements of the brine consumption showed to be complicated.

Only some of the reactors were connected to the brine system as mentioned

above. Moreover, many of the reactors using brine were connected to the brine

system so badly that no measurements were possible (e.g., too short connection

to the main pipe for the measuring equipment; see Chapter B.2 in the Appen¬

dix). This prevented the measurements of a good part of the reactors that would

guarantee a significant spot check of the different systems. The difficulty of the

measurements is discussed in Chapter B.2 in the Appendix and will not be re¬

peated here.

Measurements of the hourly average of the brine consumption were per¬

formed. The temperature of the reaction mass (IT), the temperature of the

jacket (OT), the brine flow in the jacket and the brine consumption according to

Equation (3-5) were gathered as depicted in Figure 5-3.

43

Reactors

12 112003 13 112003 14 112003 15 112003 16 112003 17 112003 18 112003 19 112003 20 112003

Date & Time

| CTT IT Brine Consumption Brine Flow

Figure 5-3: Example of brine measurements for a 10 m3 stainless steel vessel

Table D-9 summarises the measurements of the brine consumption per¬

formed.

Electricity

The electricity consumption of the stirring and circulation equipment is dis¬

cussed in Chapter 5.5 separately. In this subchapter, the electric heating of one

specific vessel (i.e., high temperature reactor) will be investigated and discussed

in detail.

The electricity consumption of the process control equipment is not meas¬

ured or investigated. According to (Schalcher et al. 2003a; Schalcher et al.

2003b), the energy consumption of this equipment is negligible compared to the

energy consumption of the controlled motors.

For one high-temperature reactor (4 m3 stainless steel reaction vessel), an

electric heating aggregate is installed with a nominal power of 400 kW. This

vessel is not heated directly, but is heated with a heating-oil14 circuit. This cir¬

cuit is heated either with steam (15 bar), or with electricity, or with both as de¬

scribed above, or is cooled with water through heat exchangers. The steam

measurements of this vessel are discussed in the steam measurement paragraphabove and will not be repeated here. Electricity measurements were performedwith the help of a Memobox described in Chapter B.3 in the Appendix. An

example of these measurements is shown in Figure 5-4. The figure shows that a

base consumption of electricity exists. This base consumption is due to the cir¬

culation pump of the system that is running continuously. At the beginning of

each batch, a high peak in electricity consumption is observed that shows the

heating of the reaction mass. After this peak, only minor consumption is ob¬

served according to the measurements (the reaction is exothermic and helpstherefore to balance the losses).

14 Marlotherm®; for details see http://www.marlotherm.com

44


E 250

07 05 2003 08 05 2003 08 05 2003 09 05 2003

12 00 00 00 12 00 00 00

05 2003 10 05 2003 10 05 2003 11 05 2003

12 00 00 00 12 00 00 00

Date & Time

-Electricity Consumption

Figure 5-4: Measurements of the electric heating of the 4 m stainless steel

high-temperature reaction vessel

5.1.3 Model and Conclusions

Different authors have dealt with the problem of heat integration or energy

optimisation of batch reactors (see e.g., (Anonymous 1986a; Aziz and Mujtaba2002; Carpenter 2001; Hessel et al. 2002; Kemp and Macdonald 1988; Uhle-

mann et al. 1996; Wardle et al. 1987)). Most of them, nevertheless, tried to

adapt the concept of thermal coupling and pinch analysis to batch operation. It

is a fact that batch reactors have not improved significantly during the last cen¬

tennials. The basic concept remained the same. Recently, new concepts arose

but did not manage to supersede the classical batch reactor until today as men¬

tioned in (Stitt 2002) (see Chapter 2.1 as well).A recent approach to model the energy consumption of batch reactors is pre¬

sented by (Bouhenchir et al. 2001). Similar to earlier models, this approach has

the drawback that it is too complicated for daily business in an existing produc¬tion plant (too much data and too many unknown parameters are required).Given the flexibility in designing a new production plant, the required meas¬

urements could be installed with minor increase in costs. In contrast to the

models described by (Bouhenchir et al. 2001), the models investigated in this

thesis are applicable to existing plants and require only minimal supplementarymeasurements when transferred from one plant to the other15.

The model for the stirrer and the circulation pump is described in detail in

Chapter 5.5.

The models for heating and cooling of reactors are based on several assump¬

tions, presented in Chapter A. 1 in the Appendix.

'

This has of course to be proven by further investigations (see Chapter 7.2)

45

Reactors

Steam

The model for the calculation of the steam consumption (either 5 or 15 bar)of the reaction vessels was postulated according to the generic model given in

Equation (3-16). The general Equation (3-5) for heating up of the vessel and the

substances, Equation (3-6) for evaporation of the solvent and the heat of reac¬

tion, Equation (3-7) for the losses, and Equation (3-9) for the heat input by the

stirrer were combined to result in the following detailed equation of the steam

consumption of a batch reactor:

E^s, = K, • [cp • ATRM + AH]+ mES • AHf )+ imA-cAP+mw-c\ATA)+(K-A-ATAm-1-yPN)-t

where EpRVSt is the production dependent steam consumption (either 5 or

15 bar) of a batch reactor in kJ, m are the masses of the reaction mass (rm), the

evaporated solvent (es), the apparatus (a), or the water in the heating/cooling-

system (H/C-system) (w) in kg respectively, Cp represents the heat capacities of

the reaction mass (rm), the material of the apparatus (a), or the water (w) in

kJ / kg / K, respectively, AT represents the temperature increases of the reaction

mass (rm), the apparatus (a) or the temperature difference of the apparatus to the

ambient temperature (Am) in K, respectively, AHr is the reaction enthalpy16 in

kJ / kg, AHv is the heat of vaporisation in kJ / kg, K is the loss coefficient in

kW / m2 / K, A is the surface area of the vessel in m2, r\ is the efficiency of the

stirrer in %, fis the relation of actual power to nominal power consumption of

the stirrer in %, Pn is the nominal power of the stirrer in kW, and t is the batch

time in s. EpRVSt may be translated to kWh by dividing kJ with 3,600 s / h.

16Defined according to Atkins, P. W. (1990). Physikalische Chemie, VCH, Weinheim., i.e..

AHR > 0 representing endothermic reactions

46


Figure 5-5: Modelling and measurements of the steam consumption of reaction

vessels

With the help of this equation, modelling of the steam consumption of a re¬

action vessel was possible and the loss coefficient was fitted for the measured

vessels according to the data. The modelling results are presented in Figure 5-5.

It can be seen that the deviations between the measured and calculated steam

consumptions are reasonable. This shows that the postulated model accordingto Equation (5-1) is valid.

With the help of this model, the steam consumption of a batch vessel may be

analysed as shown in Figure 5-6. It can be seen that the losses are responsiblefor the biggest part of the steam consumption. As found by the modelling of the

brine consumption (see next subchapter), about 50% of these losses are due to

the radiation to the environment and the other 50% are losses through the steam

traps and the pipes. This is because each kg of steam introduced to the system

requires a kg of water to leave the system through the steam pipe as depicted in

Figure 5-1. This means that the water leaves the system at the hottest point -

without ever reaching the heat transfer area of the batch reactor. The normaliza¬

tion of the loss factor to an area basis is nevertheless valid since pipe and steam

traps dimensions are proportional to the size of the vessel.

47

Reactors

Heat of Reaction

nLosses

H Evaporation

B Heating up of apparatus

s Heating up of reaction mass

x Reaction Time

-Measured steam consumption

4 5 6

Batch No.

Figure 5-6: Modelling results of the steam consumption of a 10 m stainless

steel reaction vessel (in comparison with measured steam consumption and

reaction time)

Out of the different measurements and models, different loss coefficients are

found for the investigated batch reactors. These loss coefficients are summa¬

rized in Table 5-1. It is seen that the distribution of the loss coefficients from

the best to the worst equipment is wide. An average of about

3.3-10"2 kW / m2 / K was calculated. The lower values represent batch vessels

that operate at maximum performance and "ideal" conditions (i.e., the loss coef¬

ficient is in the same order of magnitude as the loss coefficient of the brine sys¬

tem as will be discussed in the next subchapter). These "ideal" conditions are

nevertheless not attained all the time and for all the apparatus in a production

plant. The influence of cleaning of a vessel is investigated separately in Chap¬ter C.3 in the Appendix. In the random sample of the investigations, many ves¬

sels were found operating at nearly ideal conditions. Discussions with expertsfrom the production plant and other industry experts showed that in usual opera¬

tion, fewer ideal conditions would be found (compare (Dahinden 2003)). From

these discussions, a loss coefficient about 25%> higher than the average one was

assumed more realistic (to account for the not ideal conditions in daily produc¬

tion). Therefore, a loss coefficient of about 4.2-10"2 kW / m2 / K was used in the

modelling of the steam consumption of the reaction vessels and nutsche dryers.

48


Table 5-1: Calculated loss coefficients for the steam consumption of the reac¬

tion vessels and nutsche dryers investigated

Reactor Type K

[kW/m2/K]10 m3, glass lined

10 m3, stainless steel

6.3 m3, glass lined

6.3 m3, stainless steel17

6.3 m3, stainless steel18

6.3 m3, glass lined

6.3 m3, glass lined (dirty)6.3 m3, glass lined (clean)

1.8

5.2

1.2

7.5

2.2

8.3

4.2

3.7

10"2

10"2

10"2

10"2

10"2

10"3

10"2

10"2

10 m2, stainless steel nutsche17'19



4.7

4.7

8.3

10"2

10"2

10"3

Average 3.3 10"2

The measurements for the high temperature reaction vessel shown in Table

D-l8 and the discussions mentioned in the electricity subchapter below resulted

in an easier steam model for this apparatus than for the other equipment. This

vessel requires steam just for the first part of the heating-up period. This is a

period similar for all the different batches observed (reactor is filled the same

way and the heating up is performed "as fast as possible"). Because of this, the

steam consumption was modelled as a constant (base) consumption for each

batch. The value observed in the measurements (about 530 kg /batch or

430 kWh / batch) was taken for the modelling.

Brine

Brine is used either for crystallization processes (i.e., cooling crystallization)or for reactions that have to be performed at low temperatures.

The general model for the cooling process is the same as presented in Equa¬tion (5-1) above for the steam consumption, just that this time, the reaction me¬

dia is cooled down. Most of the time, no solvents evaporate and the reaction

enthalpy is for crystallization processes replaced by the crystallization enthalpy,if known. For the processes conducted in the investigated building, no crystalli¬zation enthalpy was known. Because of the unique kind of the produced mole¬

cules, it was not possible to gather the crystallization enthalpies of analoguesmolecules. It was observed, that the crystallization processes often start at

higher temperatures than the one from which on brine may be used (i.e., the

cooling crystallization often starts at about 60 °C while brine may only be used

below about 30 °C). No data was available on how much of the product is al¬

ready crystallized when the switch from water to brine cooling is performed.

17With simultaneous heating and cooling

18Without simultaneous heating and cooling

19See Chapter 5.2

49

Reactors

Several discussions with industry experts on this problem were performed.These discussions lead to the assumption, that no significant part of the crystal¬lization enthalpy is released while cooling with brine. Reaction enthalpy, never¬

theless, is considered where applicable.The investigations on whether the linear model according to the preceding

chapter and Equation (3-16) is applicable are shown in Figure 5-7. The base

consumption (i.e., cooling down of the apparatus) and the cooling down of the

reaction mass for this process are about 150 kWh per batch. This stands in goodcorrelation with the theoretical value. The figure shows that a linear model with

only one time dependent (loss) term models reasonably well the brine consump¬

tion of this apparatus and any more complicated model would not be more ade¬

quate (cf. the problem of overfitting discussed in (Stahel 1995)). The correla¬

tion coefficient is not too high, but because of the large uncertainties and errors

of the brine measurement (see Chapter B.2.2 in the Appendix), a better correla¬

tion could not be expected.

EKjCo=-1706 t-157 53

R2 = 0 13

Cooling Time t [hours]

Figure 5-7: Measurements of the brine consumption of a 10 m stainless steel

vessel (regression according to Equation (3-16))

50


The modelling results for the same 10 m stainless steel reactor shown in

Figure 5-7 are presented in Figure 5-8. It can be seen, that the cooling down of

the apparatus is only of minor importance for the brine consumption (because of

the limited temperature difference). The cooling of the reaction mass is more

important, since the media has a higher heat capacity than stainless steel and the

apparatus contains a big amount of reaction mass that needs to be cooled down

compared to the mass of stainless steel (see Table D-9). Despite the long cool¬

ing times observed for the process (see Table D-9), the losses are small com¬

pared to the losses during heating with steam. This can be explained by the

smaller loss coefficient discussed below.

D Losses

H Cooling down of apparatus

Cooling down of reaction mass

— Measured Brine Consumption

s -loo

Figure 5-8: Modelling of the brine consumption of a 10 m stainless steel vessel

(according to Equation (3-16); in comparison with measured steam consump¬

tion)

All the loss coefficients investigated and found during the modelling of the

different vessels are presented in Table 5-2. It can be seen that theses values are

significantly smaller than the loss coefficients for the steam measurements pre¬

sented in Table 5-1 in the subchapter above. This may be explained by the fact

that for the brine measurements, the system is a simple input-output system and

no losses may occur due to steam traps. Simultaneous heating and cooling is

also not possible while cooling with brine (this would be noticed instantane¬

ously by contamination of the water). It can therefore be seen, that about 50%

of the losses observed at the steam measurements were caused not by losses

through irradiation but by losses through the steam traps and other suboptimalprocedures during the heating period.

51

Reactors

Table 5-2: Loss coefficients for the brine measurements of the investigated re¬

action vessels

Reactor Type K

[kW / m2 / K]10 m3, stainless steel

10 m3, glass lined

6.3 m3, glass lined

4 m3, glass lined

3.3-10"3

5-10"3

2.2-10"2

3.3-10"3

Average 8.4-10"3

Modelling of the brine consumption with the found loss coefficients is de¬

picted in Figure 5-9. As expected, the deviations between measured and mod¬

elled consumptions are larger than for the steam measurements. Deviations of

20%) and more may occur. Because of the uncertainties in the measurement of

the brine consumption (see Chapter B.2.2 in the Appendix) and some other un¬

certainties of the parameters (e.g., neglecting of the crystallization enthalpy),this is not surprising. Nevertheless, modelling of the brine consumption with

reasonable accuracy is possible with the help of this simple equation.

-500 -450 -400 -350 -300 -250 -200 -150 -100 -50 0

Measured Brine Consumption EpuCo [kWh]

010 m3, stainless steel reaction vessel D10 m3, glass lined reaction vessel

A6 3 m3, glass lined reaction vessel 04 m3, glass lined reaction vessel

Figure 5-9: Modelling of the brine consumption (according to Equation (3-16))vs. measurements

The small values of the loss coefficient K depicted in Table 5-2 (comparedto steam as shown in Table 5-1) are nevertheless not considered as reasonable

for all of the apparatus for the same reasons discussed above for the steam

measurements. As seen in Table 5-2, one of the investigated 6.3 m3 glass lined

reactors has a significantly higher loss value. Since the brine consumptioncould only be measured for a few reaction vessels, it is not clear whether the

high value of K for the 6.3 m3 glass lined reactor is an exception or more stan-

52


dard for most of the vessels. In the 4 m3 glass lined reactor, moreover, a diffi¬

cult reaction is going on with evaporation of a reaction product that is not com¬

pletely understood by the industry experts in energetic aspects. During discus¬

sions with industry experts, it was concluded, that a value of the loss coefficient

K of about 1.7-10"2 kW / m2 / K is reasonable for reaction vessels for brine cool¬

ing. This value was taken as the standard value during modelling of the whole

building (see Chapter 6).

Electricity for the High Temperature Reaction Vessel

The electricity consumption measurements of the high temperature reaction

vessel shown above have a complicated behaviour not suitable for simple mod¬

elling. Only one reactor uses electric heating. Since always the same product is

produced in this reactor, a model according to Equation (3-8) was postulated for

the base consumption (circulation pump) and according to Equation (3-13) for

the consumption of heating energy respectively.The continuous base consumption of the circulation pump may be measured.

The base consumption is about 16.3 kW according to the measurements. It is

consumed during the whole batch time. According to the nominal power of the

circulation pump of 21.5 kW, this represents 76% of the nominal power (y). For

the programming of the BOTUMO of the whole building, it was required to be

able to use the same parameter values (i.e., values of y) for all circulation pump

calculations according to Equation (3-13) (see Chapter 5.5 as well). Therefore,the nominal power was set, virtually, to about 19.2 kW. This enables the pro¬

grammer to use the same value of y for all the circulation pump models in the

BOTUMO. The specifications of the circulation pump of the high temperaturereactor differ significantly from the specifications of the other circulation

pumps. The different specifications result in a different value of y Since this

would increase modelling effort significantly (i.e., a new circulation pump

model just for one reactor), the nominal power was adapted in a way to be able

to use the same value of fas for the other circulation pumps (i.e., to attain the

same value of P as would be the outcome of the calculation with the actual

nominal power and the actual y).For the electricity consumption dependent on heating, an average value of

about 100 kW was computed. Measurements of the batch time showed, that

only minor deviations from the average value occur (see Table D-10). The

deviations in batch time are about 4%. This is more accurate than the expected

accuracy of the model. Therefore, the mean value could be taken as constant

consumption during the whole batch time. This leads to a model similar to

Equation (3-13). Here, C is the constant consumption of electricity of about

100 kW. Although this represents not exactly reality, it is sufficient to model

the electricity consumption per batch with the required accuracy. Moreover, it

makes the model significantly easier and allows a simple model without the re¬

quirements of many parameters.

53

Nutsche Dryer

5.2 Nutsche Dryer


A general picture of the nutsche dryers is shown in Figure 5-10. The heat¬

ing/cooling-system is similar to the one of the reactors described in Chapter 5.1,

despite it has no possibility to cool with brine. The walls of the nutsches are

heated. The stirrer itself is heated too for optimising the heat transfer to the me¬

dia to be dried (similar to the horizontal vacuum rotary dryer described in Chap¬ter 5.7). Investigations had shown that about 40% of all heating energy is intro¬

duced to the system through the heated stirrer (efficient heat transfer).

Steam Water

Figure 5-10: Scheme of a nutsche dryer with its heating/cooling-system

Moreover, the nutsche dryers may be evacuated by vacuum pumps for

evaporation of the solvent of the wet product. These vacuum pumps utilize

brine to condensate the evaporated solvents and prevent, therefore, a contamina¬

tion of the exhaust air with solvents (these APOVAC pumps are described in

more detail in Chapter 5.4.2).An overview of literature on drying processes is given later on in Chap¬

ter 5.7. The principles for the operations of nutsche dryers are given in

(Perlmutter 1992). In (Martin 2003) a good overview of the development of the

description of heat and mass transport phenomena for the past 75 years is given.Different techniques for drying are described in (Gehrmann 2003). Energy as¬

pects of drying are discussed in (Strumillo et al. 1995). In this book, it is stated

as well that about 6%> of total energy used in the British and French chemical

industry is used for drying. Drying is therefore an energy intensive unit opera¬

tion with high influence on industry.

54


5.2.2 Measurements

Measurements were taken during two periods. During the first period, it

was observed, that the heating/cooling-system was badly tuned. The systemtried to hold exactly the given temperature. This resulted in a fast change be¬

tween short heating phases followed by short cooling phases. As a result, the

energy consumption increased significantly. The operation mode was therefore

changed by re-programming the steering software. After the change, the second

measuring period was performed. These second measurements clearly showed

the decrease in energy consumption and the improvement of operation time (seenext subchapter for a detailed description). The measurements are summarized

in Table D-l 1 and Table D-12.

The measurements show that the elimination of the simultaneous heatingand cooling regime saved a significant part of the steam consumption (fornutsche 2 and product 2). The steam consumptions are now rather low and ac¬

curacy is therefore not guaranteed anymore. The model consists of several un¬

certain parameters. As mentioned in Chapter B.l in the Appendix, the steam

measurement itself has an uncertainty of about ±10%. Given the small values of

steam consumption, even small deviations in other parameters may result in sig¬nificant deviations from the actual value as discussed in the following subchap¬ter.


The model introduced for the calculation of the steam consumption of the

nutsche dryers is based on Equations (3-5) (heating of the substances), (3-6)

(evaporation of the solvents), and (3-7) (losses), reduced by the use of the inputthrough mechanical energy provided by the stirrer according to Equation (3-9).With the help of these equations, the batch time, the known drying temperaturesand the specification of the apparatus, a model according to Equation (3-16) wasbuilt. The model is similar to Equation (5-1). Only here, no reaction occurs and

reaction enthalpy equals therefore zero. By adapting the loss coefficient K of

the model, modelling of the energy consumption of the nutsche dryer was pos¬

sible. The regression on the loss coefficients led to a value of

4.7-10"2 kW / m2 / K. This value is in the upper range of the values observed for

reaction vessels and may be explained with the simultaneous heating and cool¬

ing of this equipment.

55

Nutsche Dryer

The results of the modelling of the nutsche dryers are presented in Figure5-11. The deviation between the calculated and the measured steam consump¬

tion lies in the region expected for this unit operation (most of the time about

±10%) but sometimes even more). Drying is, as mentioned in Chapter 5.7, a

complicated unit operation dealing with solid, liquid and gaseous phase and dif¬

ficult transfer conditions. This simple model leads, nevertheless, to a good pre¬

diction of the steam consumption of this unit operation.

E 2000

75 500O

+20%

+10%

-10%

'20%

.'£

1000 1500 2000

Measured Steam Consumption EplNDSt [kg/batch]

O Nutsche 10m2 Product 1 ANutsche 10m2 Product2

Figure 5-11: Modelling according to Equation (5-1) and measurements of the

nutsche dryer with simultaneous heating and cooling

56


In Figure 5-12, the results of the steam modelling for nutsche 1, drying

product 1 is presented. Since the temperature stayed constant (always the same

product), the losses are directly proportional to time. It can be seen that losses

are significant and account for a big part of total steam consumption. Because

nutsche dryers are large and heavy equipment, the heating of the apparatus to

the process temperature requires a significant part of the total steam consump¬

tion. The actual production consumption (i.e., the heating of the product, the

solvent and the evaporation of the solvent) requires about 30 to 50% of the total

steam consumption. By minimising the losses of the system, the energy effi¬

ciency of the apparatus could be improved. Since losses are proportional to

time, this is especially important for products with long drying times. In gen¬

eral, drying times should be decreased for saving energy.

12 3 4 5 6

Batch No.

BHeating and Distillation of Substances B Heating of Dryer DLosses XDrying Time —Measured Steam Consumption

Figure 5-12: Modelling according to Equation (5-1) of the drying of product 1

in a 10 m2 nutsche dryer (in comparison with measured steam consumption and

drying time)

Since nutsche dryers are quite heavy equipment, optimising the design of

these apparatus could reduce energy consumption. Reducing the weight of

these apparatus and minimising the water content of the heating/cooling-system(or change to steam condensing in the heating jacket) would minimise the steam

consumption for the equipment.The loss factor determined for the nutsche dryers after the simultaneous

heating and cooling was eliminated was determined to be approximately zero.

This outcome shows the limitation of this simple model. Because several pa¬

rameters could not be determined exactly (such as the exact amount of distilled

solvent, the heat capacity of the product, the temperature inside the nutsche

dryer, etc.) and because of the measuring error of the steam measurement

equipment, the deviation became large. A determination of the exact loss coef¬

ficient was not possible anymore. It is significantly smaller than the one with

57

Heat-Chamber

simultaneous heating and cooling and is definitely larger than zero. The exact

value could nevertheless not be determined from the measurements because of

the measuring error.

The model for the brine consumption of the APOVAC pumps (see Figure5-10) is given later on in Chapter 5.4.2.

5.3 Heat-Chamber


The scheme of a heat-chamber is given in Figure 5-13. This equipment is

used for heating and melting of solids for easier transfer into the reactors. A

heat chamber is operated by direct heating using 5 bar steam, which condensates

inside the double-jacket. A small ventilator guarantees a uniform temperatureof the air inside the heat chamber.

Steam

[>o

1Condensate

Figure 5-13: Scheme of a heat-chamber

5.3.2 Measurements

The steam consumption of the heat-chamber was measured first for itself

(i.e., without any barrels inside) to investigate whether or not the losses are con¬

stant over time and proportional to the temperature. After these experiments,two 200 1 barrels filled with about 150 kg of water were put inside and the trans¬

ferred heats as well as the losses were investigated.


A model based on Equation (3-16) was postulated for the heat-chamber. It

is based on the heating up of the substances, the material of the heat-chamber

and the air volume inside the heat-chamber according to Equation (3-5) and the

losses to the ambient according to Equation (3-7). The detailed equation for the

steam consumption of a heat-chamber is shown in Equation (5-2).

58


ma^So AT

-I HC St

(mA-Cp m^Air

Air )-AV -K-A.ATAm.t

3600 s/(5-2)

h

where EpHCSt is the production dependent steam consumption of the heat-

chamber in kWh, ruß is the mass of the filled barrel in kg, Cp° is the heat capac¬

ity of the (organic) solid compound in kJ / kg / K (i.e., 2.5 kJ / kg /K for a ge¬

neric compound; see (Perry et al. 1997)), ATq is the temperature rise of the bar¬

rel and its contents in K, itia is the mass of the heating chamber (apparatus) in

kg, Cp is the heat capacity of stainless steel in kJ / kg / K, rriAir is the mass of air

inside the heating chamber in kg, Cpir is the heat capacity of air in kJ / kg / K,AT

a is the temperature rise of the heat-chamber in K, K is the heat transfer coef¬

ficient to the environment in kW / m2 / K (loss coefficient), A is the surface area

of the heat-chamber in m2, ATAm is the temperature difference between the out¬

side wall and the environment in K, and t is the time in s. The scaling factor of

3,600 s / h is required for converting kJ in kWh.

It is assumed that the chamber stays at the set temperature for the whole

time, that it consists of stainless steel and that only half of the chamber material

is heated to the set temperature (the temperature of the outside walls is alwayssignificantly lower than the set temperature (i.e., temperature inside the heat-

chamber). Moreover, it is assumed that the steam trap works correctly and that

the steam is entering as saturated 5 bar steam.

The results of the experiments are presented in Figure 5-14 and in Figure5-15.

s

st so

u 5

Sj S 40

5 "

- £

E 5 30

10 -

;0 0

X

'-- 0

- 0 0 0

O 0

0

-

.

-- 0 II

:m r*i n n

mn

niX 1

—1— —1—

-

3a 3b 2 4 5

Experiment ID

Steam for Heating Chamber a Steam for Heating ofWater

[I] Steam for Losses x Experiment Duration

oTemperature Increase of the Heat-Chamber internal —Measured Steam Consumption

16

14

12

10 S

m

D

6 I4

2

0

Figure 5-14: Measured and modelled (according to Equation (5-2)) steam con¬

sumption and experiment duration for the heat-chamber

59

Heat-Chamber

The heat transfer between the hot air inside the heat-chamber and the barrels

is rather low. Moreover, the content of the barrels is not stirred, what decreases

the heat transfer coefficient too. This results in high occupation time of the ap¬

paratus. This, again, results in high losses compared to the actual energy trans¬

ferred to the barrels (according to Equations (3-7) and (5-2)).

0 10 20 30 40 50 60 70 80

Measured Steam Consumption E lHcst [kWh]

Figure 5-15: Measured vs. modelled steam consumption of the heat-chamber

(according to Equation (5-2))

As discussed in (Dahinden 2003), the model for the heat-chambers is still

too complicated for industrial practice. Since this apparatus consumes only a

minor amount of the steam consumption of the building, an easier, though not

equal accurate model is justified. The new model is still based on Equa¬tion (5-2) mentioned above. Some of the parameters, nevertheless, are fixed for

simplicity reasons (i.e., not all parameters have to be investigated each time the

model is used). The mass and the heat capacity of the substance in the barrels is

fixed at a value of 1,200 kg and 2.5 kJ/ (kg • K), respectively and as outside

temperature of the wall a temperature about 40 °C lower than the inside tem¬

perature is taken as standard. More, the air inside the heat-chamber is ne¬

glected. The melting enthalpies of the substances are neglected too (see the as¬

sumptions in Chapter A. 1.1 in the Appendix). As Figure 5-14 shows, the main

steam consumption is caused by the losses of the system. Neglecting the melt¬

ing enthalpies is therefore resulting in only minor deviations (see (Lide 1995;

Perry et al. 1997) as well).

60


5.4 Vacuum Pumps

Different vacuum pumps are present in the examined production plant. The

three main types of vacuum pumps, which are used in the building, will be de¬

scribed shortly in the next paragraphs.

5.4.1 General Vacuum Pumps

Description of the Equipment

A ventilator and a compressor are built together in one apparatus to reach

lower levels of vacuum. Evaporated solvents are condensed either before or

after the vacuum pump or discharged to the vent. The vacuum pumps are not

regulated but either turned off or operated at full capacity. The vacuum pumps

discussed here are operated especially for the one or the other process and there¬

fore measured separately. The general vacuum pumps running for all the proc¬

esses are accounted for in the infrastructure consumption as discussed in Chap¬ter 5.6.1.

Measurements

Electricity measurements were performed for different vacuum pumps.

During several days of operation, the electricity consumption of each vacuum

pump was measured. Several start-up as well as shutdown procedures were

considered for each pump. Different levels of vacuum and gas flow are consid¬

ered. A typical measurement is shown in Figure 5-16. It can be seen from this

picture, that, besides some peaks, the electricity consumption of vacuum pumps

is quite stable and independent from the actual level of vacuum. The peaksmight occur because of synthesis of high gas volumes in the reaction or because

of fluctuation in voltage.

15r~

14:-

13:-

12:-

11 :-

10:-

9:-

w 7;-c

o

O 6:-

Io

°-

Q.

4:-

3:-

2:-

1 :-

Mean = 0 47 *

P» XÏNU

Measurements^j

16 05 2003 17 05 2003 18 05 2003

Date &Time

Figure 5-16: Typical measurement of the electricity consumption of a vacuum

pump (here: Pn= 16.5 kW)

61

Vacuum Pumps

Equation (3-10) explains why the power consumption of the vacuum pump

stays constant over time. The better the vacuum, the higher is the volume that is

required to be extracted from the vessel. The product of pressure and volume,

nevertheless, stays constant.

The different measurements are summarized in Table D-13. It can be seen

that the power consumption of the different vacuum pumps compared to their

nominal power is about 50%. The model should reflect this fact. It is discussed

in the next paragraph.

Model and Conclusions

As can be seen in the preceding subchapter, the electricity consumption of

vacuum pumps stays constant over the time of one batch and between different

batches. Therefore, a linear model, only dependent on nominal power and op¬

eration time is postulated for these unit operations according to Equation (3-8)with y being equal to 50%.

5.4.2 Anti Pollution Vacuum Pumps (APOVAC)


The APOVAC pumps are vacuum pumps designed for releasing a minimum

amount of solvents in the vent20. They are based on the principle of water ringvacuum pumps using the main solvent in the vent as ring fluid. Moreover, theyhave a condenser operated with brine at low temperatures that condenses most

of the solvent content of the gas stream.

Therefore, these vacuum pumps require not only electricity for their opera¬

tion but brine as well. Because of the different construction (water ring system),they have a different electric efficiency compared to the other vacuum pumps

presented in Chapter 5.4.1. This is shown in the next paragraph.

Measurements

Some measurements were made of an APOVAC vacuum pump. The meas¬

urement taken over the longest period is shown in Figure 5-17. It can be seen

that these pumps have high peak consumption at the beginning of their opera¬

tion. This is quite common for electric motors (i.e., the power consumption is

highest at the beginning of operation because of acceleration starting from

standstill). This peak, nevertheless, does not influence the mean consumption

greatly. The relation of the mean consumption to the nominal power is, as can

be seen in Figure 5-17, higher than the one of the other vacuum pumps (i.e.,

bigger value of y in Equation (3-8)). This can be explained by the fact, that

these vacuum pumps are designed for the purpose of the nutsche dryers. Be¬

cause the actual purpose of the vacuum pumps is known, better sizing was pos¬

sible during the design phase. This results in a higher ratio of actual to nominal

power of about 62% as shown in Figure 5-17.

See e.g., http://www.ddpsinc.com/GasLiquid.htm, or

http://www.rosenmund.com/vacsyssolvrec.html

62


'

Mean = 0 62 *

P„ 1 1 II

\

-

Jr4

i—i—i—i— 1 1— 1 1

01 11 2002 02 11 2002 03 11 2002 04 11 2002 05 11 2002 06 11 2002

Date & Time

Figure 5-17: Measurements of the electricity consumption of the APOVAC

pumps (P/v = 27 kW)

Measurements for the brine consumption of the APOVAC pumps were also

undertaken. As mentioned in Chapter B.2 in the Appendix, the flow meter re¬

quires a certain straight length of the pipe for accurate measurements. Since this

was not available directly at the pump, the flow in the main pipe (feeding pipefor all of the four APOVAC pumps) was measured, while only one pump was

working. Unfortunately, this reduces the accuracy of the whole measurement.

A selection of the measurements (temperature of the brine inflow and outflow

(InT and OuT, respectively and the brine flow m) is shown in Figure 5-18 to¬

gether with the calculated brine consumption EPAP0VAC Co according to Equa¬

tion (3-5). All the measurements and calculations are summarized in Table

D-l5. The APOVAC pumps are only used during the thermal drying in the

nutsche dryers as indicated in Figure 5-18.

63

Vacuum Pumps

Dates Time

| InT — —OuT Brine Flow Cooling Energy

Figure 5-18: Measurements of the cooling media consumption of the APOVAC

pumps (calculation of the cooling energy consumption according to Equa¬tion (3-5))

The measurements indicate, not surprisingly for brine measurements, a high

variability (high standard deviation) of the brine consumption as shown in Table

D-l5. Considering the high inaccuracies inherent in the measuring method, as

stated in ChapterB.2.2 in the Appendix, a value of about 30 kWh/h (i.e.,30 kW) of operation seems feasible for the brine consumption of an APOVAC

pump as it is installed in the investigated building.


For electricity, Equation (3-8) was used as for the other vacuum pumps.

The only difference was that a value for y of 62% was used for the modelling as

discussed above.

For the brine consumption, a model with constant consumption over time

was postulated and congruent with the measurements. The basic equation for

the brine consumption model is analogous to Equation (3-13) discussed above,

only this time, the constant C is the brine consumption of 30 kWh/h or

8.3 Wh /s.

The electricity and the brine consumption of the APOVAC pumps may not

be neglected and are significant for the consumption of the whole building.This will be discussed in more detail in Chapter 6.

64


5.4.3 Steam-Jet Vacuum Pumps


The principle of a steam-jet vacuum pump is presented in Figure 5-19 and

may be found in (El-Dessouky et al. 2002) or in (Baier 1989; GEA.a; GEA.b;

GEA.c; Hinrichs 1991). A jet of steam flows through a venturi-kind nozzle. By

flowing through this nozzle, the steam jet drags gas from the vacuum side with

it, producing a vacuum. By connecting several such units, high vacuum levels

may be attained.

IPrmmxj

lUxmttQM m th« Ejaetor

Figure 5-19: Principle of a steam-jet vacuum pump (El-Dessouky et al. 2002)

The big advantage of these vacuum pumps is their resistance to acids (theyare usually constructed of glass or ceramic) and their low maintenance level,since no moving parts are required.

In (Heuser 2003), the interesting history of the main producer company of

steam jet pumps (GEA Wiegand) is presented.

Measurements

Before measurements were conducted, discussions with a main producer

company of steam jet pumps were performed21. These discussions showed that

measurements were not required for these equipments. A steam-jet vacuum

pump is an apparatus that is designed and produced for one specific task (i.e.,

specific gas flow rate and vacuum level). It is not possible and useful to try to

adjust the flow of steam though the equipment since the nozzle has only one

point of optimal operation. The steam consumption is not dependent on the

21Discussions with Mr. A. Riatti from GEA-Wiegand on July, 31th 2002

65

Stirrers and Motors

level of vacuum. An example of different steam consumptions of different

kinds of steam-jet vacuum pumps can be seen in Table D-14.

The steam-jet vacuum pumps used in the plant investigated required a flow

of 115 kg / h of 5 bar steam for their operation.


The general model for the steam consumption of steam-jet vacuum pumps is

presented according to Equation (3-13). Here, the constant C reflects the con¬

stant consumption of one steam-jet vacuum pump, given by industry data as

115 kg / h (i.e., 0.032 kg / s or 2.58 kWh / s).

Although this model is quite simple, compared to the models of the other

unit operations, it is possible to predict the steam consumption of this unit op¬

eration quite accurate according to industry experts. Moreover, steam-jet vac¬

uum pumps are only operated when needed and shutdown if not in use for en¬

ergy saving reasons. Therefore, this unit operation has only minor contributions

to the steam consumption of the whole plant and models that are more detailed

are not needed (see Chapter 6).

5.5 Stirrers and Motors


The motors and the actual stirring equipment (stirrers) are present in every

reaction vessel and in some storage vessels (see Figure 5-1). Many different

stirrer types exist. The motors, on the other hand, differ only in their nominal

power. Standardization is common in industry. The different stirrer forms in

use in the investigated building are presented in Table 5-3. These stirrer types

represent the most often used in chemical industry.For motors, two different general types of regulations are possible: fre¬

quency converters and staged motors. In the investigated building, frequencyconverters are installed at about 75% of all stirrer motors. The other 25% are

two-staged motors (i.e., motors with two different coils allowing them to oper¬

ate at two different speeds).

66


Table 5-3: Different kind of stirrers used in the investigated building22

Stirrer ID Name Picture

Cross-Blade

(Double) Anchor

Blade or Paddle (1 or 2 stages)

Impeller

Intermig

5.5.2 Measurements

Electricity consumption for all abovementioned stirrer types (some glasslined, other stainless steel) for different vessel sizes and different motor types

are measured with the Memobox described in Chapter B.3 in the Appendix.These measurements were either taken directly at the plug of the motor (behindthe fuse) for the two-staged motors or after the frequency converter for the mo¬

tors that had a frequency converter. The measurements had to be taken after the

frequency converter (although this device uses electricity as well) because the

Source: http://www.rvt-systeme.de/html/ruhrorgane.html,

http://www.gotoppi.com/images/mixer/mixl23.gif, and company internal datasheets

67

Stirrers and Motors

frequency converter corrupts the sinusoid curve of the alternating current and

this has a significant influence on the characteristics of the motor consumption.Measurements were also taken by the standard installations inside the plant

where available. The electricity consumption of some motors is used to super¬

vise the stirring performance. For these motors, a measurement of the electric

current is installed permanently. The assumption that the voltage stays constant

was proven by measurements. The power consumption may be calculated from

the current and the voltage according to Equation (3-11).All motors in the building related to production operations are using a volt¬

age of 500 V. Only some parts of the infrastructure are working at a voltage of

400 V. This will be discussed in Chapter 5.6.1.

The motors not measured directly but with the permanent measurement in¬

stallations, have, nevertheless, a higher inaccuracy (e.g., unsteadiness of the (not

measured) voltage, less accurate measurements of the current, COScp that is not

constant (and not measured) but required for Equation (3-11)). Therefore, the

direct measurements of the electricity consumption with the Memobox are con¬

sidered as the more significant measurements. A typical measurement of a stir¬

ring process over several batches for different vessels and stirring equipment is

shown in Figure 5-20, in Figure 5-21, and in Figure 5-22.

»...-. f

r

"40 j?

20 03 2003 12 00 21 03 2003 12 00 22 03 2003 12 00 23 03 2003 12 00 24 03 2003 12 00 25 03 2003 12 00

Date & Time

-Povrer Consumption ^^~|T

Figure 5-20: Power consumption (P) and temperature of the reaction mass (IT)for a 6.3 m3 stainless steel vessel with an Intermig stirrer

From Figure 5-20 it can be seen that the energy consumption of the stirrer

motors is well reproducible between different batches. The motor uses about

27%) of its nominal power at each batch. Nothing can be said about the tem¬

perature (viscosity) dependence of the power consumption. This is seen more

clearly in Figure 5-21. Although the temperature is changing over a wide range,

the power consumption has no clear dependence from this change in tempera¬

ture (and viscosity).

68


26 02 2003 00 00 27 02 2003 00 00 28 02 2003 00 00 01 03 2003 00 00 02 03 2003 00 00 03 03 2003 00 00

Date & Time

Power Consumption IjJ

Figure 5-21: Power consumption (P) and temperature of the reaction mass (IT)of a 6.3 m3 glass lined vessel with an Intermig stirrer

In Figure 5-22, a temperature dependence of the electricity consumption of

the stirrer motor is also not seen clearly (i.e., given the inaccuracy of the meas¬

urements). Moreover, no clear and significant dependence of the power con¬

sumption from the revolution velocity of the stirrer can be seen.

15 05 2002 16 05 2002 17 05 2002 18 05 2002 19 05 2002 20 05 2002 21 05 2002 22 05 2002

Dates Time

Power Consumption Rounds per Minute ^^^IT

Figure 5-22: Power consumption (P), rounds per minute, and temperature of

the reaction mass (IT) of a 6.3 m3 stainless steel vessel with a Cross-Blade stir¬

rer

69

Stirrers and Motors

Although these pictures are only examples, the same results were found for

all the measured stirrers. This stays in contradiction to the general equationgiven in Equation (3-12). This finding is, nevertheless, congruent with many

other studies if only the turbulent region of mixing is considered (see e.g.,

(Ellermann 1991)). Several authors state, and have measured, that the power

consumption stays constant in the turbulent region, even with increasing revolu¬

tions per minute or decreasing viscosity (see e.g., (Benz 2003; Bertrand et al.

1980; Liepe et al. 1998; Ng and Yianneskis 2000; Taca and Paunescu 2001;Zehner 2003)). Industry experts agreed on the assumption (see Chapter A. 1.1 in

the Appendix) that turbulent behaviour of the reaction media is guaranteed in

the baffled reaction vessels used in the investigated building.Another parameter of influence in Equation (3-12) is the diameter d of the

stirrer. In changing the diameter of a stirrer, the power consumption is increas¬

ing significantly according to industry experts. Since all the stirrers in the in¬

vestigated buildings are dimensioned similarly (i.e., the ratios between the di¬

ameter of the vessel and the stirrer are similar), no influence of the stirrer di¬

ameter is observed in the investigations. This should be considered, neverthe¬

less, when the model is transferred to another building.All the measurements (mean values of the power consumption) are summa¬

rized in Figure 5-23. It can be seen that the measurements of the Memobox are

comparable to the ones with the installed equipment except for some outliers for

stirrer type 3 (see Table 5-3). The average of the actual power consumption in

relation to the nominal power of all of the Memobox measurements results in a

relation of P to P/v of 28%. This value is required for the model described in

the next paragraph.

o

o

X

8

X

X

X

X

12 3 4 5

Stirrer ID (see Table 5-3)

XMemobox Olnstalled Measurements

Figure 5-23: Measurements of the relation of power consumption to nominal

power P/Pn of different stirrer types

E 30 -

<

<

Ü 20 :a»

Î

70


Another electric equipment required for the operation of a batch reaction is

the circulation pump of the heating/cooling-system. This pump is constantlyrunning whenever a vessel is heated or cooled. Therefore, it could result in sig¬nificant power consumption although circulation pumps are most of the time

small equipment items. Measurements showed that the power consumption is

constant and that circulation pumps have a high ratio y between actual power

consumption and nominal power of about 85%.


A model according to Equation (3-8) was postulated for the stirrer motors

and the circulation pumps of the heating/cooling-system. The value of y was

found by the measurements discussed above. For the stirrer motors, it was

found to be about 28% and for the circulation pumps, it was found to be about

85%.

The circulation pumps have a higher relation between actual power con¬

sumption and nominal power. This can be explained by the fact that the differ¬

ent fluids that need to be circulated (i.e., brine and water) are well known and

may be assumed as staying constant over the lifetime of a vessel. Therefore, the

engineer is able to design the circulation pumps accurately. The reaction media,on the other hand, may vary heavily in terms of viscosity, stirring requirementsand solid contents (probably there is once in the future a crystallization going on

in the vessel). The worst thing that could happen is to undersize a motor for

stirring so that the reaction media sticks in the vessel without stirring (danger of

thermal runaway, etc.). Secondly, high power may be required if the stirrer is

shutdown accidentally during operation (e.g., electric power outage). To restart

stirring the reaction mass, high power is required. Otherwise, restarting the op¬

eration would be impossible. To prevent these worst-case scenarios, the motors

are oversized to keep on the safe side. This is reflected by the low relation of

actual power consumption to nominal power in daily operation as seen by the

measurements.

71

Continuous Equipment

5.6 Continuous Equipment

Although the investigated plant is a batch plant, some continuous equipmentis run. These equipment units may be responsible for a significant amount of

the total energy consumption. Therefore, special models, accounting for the

specific operation and specifications of the continuous equipment were elabo¬

rated to guarantee an accurate modelling of the whole plant.

5.6.1 Infrastructure


As infrastructure, the following equipment of the building was regarded:Ventilation and waste air systemsGeneral vacuum pumps

LightsBrine pumps for the whole buildingOffice computers

HeatingLosses of the steam and brine systemMinor general electricity consumers like the accumulators for the

forklifts

Most of this equipment uses electricity as the major energy source. Actu¬

ally, only the losses of the steam and brine system (radiation from the pipes and

losses through steam traps) influence the steam and brine consumption. Elec¬

tricity on the other hand offers no possibility for losses. The only possibility to

save electricity is to switch-off unused equipment and to dimension the equip¬ment (e.g., air ventilators) accurately.

Measurements

The losses through the steam pipe and the brine system were measured dur¬

ing a shutdown period as described in Chapter 4.3.

The measurements of the electricity consumption are described in Chap¬ter 4.3.2. A short feasibility calculation will now show whether the measured

values are higher, lower or comparable to industry standards. As mentioned

above, the power consumption of a ventilator (or vacuum pump) may be calcu¬

lated according to Equation (3-10). A feasible pressure difference for an air

ventilator is about 2000-3000 Pa according to discussions with industry experts.This results in a power consumption of 0.6-0.8 W / (m3 / h) of air change. In the

investigated building, the ventilation system delivers about 30,200 m3 / h of

fresh air. This results in a power consumption of about 18-24 kW of the ventila¬

tion system alone. In one day, this leads to an electricity consumption of 432-

576 kWh. The measurements showed a consumption of the ventilation systemof about 40-50 kW, resulting in a daily consumption of about 960-1200 kWh.

This is in the same order of magnitude and therefore reasonable.

72


All the measurements of the 500 V electric infrastructure consumption are

shown in Table D-l6 (the lighting uses 400 V electricity and is discussed in

Chapter 4.3.2).


The measurements confirm the constant behaviour of the infrastructure con¬

sumption. A model similar to Equation (3-13) was postulated. This time C was

defined according to the values given in Table 5-4.

Table 5-4: Base Consumption of the investigated building

Energy Form Base Consumption

[kW]

Electricity (500 V)

Electricity (400 V)Brine

Steam (5 bar)Steam (15 bar)

150

30

20

< 100

< 100

5.6.2 Short-Path Distillation Column


Feed »I

Heating |:Jacket H

1

"

Wiper-

system I 11 1

Condenser—P 11 >*—11

11 ^ ^y mm

J*-Vacuum

jk JJ— Cooling

f"Residue

tDisti ate

A short path distillation column as

depicted in Figure 5-24 is used for

splitting mixtures with high boilingpoints. High vacuum levels lower the

boiling point to a reasonable value.

The advantage of this equipment is

that the liquids remain only short time

at the high evaporation temperatures.

Immediately after evaporation, theyare cooled down again. The compact

design is also advantageous. More¬

over, absolute temperature of evapora¬

tion is reduced, since high vacuum

levels are achieved with this equip¬ment.

Figure 5-24: Scheme of a short pathdistillation column

23

Source: http://www.reactorvessels.demon.co.uyShort_Path_Distillation/a_VTAShrtPthDr.gif

73


Measurements

This equipment runs under continuous conditions. Therefore, start-up and

shutdown behaviour is not regarded. The short-path distillation is heated with

electricity and requires brine for cooling down of the solvents. The whole (elec¬tric) equipment of the distillation column includes circulation pumps, vacuum

pumps, the electric heating, and the stirrer.

Measurements were taken for the brine consumption and the total electricity

consumption. The total electricity consumption of the equipment was measured

since the continuous operation of this equipment results in constant consump¬

tion of all of the different parts of the distillation column.

The measurements of the electric consumption are shown in Figure 5-25.

The measurements of the brine consumption can be seen in Figure 5-26. The

brine consumption was calculated according to Equation (3-5).

50

45

40

W 355^

0.

c 30o

Q.

E 25

(/)

8 20

Ol

I 15

10

5

>'

/^

^/

tiiiiiiinui.i.i.ii..i LiiiiJhiiBiikniÉib iiMil^^ttiiiMtajiJjJiyiiBiiiiiiiiiiiiÉjyiy>iMiitoLÉiilJlMlMlhyMUiMJi I H^uUujtiu

o

08 07 2003 09 07 2003 10 07 2003 11 07 2003 12 07 2003 13 07 2003 14 07 2003 15 07 2003

Date & Time

Power Consumption Distillate Temperature - - - Heating Fluid Temperature — —Temperature of Raw Material

Figure 5-25: Measurements of the total electricity consumption of the short

path distillation column

Figure 5-25 shows the measured total electricity consumption of the short

path distillation column. Except for the period during July 13th, the electricityconsumption is constant at about 22.9 kW. During this period, the set tempera¬

ture of the electric heating was lowered for experimental reasons (and neglectedfor calculating the mean). A clear influence on the electricity consumption can

be seen. During normal operation, nevertheless, the temperature stays constant

and the mean electricity consumption may be regarded as the correct value for

modelling.

74

0 j-

-2 :-

-4 :-

-6 :-

a»

| -10 -

a»a.

I -12 :-

t-

-14 :-

-16 :-

-18 :-

-20 :-

08 07 200

Brine Outlet Temperature Brine Inlet Temperature — —Brine Flow Rate Cooling Energy Consumption

Figure 5-26: Measured brine consumption of the short path distillation column

As can be seen from Figure 5-26, the mean value of the brine consumptionis about 3.6 kW. This value is used for the cooling energy model described in

the next paragraph. The peak at the beginning of the measurements was ne¬

glected because it was regarded as an outlier in the temperature measurements.


The model for the calculation of the electricity consumption is postulatedaccording to Equation (3-8). The measured average power of this apparatus

(i.e., 22.9 kW as mentioned in the preceding chapter) is about 96% of the nomi¬

nal power of all the electric equipment of this apparatus. Therefore, the value

for y is set to 96% for the modelling of this apparatus.The brine consumption is modelled with the same equation as well; only this

time, C represents the standard mean cooling energy consumption of 3.6 kW.

These simple models allow an accurate modelling of this continuous unit

operation. By changing only the value of one parameter, the model remains

simple (i.e., needs no implementation of further equations). This is of great helpfor the implementation of all the single unit operation models into a large, uni¬

fied model of the whole plant.

75

Modelling of Single Umt Operations

\ \. * ' V

09 07 2003 12 00

Date & Time

"6 o

n Tl

4 3 £

c M

zi tu

TJ

10 07 2003 12 00


5.6.3 Falling-Film Evaporator


Solvent Inlet

Vent P

Steam

Inlet F

Condensate

Outlet

Vent

Distillate

Outlet

-

_^-^-

Retenate

Outlet

Drain

Figure 5-27: Scheme of a falling-film24

evaporator

Measurements

A scheme of a falling-film evapo¬

rator is given in Figure 5-27. The sol¬

vent enters the equipment from the topand builds a film around the heated

pipes. In the pipes, steam is condens¬

ing. On the outside, the solvent film is

boiling and evaporating. The bottoms

are collected and discharged and the

vapour is condensed either inside the

falling-film evaporator or outside in a

separate condenser. The advantage of

the falling-film evaporator is the short

heating and residence time of the sol¬

vent and the good heat exchange be¬

cause of film boiling.This apparatus cannot be operated

batch-wise. Solvent is therefore col¬

lected in a tank (buffer tank) from the

beginning of a campaign on. If

enough solvent is collected to guaran¬

tee a large-enough feedstock, the dis¬

tillation is started.

No measurements were performed for this unit operation. The specifica¬tions of this unit operation were well known and investigations on distillation

processes were performed elsewhere (see Chapter 5.8). It was assumed that also

for this unit operation, the loss coefficient is the same as for the vessels operatedin the building (same kind of steam traps and isolation; see Chapter 5.1). The

only measured values are the constant temperature of the inlet stream and the

evaporation temperature as well as the amount of waste solvent entering the

evaporator and the amount of regenerated (evaporated) solvent.

A circulation pump that is pumping the waste solvent into the evaporator

was also active. All the required parameters are presented in Table 5-5.

24Source: adapted from http://www.tespl.com/tech_information/waste_wat/ver_eva.jpg

76


Table 5-5: Parameter values of the falling-film evaporator

Parameter Value Unit

Volume of the Steel Evaporator 0.0357 m3

Outside Surface of the Evaporator 6.14 m2

Density of Stainless Steel (Lide 1995) 7900 kg/m3Heat Capacity of Stainless Steel (Lide 1995) 0.502 kJ/(kgK)

Temperature of the boiling solvent 80 °C

Volume inflow of solvent 800 1/h

Distillate outflow of solvent 700 1/h


The base equation for the modelling of the production dependent steam con¬

sumption of the falling-film evaporator is a combination of the heating up of the

solvent to the boiling temperature (Equation (3-5)), the evaporation of the sol¬

vent without any reflux ratio (Equation (3-6)), and the losses of a heated system

according to Equation (3-7). Since heating of the vessel was only required at

the beginning of a campaign, it was neglected. This results in an equation simi¬

lar to Equation (3-16) or Equation (5-1); only this time, no reaction is going on.

For the circulation pump, Equation (3-8) was used with the value of y for

the circulation pumps found in the measurements mentioned in Chapter 5.5 (i.e.,

85%).The results of the modelling of one day of operation (i.e., about 19 m3 of

treated waste solvent) can be found in Figure 5-28. It can be seen that evapora¬

tion of the solvent holds responsible for the major energy consumption while

heating of the waste solvent and the losses have minor contributions of about

10%) to total heating steam consumption. This finding is in agreement with the

modelling of the batch distillation column, where also evaporation held respon¬

sible for the biggest part of the heating steam consumption (see Chapter 5.8).

3500 -, 1

3000 1 1

2500

2000

1500

1000

500

Heating of the Vessel Heating of the Substance Evaporation of the Substance Losses

Heating Operation

Figure 5-28: Modelled energy consumption of a one-day operation of the fal¬

ling-film evaporator (according to Equation (5-1); parameters see Table 5-5)

77

Horizontal Vacuum Rotary Dryer

5.7 Horizontal Vacuum Rotary Dryer


A picture of a horizontal vacuum rotary dryer is given in Figure 5-29 and a

detailed description of the unit operation can be found in (Mujumdar 1995).This apparatus is widely used in chemical industry because of its flexibility and

the ease of its operation. Because of its importance to chemical industry, this

unit operation was investigated and measurements were taken although it was

not available in the investigated building. Several apparatus were, nevertheless,available in the drying plant mentioned in Chapter 4 (i.e., Building 4). There,the measurements of the horizontal vacuum rotary dryer were performed.

Contrary to reaction vessels that have a heating jacket, the horizontal vac¬

uum rotary dryers investigated have heating pockets on the outside of their walls

that serve the same duty. These heating pockets cover about half of the total

surface of the apparatus, diminishing therefore the heated part of the apparatusand the heat-loss area at the outside.

Figure 5-29: Typical horizontal vacuum rotary dryer with agitator being in¬

stalled into shell (from (Mujumdar 1995))

5.7.2 Measurements

The assumptions for the measurements and the model are the same as men¬

tioned in Chapter 5.1.

Measurements were undertaken for four different solvents and products in

two different apparatus. Both of the apparatus were 4 m3 horizontal vacuum

rotary dryers with an outer surface of 17.6 and 19.5 m2 respectively. The front

and the rear door are heated as well with the same heating pockets as the other

parts of the apparatus. A typical steam measurement of several batches of the

same product in the same horizontal vacuum rotary dryer is shown in Figure5-30. It can be seen that at the beginning of the drying process, the most steam

is required and that the steam requirements fall towards the end of the drying

process to a constant value. As mentioned in Chapter 1.3, the timely depend¬ence of the steam consumption is not the primary interest of this thesis. The

main goal is to model the total consumption over a certain period (e.g., one dayor longer). Therefore, the model does not incorporate the timely dependence of

78


steam consumption, but only the total consumption per batch. This will be ex¬

plained in more detail in the next chapter.

Z 150

150 200 250

Drying Time [mm]

Figure 5-30: Steam measurements of a 4 m horizontal vacuum rotary dryer


The model of the horizontal vacuum rotary dryer consists of several parts as

the one of the nutsche dryers (Chapter 5.2) or the reaction vessels (Chapter 5.1).The first part is the heating of the substances (i.e., dry product and solvent) de¬

scribed by Equation (3-5). Furthermore, the solvent has to be evaporated. Sev¬

eral detailed models describe the behaviour of a drying process (see e.g.,

(Courtois et al. 1992; Mujumdar 1995; Parti and Palâncz 1974; Tsotsas 1992)).These models are, nevertheless, most of the time not useable for industrial

praxis because of the lack of data (e.g., the lack of the knowledge of the exact

particle shape and the partial pressure of the solvent inside the holes of the par¬

ticles). Simpler models are therefore needed. In this thesis, the basic unit op¬

eration of evaporating the solvents from the solid particles is considered as a

common boiling and evaporating operation described by Equation (3-6). More¬

over, the apparatus (consisting of stainless steel) has to be heated according to

Equation (3-5). The final part is the loss coefficient as described in Equa¬tion (3-7), which is diminished by the input of energy by the stirrer. This stirrer

input is calculated according to Equation (3-9), diminished by an efficiency fac¬

tor (rj) of 60%), accounting for the efficiency of the motor, the frequency con¬

verter and the losses in the seal as described in Table D-7 and a power factor (y)of about 30%o according to the measurements presented in Chapter 5.5. This

results in an equation for the production dependent steam consumption of a

horizontal vacuum rotary dryer according to Equation (3-16).The loss coefficient (K in Equation (3-7)) was fitted to the measurements for

each of the dryers separately. The results of the modelling are presented in

79

Horizontal Vacuum Rotary Dryer

Figure 5-31. For one dryer, K was found to be 4.5-10"2 kW/m2/K and

1.3-10"2 kW / m2 / K for the other one. The same factors as mentioned in Chap¬ter 5.1 influence the loss coefficient of the horizontal vacuum rotary dryers. Es¬

pecially the state of the steam traps and the complete heating/cooling-systemhas a big influence on the steam consumption. Furthermore, the heating of the

two stirrers is not performed in the same way: while in one dryer, the stirrer is

just filled with hot water; the other one (the one with the lower loss factor) has a

forced circulation inside the stirrer. This results in better heat transfer and more

efficient drying as well as better and more uniform drying conditions. This

shows again the range of fluctuation of the loss factor.

Measured Steam Consumption EplRDSt [kg/batch]

O Dryer 1 Dryer A A Dryer 1 Product B O Dryer 2 Product C D Dryer 2 Product D

Figure 5-31: Measured and calculated steam consumption for two 4 m3 hori¬

zontal vacuum rotary dryers (according to Equation (3-16))

With the help of the model equations, it is possible to analyse where the heat

delivered by the steam to the heating/cooling-system is going. This is shown, as

an example for Dryer 1, drying Product 2 in Figure 5-32. The long drying times

usual in drying products in a horizontal vacuum rotary dryer result in high loss

terms (proportional to time). The heating of the dryer is not a constant because

there were shutdown times during the drying periods after which the dryer had

to be heated again. Since the plant is operating five days per week, the dryersare shutdown over weekend even if the drying processes are still on their way.

The next Monday, the dryers need to be restarted and reheated to process tem¬

perature. This doubles the heat required for heating up the apparatus.

80


3000 -, r 48

Heating of Product and Solvent a Heating of Dryer m Evaporation of Solvent D Losses

xDrymgTime —Measured Steam Consumption AWetWeight oSolventMass

Figure 5-32: Modelled steam consumption for the 4 m3 Dryer 1 drying Prod¬

uct 2 (calculated according to Equation (3-16); in comparison with measured

steam consumption and drying time)

Figure 5-32 shows that by optimising the insulation of the dryers and bymodifying the steam traps (changing from swimmer-type to thermodynamic-

type), large savings could be obtained. Care must also be taken not to "over

dry" the products (i.e., keeping the already dried product in the dryer). This in¬

creases the drying time and therefore the losses while not improving the productquality.

Another possibility to save energy would be not to interrupt the drying proc¬

ess e.g., because of weekend shutdown of the plant. In such a case, it would be

preferable from an energetic point of view to wait with the start of the process

until the next week.

Improving the vacuum quality could be another possibility to save steam en¬

ergy. Care needs, nevertheless, to be taken not to save steam energy by the

price of increasing electric energy consumption of the vacuum pumps signifi¬cantly (see Chapter 5.4).

The rather good correlation shown in Figure 5-31 and the applicability of the

model on two different apparatus discussed above, show the model applicabilityto this equipment unit.

81

Batch Distillation Column

5.8 Batch Distillation Column


CoolingWater

Figure 5-33: Scheme of a batch distil¬

lation column

A batch distillation column is de¬

picted in Figure 5-33. It is, basically,a batch reactor (see also Chapter 5.1)with a distillation column on top

(i.e., only the enriching part of a recti¬

fication column). It is furthermore

equipped with a stirrer, a heatingjacket (usually without the possibilityto cool) a cooler and a possibility to

split the evaporated and condensed

solvent flow in a reflux and a distillate

flow. Unlike a continuous distillation,the solvent to be distilled is filled into

the reboiler at the beginning of the

process, heated-up, and distilled until

certain content either of the distillate

or the bottom product is reached.

Then the vessel is emptied and the

next batch may be started. A detailed

description of the concept of batch dis¬

tillation may be found in (Grassmannet al. 1998).

5.8.2 Measurements

In the investigated building, no batch distillation column is present. Since

batch distillation, because of its flexibility and efficiency, is one of the most im¬

portant unit operations in chemical industry, another industrial partner was

searched for and found, for performing the measurements within his distillation

facility.The measurements were performed with a 16 m3 batch distillation vessel

(stainless steel), heated with 15 bar steam condensing in the heating jacket (half-pipes). A distillation column with 40 perforated trays was on top of the vessel

for distillation.

Table D-17 shows the steam measurements of different batches in the distil¬

lation column for two different solvents.

82



Different modelling approaches for batch distillation are present in the

literature (see e.g., (Arellano-Garcia et al. 2002; Charalambides et al. 1995; Ga-

lindez and Friedenslund 1988; Nakaiwa et al. 2003; Oppenheimer and Sorensen

1997; Seader et al. 1997; Silva et al. 2003; Tapp et al. 2003; Venkateswarlu and

Avantika 2001)). The most extensive discussion on batch distillation and spe¬

cific models applicable can be found in (Barolo 2000; Diwekar 1996). These

models are, nevertheless, too complicated for daily use in chemical industry.

Again, simpler models are required for conducting an energy analysis.The model postulated was similar to the model for a batch reactor (see

Chapter 5.1 and Equation (5-1)), except that no reaction was going on and that

the distilled amount of solvent was multiplied by the reflux ratio plus one, since

this much solvent needed to be evaporated. The operation times were extracted

from the production record (PR) or the process step procedure (PSP). The sur¬

face of the equipment (still and column) and the other equipment specificationsare taken from the working plans of the apparatus. This leads to the following

equation:

Eisest = (ms [4 ATS]+ {1 + RR}- mES • AHf )+ mA-cAp-ATA+(K-A-ATAm-71-yPN)-t

where EpBCSt is the production dependent steam consumption of a batch

distillation column in kJ, m are the masses of the total solvent (s), the evapo¬

rated solvent (es), or the apparatus (a) in kg respectively, Cp represents the heat

capacities of the total solvent (s) or the material of the apparatus (a) in

kJ/kg/K, respectively, AT represents the temperature increases of the total

solvent (s), the apparatus (a) or the temperature difference of the apparatus to

the ambient temperature (Am) in K, respectively, AHv is the heat of vaporisationin kJ / kg, RR is the dimension free reflux ratio, K is the loss coefficient in

kW / m2 / K, A is the total outside surface area of the vessel in m2, r\ is the effi¬

ciency of the stirrer in %>, y is the relation of the actual power to the nominal

power consumption of the stirrer motor in %, P/v is the nominal power of the

stirrer motor in kW, and t is the batch time in s. EPBCSt may be translated to

kWh by dividing kJ with 3,600 s / h.

The measured and the calculated steam consumptions of the batch distilla¬

tion column are presented in Figure 5-34. It can be seen that a good correlation

between measured and calculated steam consumption exists for this simple to

use model. The model is therefore applicable to the batch distillation column.

83

Batch Distillation Column

0 2000 4'000 6'000 8'000 10'000 12000 14'000 16'000 18'000 20'000

Measured Steam Consumption EplBcst [kg/batch]

[cTsolventl DSolvent 2

Figure 5-34: Measured and calculated steam consumption for the investigatedbatch distillation column (according to Equation (5-3))

The model was fitted with the measurements taken for Solvent 1 and tested

by calculating with the same parameters for the steam consumption of the re¬

generation process of Solvent 2. It can be seen from the data provided in Figure5-34 and Table D-17, that although the process of Solvent 2 is quite different

from the one of Solvent 1, the model is still applicable for this solvent. It is

therefore assumed that the model is of general use for calculating distillation

processes.

The loss coefficient (K) for the investigated batch distillation column was

found to be about 2.5-10"2 kW / m2 / K. This is significantly lower than the loss

coefficient found for the reactors and nutsche dryers (see Chapters 5.1 and 5.2).This could be explained by the more direct heating of the vessel with steam

condensing in the jacket, the good insulation of the equipment and possibly the

well functioning steam traps of this unit operation since the equipment was justchecked during a revision period.

The deviation of about ±10% for all of the measurements shows the good

accuracy of the model, although it is an easy-to-use model without extensive

differential equations and physical data input.

Figure 5-35 shows the results of the modelling of the batch distillation col¬

umn for Solvent 1. The largest part of the steam consumption is caused by the

evaporation of the solvents, followed by the losses. This is according to the fact

that the evaporation of solvents is a rather energy-intensive unit operation. The

efficient energy usage reflected by the low loss coefficient keeps the losses

small despite the long operation times and difficult operation.

84


16 22 17

Batch No.

B Heating Up of Reaction Mass Heating Up of Apparatus m Heat for Vaporization D Heat for Losses

— Measured Steam xTotal Distilled Solvent +BatchTime

Figure 5-35: Modelling results of the batch distillation column (according to

Equation (5-3)) in comparison with measured steam consumption and distilla¬

tion time

It has to be considered, nevertheless, that about 70% of total steam (accord¬ing to the measurements) are required for the first running and the intermediate

cuts of the distillation and are therefore not directly used for the actual product.Since this is nearly unavoidable for attaining the purity of the final product, no

large savings could be expected from this side, except by changing the whole

process and equipment to a more efficient distillation column. This contradicts

the flexibility a batch distillation column offers.

85

Centrifuge

5.9 Centrifuge


A scheme of a centrifuge is pre¬

sented in Figure 5-36. A centrifuge is

fed by the slurry of a reaction or crys¬

tallization operation. This slurry is

then rotated fast by the movement of a

sieve tray. The liquid leaves the solid

by passing through the sieve tray re¬

sulting in a wet filter cake. The filter

cake has usually to be dried in a ther¬

mal dryer (see e.g., Chapter 5.2 or

5.7).A detailed overview of the differ¬

ent types of centrifuges available to¬

day may be found in (Anlauf 2003).

5.9.2 Measurements

No measurements were performed for this unit operation since no such

equipment was available in the investigated plant. Therefore, discussions with a

representative of Ferrum AG (http://www.ferrum.ch/edefault.htm)25 were con¬

ducted to come up with a suitable model for this important unit operation.


Centrifuges are operated batchwise most of the time (except push-type cen¬

trifuges that are operated continuously). A usual batch is conducted as follows:

Filling of the centrifuge with suspension

Centrifuging of the mother liquidWashing of the cake

Centrifuging of the washing liquidChipping-off of the cake

Cleaning in place (CIP)

A usual batch time is about 20 min for products easy to centrifuge and up to

about 4 h for products hard to centrifuge.First, the suspension has to be accelerated to the filling speed of about

40 m / s circumferential speed. This requires about 750 W /1. While running,the centrifuge requires about 5 to 15 kW (depending on its size). About 2 kW

are required for the hydraulic pumps. Break power of the centrifuges is about

1.8 kW /1 solids of which about 20% can be recuperated. The size of standard

centrifuges is presented in Table 5-6.

Mr. H. Reinach - discussions on August 29,2003

Mother

Liquid

Figure 5-36: Scheme of a centrifuge

86


Table 5-6: Standard sizes of centrifuges26

Diameter Length Batch Size

[mm] [mm] [kg suspension]630 320 51

800 400 100

1000 500 215

1250 600 389

The electricity consumption of centrifuges may be calculated as presented in

Equation (5-4) according to the above-mentioned estimations.

EpI.Z.EI

= PF mSu-tF + (P° + PPu) to - 0.2 P*r mSo tBr (5-4)

here, EPZEI is the total production dependent electricity consumption of a

centrifuge, P is the power required for the feed in kW /1 suspension, rrisu is the

amount of suspension in tons per batch, tp is the feed time in s, P is the power

consumption during operation in kW, r is the power consumption of the

pumps in kW, to is the operation time in s, r is the break power in kW /1 sol¬

ids, rriso is the mass of solids in tons per batch, and tßr is the breaking time in s.

With the help of this equation, the electricity consumption of a centrifuge

may be calculated. If centrifuge equipment is available, measurements should

be performed to confirm Equation (5-4).

'

Mr. H. Reinach; Ferrum AG; http://www.ferrum.ch/edefault.htm

87

Conclusions

5.10 Conclusions

The parameter values found in the above-mentioned investigations and the

corresponding modelling equations are summarised in Table 5-7. The heat of

vaporisation for both 5 and 15 bar steam (including cooling of the condensed

water to average jacket temperature) is found to be about 0.65 kWh / kg accord¬

ing to values given in (Lide 1995) and discussions with industry experts.

Table 5-7: Summary of the Equations and Parameters for the SUOM

Apparatus Utility Modelling

Equation -

see Page

Paramet(

K

[kW/m2/K]

srs &

n

[%1

Vali

7

[%1

tes

C

[kWl

Reactor

Steam

Brine

Electricity28Electricity29Electricity30

(5-1)-46

(5-1)-46

(3-13)-17

(3-8)-16

(3-8)-16

4.2-10"2

1.7-10-2

60

60

28

28

85

28

100

Nutsche Dryer

Steam

Electricity29Electricity30

(5-1)-46

(3-8)-16

(3-8)-16

4.2-10"2 60 28

85

28

-

Heat-ChamberSteam31

Electricity

(5-2) - 59

(3-8) - 16

4.2-10"2 -

64

-

Vacuum Pump Electricity (3-8)- 16 - - 52 -

APOVAC pumpsElectricityBrine

(3-8) - 16

(3-13)-17

- - 62

30

Steam Jet Pump Steam (3-13)- 17 - - - 93

Stirrer & Motor32 Electricity (3-8) - 16 - - 28 -

Infrastructure

& Losses

ElectricitySteam

Brine

(3-13)- 17

(3-13)- 17

(3-13)-17

- - -

180

200

20

Short Path DistillationBrine

Electricity

(3-13)-17

(3-8) - 16

- -

96

3.6

Falling-Film

Evaporator

Steam

Electricity

(5-1)-46

(3-8)-16

4.2-10"2 60 85

85

-

Horizontal Vacuum

Rotary Dryer

Steam

Electricity29Electricity30

(5-1)-46

(3-8)-16

(3-8)-16

4.2-10"2 60 28

85

28

-

Batch Distillation

Column

Steam

Electricity

(5-3) - 83

(3-8) - 16

2.5-10"2 60 28

28

-

Centrifuge Electricity (5-4) - 87PF=750W/t, P°=5-15

PPü=2kW;P3r=1.8k\^kW;lit

Bold values are absolute values, specific for the apparatus of the investigated buildingFor heating of the high temperature reaction vessel

Circulation pump

Other equipment

Other fixed values: CSp° = 2.5 kJ/(kg K); mB = 1.21 (see Chapter 5.3)

For all apparatus

88


With the help of these parameters and the modelling equations for singleunit operations, the modelling of a whole production plant according to Equa¬tions (3-14), (3-15) and (3-26) will be performed in the next chapter.

The investigations on single apparatus level showed, that simple models ac¬

cording to the base Equation (3-16) are applicable to model the energy con¬

sumption of these apparatus.For the generation of these models, extensive measurements had to be per¬

formed. These measurements were responsible for a big part of the work of this

thesis. Measurements were not possible for all apparatus available in the build¬

ing and extensive assumptions had to be made (see Chapter A. 1 in the Appen¬

dix). These assumptions were required to keep the models easy enough to be of

use for daily business. The models should be easy enough for being applicablewith the few data available in a standard way for most of the chemicals used in a

batch production facility.Differential equations were avoided in the models because not the timely

energy consumption but the total consumption per batch is of main interest for

production. This value is required not only for accounting the (standard) costs

of a batch but also for comparing the actual utility consumption to the calculated

utility consumption according to the production mass (see Chapter 4). If this is

not possible with a TODOMO, the BOTUMO elaborated in this and the next

chapter has to be applied. Deviations between reality and model could lead to

the investigation of batches that performed badly or equipment failure.

The models of the equipment units and the whole plant show where the en¬

ergy is consumed. With the help of this knowledge, optimisation potentials can

be revealed. Changes in energy consumption caused by changes in the produc¬tion mix will also be shown and accounted for more accurately than it is done

until now.

Optimisation potential for the investigated processes lies mainly in two

fields: the loss coefficient of the reaction vessels (steam and brine consumption)and in the nominal power of the stirring motors.

The loss coefficient of the reaction vessels influences directly the brine and

steam consumption of these unit operations. As seen in the measurements men¬

tioned in the preceding subchapters, a big part (sometimes about 50% of total

utility consumption) is lost. This loss is due to the stirrer and the circulation

pump introducing heat to the system (for brine usage only), the radiation of heat

from and to the environment, and the loss through the pipes and the steam traps

(for steam usage only). The stirrer and the circulation pump have to providemechanical energy to the system. This energy is converted to waste heat

through friction. Stirrers operated at low percentage of nominal power as the

ones usually found in chemical industry have a poor efficiency, resulting in highamounts of waste heat. Better design of the motors could therefore lead not

only to lower installation costs but also to lower operating costs. The losses

through the walls of the apparatus could be minimised by improving the insula¬

tion of the apparatus. In the investigated plant, most of the reaction vessels

were not insulated at the top because of flexibility reasons. This is definitely a

significant factor for the losses. With the help of a flexible insulation that is

easy to remove, the top could be insulated as well and the losses through the

89

Conclusions

wall would decrease. The losses through the steam traps and piping system are

significant and inherently related to the design of the reaction vessel depicted in

Figure 5-1. Any other constellation could result in cavitation within the pump.

It could be investigated, nevertheless, if installation of the steam inlet directlybefore the inlet to the vessel (i.e., after the circulation pump) would be possible.Heat transfer could be improved by this installation while not affecting the cir¬

culation pump. A drawback could be the occurrence of hotspots in the heating

jacket and a larger temperature difference from the inlet of the jacket to the out¬

let. Detailed investigations are therefore required for this possibility. Another

possibility is staying with the design, as it is today and installing different kinds

of steam traps. The main steam trap installed today is a steam trap with a float¬

ing ball. This type of steam trap is easily corrupted. Furthermore, it does not

exactly divide steam and hot water. The other type is a thermodynamic steam

trap. Here, a bimetal part opens and closes due to temperature and divides

therefore clearly between steam and liquid. The thermodynamic steam trapsknown today in industry are not too practical for this purpose because they need

to be adapted to the desired temperature manually. Changing process tempera¬

tures would therefore require manual adaptation of the setpoints of the steam

traps. Probably in the future, an electronic solution to this problem is provided.With the help of these improvements, the losses of the system could be mini¬

mised.

The model equations for the different apparatus are summarised togetherwith their parameters in Table 5-7. Some values are of general concern while

others should be investigated again in a new building. The parameters specificto the equipment of the investigated building are shaded in the table. Althoughthe development of the models required extensive measurements, the models are

built to be adaptable to different unit operations, processes and buildings. For

the modelling of a new building with new processes, only minor measurements

for verification of the models and for investigating the base consumption of the

building have to be performed. This is a big advantage when trying to providethe models company-wide while the basic unit operations and apparatus stay the

same.

As mentioned above, the models developed in this section can be used for

modelling the energy consumption of a whole plant according to the equationsprovided in Chapter 3.2. This is described and shown in the next chapter.

90

Bottom-Up Modelling of Multipurpose Batch Plants

6 Bottom-Up Modelling of Multipurpose Batch Plants

The multipurpose batch plant investigated in this chapter is, as in the preced¬ing chapter, Building 1 discussed in Chapter 4. The specifications of the build¬

ing are presented in Table 4-1. In this chapter, it will be shown that the bottom-

up approach is valid for this multipurpose batch plant for which the top-down

modelling delivered insufficient results as discussed in Chapter 4.

The steam for heating of the building is not investigated further in this chap¬

ter, since the top-down modelling according to Equation (3-3) and discussed in

Chapter 4.3.1 was applicable as presented in Figure 4-3.

6.1 Combining the Different Unit Operation Models to a Plant

Model (BOTUMO)

6.1.1 Description of the Program for Modelling Multipurpose Batch

Plants

The different unit operation models developed and postulated in Chapter 5

are based on the equations given in Chapter 3.2. As depicted in Figure 3-2, the

single unit operation models have to be combined according to production data.

These single models are then added up with the base consumption of the build¬

ing (according to Equation (3-14)). This results in a model of the whole plant(BOTUMO).

This task is performed with the help of a dedicated Excel® model (called

program in the proceeding of this thesis; see (Dahinden 2003) as well). The

program will be shortly explained for better understanding.The program is split in four layers as shown in Figure 6-1. The base data

layer consists of the specifications of the standard substances, the apparatusused and the general modelling parameters given in Table 5-7.

Base Data• Thermodynamic Data

of the Substances

• Specifications of the

Apparatus• Base Consumption of

the Building

Production Data• Recipes• Heat of Reactions

Figure 6-1: The four layers of the program for modelling the energy consump¬

tion of chemical batch plants

The production data layer contains the input sheet for the production data

(either from production record (PR) or from process step procedure (PSP)), and

the input and calculation sheets for the reactions, the heat chambers and for

other special equipment (e.g., vacuum pumps).

Calculations \• Calculations and \ R@SUltS

summations according /to the bottom-up model /

91

Combimng the Different Umt Operation Models to a Plant Model (BOTUMO)

The calculation layer contains the calculation sheets for the heating and

cooling of the substances, the evaporation of substances and all the other calcu¬

lations according to Equation (3-14) and (3-15)The results are finally summarised and presented in the results report layerAll these different layers and the interconnection of the different sheets

available on each layer are shown in Figure 6-2

The required input data for the different sheets for modelling (not includingthe input on the base data sheet since this is only required once for a specificbuilding and should not be changed by a user of the program) are presented and

explained in Chapter A 2 2 in the Appendix

Base Data

Substances

Data

ApparatusParameters

ApparatusData

Production

Data

Heats of

Reactions

Special

Equipment

Production

Data (Recipe)

Calculations

Heating/CoolingCalculations

General

Calculations

EvaporationCalculations

Results

Report

Results

Report

Figure 6-2: The different layers and the structure of the BOTUMO program and

their contents33

Hatched from bottom to top input sheets, hatched from top to bottom calculation sheets,crossed input and calculation sheets, blank results sheet, Special Equipment (e g ,

heat-

chambers) is listed separately and not together with the reactors in the sheet General Calcula¬

tions for ease of calculation

92


The different sheets and the underlying equations are presented in Chap¬ter A.2.1 in the Appendix in detail (see (Dahinden 2003) as well).

6.1.2 Modelling and Report Generation

For modelling of a production period, the input data is provided to the pro¬

duction data layer mentioned above. The input data consists of the production

steps for the different chemicals produced. Not all production steps recorded in

a PR or PSP are required as data input for the program. Only the most impor¬tant process steps such as inputs or removals of substances, the heating or cool¬

ing of the reaction media and the holding times in-between different process

steps are required. This reduces a long PR (or PSP) to a few lines in the work¬

sheets of the production data layer (see Table 6-1 for an example of a genericPSP). The times not given in the PSP (e.g., required for the input of substances,see Table 6-1) were found by discussions with production officers or by measur¬

ing the corresponding times of several batches and computing the average time.

It is also possible to compute the operation times according to theoretical data as

discussed in the outlook in Chapter 7.2. For a short production period, the pro¬

duction data of the days of interest is directly introduced to the according sheets

(see Chapter A.2.1 in the Appendix). This is, nevertheless, a big effort even for

a small number of days in a medium building, if no electronic form of the PR is

available (as in the investigated building). The results for such a modelling,

presented in the results report, are of the kind of Epm presented in Equa¬

tion (3-25). These can be added up according to Equation (3-26) to result in E

or split to give all the different E 's mentioned in Chapter 3.2.3.

93

Combining the Different Unit Operation Models to a Plant Model (BOTUMO)

Table 6-1: Example of a generic PSP and its translation for the data input to the

program34

Task Amount Time

PSP Model35 PSP Model

Check if vessel is empty -

*- -

Inerting the vessel - - - -

Input of solvent S S kg s kg - 15 min

Start stirrer (stage 1) -

*- -

Input of reactant A A kg a kg - 15 min

Input of reactant B bkg bkg - 15 min

Put stirrer to stage 2 - - - -

Heat to 70 °C - - - 45 min

React and hold at 70 °C - - 3h 3h

Take sample & send to lab - - - -

Put stirrer to stage 1 - - - -

Distil by-product C ckg -ckg - 2h

Check transfer pipe - - - -

Cool down to 30 °C - - - 45 min

Transfer mixture - -(s+a+b-c) kg - 30 min

End -

*- -

When modelling a longer period than a couple of days (e.g., a week or a

month), the modelling has to be performed with theoretical data extracted from

the PSP and not with the raw data found in the PR. Therefore, a program (i.e.,one Excel® workbook) is favourably built for each PSP performed during an

investigated period. This makes the data input easier and increases the flexibil¬

ity of the whole model. In each worksheet, only data of one batch of the spe¬

cific PSP is entered, resulting in a condensed, electronic version of the PSP.

This shows that each worksheet models a specific Epm (see Equation (3-20)). It

is possible to divide the model further to investigate Epm (Equation (3-16)). All

these single PSP models are then summarised in a single sheet where the sum¬

mation according to the different equations provided in Chapter 3.2.3 is per¬

formed. This modelling makes the program highly flexible and adaptable. A

drawback is the speed of the calculation due to the many links between the dif¬

ferent sheets that need to be updated. This could be overcome by a different

programming method as described in Chapter 7.2.

Data required for the program input are printed in bold face

A * signifies that the step is mentioned in the model without time tag

94


6.2 Results of the BOTUMO

The results of the BOTUMO and the analysis thereof will be presented in

the following chapters. Modelling was performed at different levels of detail.

The sensitivity of the BOTUMO to the input variables, the results and the con¬

clusions will be discussed in this chapter.The modelling periods mentioned in the following paragraphs were taken

during the year 2003 according to Table 6-2. For the period of one and of two

days, both the modelling according to PR and PSP data were performed. For

the longer periods (i.e., one week and one month), only the PSP data was used

as data input.

Table 6-2: Investigated periods

Period Name Starting Time End Time

One dayTwo daysOne week

One month

04.05.2003 06:00

04.05.2003 06:00

05.05.2003 06:00

06.01.2003 06:00

05.05.2003 06:00

06.05.2003 06:00

12.05.2003 06:00

10.02.2003 06:00

The BOTUMO of the whole plant incorporates not only models for each dif¬

ferent product as will be shown in Chapter 6.2.4, but also for some special op¬

eration like re-concentration and distillation of the used brine, ethanol distilla¬

tion in the falling-film evaporator, decalcification, cleaning, and preparation of

the reaction vessels. These tasks are important for the total consumption and are

modelled in the BOTUMO as depicted in the tables presented in Chapter A.3 in

the Appendix. For the discussions in Chapter 6.2.4, concerning the different

products of the building, these tasks are not considered.

The absolute values of energy consumption are modelled according to the

equations found in Chapter 3.2. The relative values found in the following

chapters are calculated by dividing the absolute consumption of each product bythe produced mass of this product. Infrastructure values are divided by the total

amount of products produced in the building. For summation, the different rela¬

tive values are summarised by weighting them with their production mass.

6.2.1 Modelling of Different Periods

The modelling of different periods with the BOTUMO is possible with two

different degrees of detail. In this chapter, the energy consumption of the dif¬

ferent energy carriers for the whole building is presented (i.e., EPm presented in

Equation (3-25)). The exact production data (extracted from the PR as men¬

tioned above) can be used as input. This is tedious and time consuming even for

short periods if no electronic version of the PR is available (see (Dahinden2003)). For most of the multipurpose batch plants known to the author, no elec¬

tronic data exist. Therefore, this approach is suitable for showing the accuracy

of the model by comparing modelled results and measured data over short peri¬ods as discussed below. Nevertheless, it is not suitable for continuous control

95

Results of the BOTUMO

and prediction of the energy consumption of a production plant, since PR data is

only available after a PSP is performed.The production data extracted from the PSP on the other hand can be used

for modelling of periods longer than a few days (e.g., one week or one month).This data is entered only once for the modelling of longer periods according to

the PSP data of each product. As shown in Equation (3-26), the modelling of

the whole plant may then be performed by multiplying the consumption of one

batch by the numbers of batches produced during the investigated period and

adding the results. A sensitivity or uncertainty analysis (see e.g., (Stahel 1995;Vose 1996)) may then be performed to investigate the influence of the uncertain

parameters to the result of the calculations (see Chapter 6.3).The possibility and accuracy of the BOTUMO of a building is based on the

accuracy of the single unit operation models discussed in Chapter 5 (see Chap¬ter 3.2 for the generic equations). In Chapter 5, the models were tested and de¬

veloped. The actual BOTUMO may only be tested on a total building level be¬

cause of the lack of measurements of smaller parts of the plant.For comparison, the consumption of the whole building was measured and

calculated by three different methods for one and two days and by two different

methods for one week and one month. For one day, the BOTUMO was pro¬

vided with the data of the PR as well as with the data extracted from the PSP.

The third method was the utility calculation method used by the company this

case study was dealing with (called CPM: Company Proprietary Method). For

periods longer than a few days, the data of the PR could not be extracted from

the paper form because of lack of manpower. Therefore, the periods of one

week and one month were modelled by the BOTUMO based on the PSP and the

CPM only. The results of the modelling are summarised in Table A-8. For one

common day of production, the results are depicted in Figure 6-3.

6 75E+03

% 5 75E+033

Ew 4 75E+03

o

"a.

E

E> 2 75E+03a>

C

LU

^ 1 75E+03a»

Q.

m

o 7 50E+02I-

2 50E+02

Steam Electricity Brine

Utilities

D Modelling according to PSP m Model According to PR a Measurement bCPM|

Figure 6-3: Modelling of the specific utility consumption (per t of product) of

the whole building for one day of production according to Equation (3-14) (incomparison with measured consumption and modelled data according to CPM)

96


It is obvious that the modelling according to the CPM for the modelling of

steam is the most inaccurate one. This model includes heating steam for the

building as well. This explains why the modelled steam consumption is that

much higher with the CPM than the measured production dependent one. Nev¬

ertheless, the deviation is bigger than the average heating steam consumption.The CPM model is based on experience of daily production. The main interest

is to give an approximate number of the (total) product cost. Since only total

utility costs are considered, and since steam is cheaper than electricity and brine,the higher value is not that disastrous for the product cost but not satisfying

anyway. The different deviations in the modelling of the different energy carri¬

ers could level out each other in terms of costs. Moreover, deviations could

level out between different months (consolidating over the year). Nevertheless,this is not satisfying since each product should be accounted for its specificcosts. The detailed BOTUMO delivers results much more accurate than the

CPM. The model based on data extracted from the PR deviates only slightlyfrom the measured value. This is a good control for the accuracy of the model.

The BOTUMO based on PSP data has a significant higher deviation from the

measured value for the one-day period. This has several reasons. First, it is

based on the standard values of the PSP. These standard times, temperatures,

masses, etc. do not correlate fully with the actual ones. This implies a deviation

inherent in the model especially for shorter periods, where different deviations

from the standard parameter values are not levelling out. The sensitivity analy¬sis described in Chapter 6.3 will give an interpretation for this fact. On the

other hand, the model based on PSP data requires the number of batches pro¬

duced (see Equation (3-26)). For one day, this number is a highly inaccurate

and uncertain fraction of a whole batch, since products usually do not have

batch times of one day. Some are started during the day and last probablylonger, while others, started earlier are finished during the day investigated. An

assumption on how many batches were produced (probably Batch (n+1) is actu¬

ally starting, while Batch (n) is operating in the middle part of its PSP and

Batch (n-1) is finishing) had to be made. The different parts of the batches un¬

der production were therefore summed up to result in a part of a standard batch

(preferably one, but also 50% or other fractions were found) resulting in 'vir¬

tual' batches. These 'virtual' batches were used for the modelling (i.e., con¬

sumption of one standard batch times the number of virtual batches according to

Equation (3-26)). This implies of course a deviation from reality that is re¬

flected by the deviation from the measured value being bigger for the model

based on PSP data than for the model based on PR data. Nevertheless, this al¬

lows keeping the same model for all the periods investigated. Since longer pe¬

riods than one day are in the focus of this thesis, a slightly larger deviation at

short periods is acceptable.The relative deviations of the different models are presented in Table 6-3.

The modelling of the brine consumption shows the largest deviation between

the BOTUMO and the measured values. This is explained by the difficulties

and high inaccuracies inherent in the single apparatus measurements described

in Chapter B in the Appendix.

97


Table 6-3: Relative deviations of the different modelling methods for the inves¬

tigated utilities according to Equation (3-14)

PeriodModellingMethod

Steam

[%1

Electricity

[%]

Brine

[%1

IDay

PSP

PR

CPM

16

-2

175

14

3

-12

-12

-18

-67

2 Days

PSP

PR

CPM

8

-3

147

5

0

-19

-19

-5

-70

1 WeekPSP

CPM

1

138

-2

-28

-27

-73

1 MonthPSP

CPM

-5

145

5

10

-16

-51

The deviations between the model and the reality decrease when modelling

larger periods. This is shown in Figure 6-4 for the modelling of one month.

The relative deviations are given in Table 6-3. As discussed above, the model

according to PR was performed for the period of one and two days only.

-3 OE+02 -1


Utilities

[Modelling according to PSP a Measurement BCPM

Figure 6-4: Modelling of the specific utility consumption (per t of product) of

the investigated building for one month of production according to Equa¬tion (3-14) (in comparison with measured consumption and modelled data ac¬

cording to CPM)

In Figure 6-4, the deviations for the CPM are high. The reasons for these

deviations are discussed above. The deviations of the PSP based model from

the measurements are smaller for the longer period. This is due to several facts

that will be discussed shortly. If the assumption that the batch time given in the

98


PSP equals the mean time of the batch operation is correct (see Chapters A.l

and C.l in the Appendix), the deviations due to time level out over long periods.For short periods, they could nevertheless account for significant deviations

(i.e., levelling is not possible since only one batch is produced). The determina¬

tion of the performed batches results in some batches not fully performed at the

end and the start of each period. This is a drawback inherent in the BOTUMO

as it is programmed here (resulting in an easy and fast data input for long peri¬

ods). The produced mass should be allocated according to the part of the total

energy consumed during the part of the operation (i.e., if x% of the total energy

of one batch are consumed in one day, x% of the total mass of the batch should

be allocated to this period). Alternatively, the exact production reports providethe information of which parts of single batches are performed during the inves¬

tigated period. Modelling of short periods is therefore preferably performedwith PR data. The longer the period, the less influence these "edge-effects"have (the smaller is the contribution to the total consumption over the period).These are the two main influences for the decreasing of the deviation from the

measurements for longer production periods.

Significant fluctuations are observed for the specific consumption for the

different periods as shown in Table A-8. This supports the finding of Chapter 4

that modelling according to simple, time dependent specific utility consump¬

tions (see Equation (3-1)) is not applicable for multiproduct batch plants with

varying production. No average standard product exists for these plants and

therefore, no general applicable specific utility consumption exists (i.e., no top-down approach is possible). This is discussed in more detail in Chapter 6.2.4.

For the modelling of brine, the deviations from the measured values are of

the same order of magnitude for all of the investigated periods. This is due to

the fact of the large uncertainty of the single unit operation models as mentioned

in Chapters 5.1.2 and B.2. It could be also a hint (since measurements are al¬

ways higher than the modelled values) that the loss coefficient is larger for a

standard batch vessel than the results of the single unit operation measurements

indicate. Another possibility could be that the assumption of a minor contribu¬

tion of safety cooling systems of some batch reactors was not correct and that

these consumptions are higher than expected. Because of the impossibility of

the measuring of these equipments, this could not be proven. The lack of a

model for the enthalpy of crystallisation to the brine consumption could be an¬

other small contribution. For reaching at a solution, detailed measurements of

crystallisations with known enthalpies of crystallisation should be performed.The measurement equipment could also be optimised to minimise the uncertain¬

ties in the temperature and flow measurements.

99


Table 6-4: Comparison of Measurements, TODOMO results according to

Chapter 4 and BOTUMO results according to Chapter 6.1 for one month of

normal production

Model TypeSteam

kWh/t

ElectricitykWh/t

Brine

kWh/t

TODOMO

PSP

CPM

ca. 2,500

2,190

5,680

ca. 625

510

531

ca. -150

-161

-94

Measurements 2,315 484 -193

Table 6-4 compares the values found in the TODOMO investigations pre¬

sented in Chapter 4 with the measurements of the one month period (see Table

6-2) and the BOTUMO according to PSP data and the CPM model. It can be

seen that the measured values lie in the same region as the mean values found in

the TODOMO. This implies that the investigated month is, as postulated, a rep¬

resentative average month. Predictions could, nevertheless, not be made

according to the TODOMO because of the high variability of the model

outcome (no linear model is possible).

6.2.2 Analysis of the Energy Consumption of the Building

The analysis of the whole building for different periods as presented in the

preceding chapter shows the applicability of the BOTUMO for energy model¬

ling of whole production plants. The model is as accurate as could be expected

considering the limitations of the measuring equipment (see Chapter B in the

Appendix) and the straightforward modelling equations used (see Chapter 3.2).The BOTUMO offers now the possibility to analyse the energy consumption

of the modelled building in detail. This is analogous to Figure 3-2 and the equa¬

tions presented in Chapter 3.2.3 only this time, the energy consumption is not

summarised to result in the energy consumption of the whole building but di¬

vided to result in the energy consumption of single parts of interest. The analy¬sis starts with the summation of the single energy carriers (i.e., steam, brine, and

electricity) according to Equation (3-25) to find Ep. With the help of this

analysis, it is possible to break down the total consumption per energy carrier to

the different apparatus groups consuming energy in the building. This enables

the analyser to put focus of energy analysis and optimisation on the apparatus

group with the highest energy consumption.The apparatus groups of the production plant requiring energy that were in¬

vestigated separately are: the reactors and nutsche dryers, the consumption of

the heat-chambers, the steam jet pumps (considered by production representa¬tives as large steam consumers), the external vacuum pumps (the APOVAC

pumps are considered directly with the nutsche dryers; the general vacuum

pumps are considered as infrastructure consumption), and the base consumption

(building infrastructure).Because of the accurate results received for the modelling according to the

PSP data and the reduced "edge-effects" encountered for longer periods, the fol¬

lowing investigations are performed for one month and sometimes for one

100


week. The modelling for one week will be presented only for reasons of com¬

parison with the results found for one month.

The modelling results are presented in Table A-9. For the relative values,the modelled consumption was divided by the total amount of products actuallyrequiring the specific utility, while the base consumption was divided by the

total amount of chemicals produced (see explanation in the introduction of

Chapter 6.2).

Figure 6-5 presents the results for the steam modelling. It is seen that the

absolute value of the steam consumption of the reactors and nutsche dryers is

the largest one. These apparatus should therefore be investigated in more detail

(see Chapter 6.2.3). Furthermore, it can be seen that the base consumption is

not as high as expected since no infrastructure equipment is using steam. The

steam jet pumps on the other hand have almost no influence on total steam con¬

sumption either. This is due to the fact that these machines are only workingwhen required and are shutdown if not in use.

1 200'000

Total Reactors & Nutsches Heat-Chamber Steam Jet

Apparatus Group

Base Consumption

§ 5 bar Steam B 15 bar Steam D Total Steam

Figure 6-5: Absolute modelled steam consumption of the building during one

month according to Equation (3-14) (PSP data)

101


In Figure 6-6, the modelled specific steam consumption for one month of

production according to PSP data is presented. The specific base consumptiondiminishes compared to the production dependent steam consumption. The heat

of reaction provides a significant, though not large contribution to the heatingsteam (and reduces the heating steam consumption therefore as shown in Figure

6-9). In terms of specific energy consumption, the reactors and nutsche dryersare the largest consumers as well. Focus has therefore to be put on these appa¬

ratus group. It will be analysed in more detail in the next chapter.

2'000

Total Reactors & Nutsches Heat-Chamber Steam Jet

Apparatus Group

5 bar Steam a 15 bar Steam Total Steam

Consumption

Figure 6-6: Specific modelled steam consumption of the building during one

month according to Equation (3-14) (PSP data)

102


The absolute electricity and brine consumption for the production period of

one month is depicted in Figure 6-7.

For electricity, base consumption is significant and is responsible for about

50% of total consumption. This is explained by the intense use of electricity bysome infrastructure equipment like the ventilation system that is running con¬

tinuously. The vacuum pumps responsible for specific products also use a sig¬nificant part of electricity. The high nominal power of these motors and the

long activity times result in high consumption of this apparatus group.

The brine consumption is mostly dependent on the reactors and nutsche dry¬ers as well. Here, heat of reaction is only a minor part of the total absolute con¬

sumption (see Figure 6-11). Since no crystallisation enthalpy was included in

the model, this is not surprising (i.e., only few reactions occur at low tempera¬ture as may be seen in Figure 6-14). The base consumption, nevertheless, is

significant. This means that the circulation of the brine inside the building has a

significant impact on total consumption and optimisation of the circulation sys¬

tem would result in significant savings.

Total Reactors & Nutsches Heat-Chamber Vacuum Pumps

Apparatus Group

Base Consumption

§ Electricity D Brine |

Figure 6-7: Absolute modelled electricity and brine consumption of the build¬

ing during one month according to Equation (3-14) (PSP data)

103


In Figure 6-8, the specific modelled electricity and brine consumption of one

month of production for the different apparatus groups is presented. It is recog¬

nised again that the importance of the specific base consumption decreases

compared to the absolute one because it is specific to the total amount of chemi¬

cals produced while the energy consumptions of the other apparatus groups are

specific to the produced amount of chemicals that actually use these energiesand apparatus (see above). Nevertheless, base consumption of electricity is still

significant. The vacuum pumps stay significant but not as important than the

base consumption of the building.For brine, the specific energy consumption of the vacuum pumps may be

neglected. Only the apparatus group reactors and nutsche dryers are of impor¬tance when modelling the brine consumption.

c. 200 -

Total Reactors & Nutsches Heat-Chamber Vacuum Pumps Base Consumption

Apparatus Group

i Electricity D Brine

Figure 6-8: Specific modelled electricity and brine consumption of the building

during one month according to Equation (3-14) (PSP data)

6.2.3 Modelling of Different Aspects of the Reactors and Nutsche

Dryers

As in the preceding chapter, the model calculations presented in this chapterare all based on PSP data, since long periods are modelled. The results pre¬

sented here are according to Ep depicted in Equation (3-25). Not the whole

Epm is considered, but only the biggest consumer of energy, namely the appara¬

tus group Reactors and Nutsche Dryers is investigated. Different aspects of the

energy consumption of this apparatus group are investigated in the followingparagraphs.

From the analyses in the preceding chapter it is seen, that the reactors and

nutsche dryers are the most important energy consumers of the building (besidesthe building infrastructure for electricity consumption). Base or infrastructure

104


consumption is extensively discussed in Chapter 4 and 5.6.1. No focus will be

put on optimising or modelling the electricity consumption of the infrastructure

of a production building in more detail in this chapter. For this continuous op¬

eration, models exist and industry has a significant expertise in optimising the

infrastructure consumption of buildings (see e.g., (SIA 1992; SIA 1995; SIA

1997) or (Gränicher 1997) or (Severson 1996; Sulzer 2003; Thumann 1983;Turner 1982)).

Furthermore, the continuous distillation will not be discussed in this para¬

graph. It is a significant consumer of steam in the investigated building but is

not a batch apparatus this thesis is dealing with. It is discussed separately in

Chapter 5.6.3.

The BOTUMO (based on PSP data) was used for a detailed analysis of the

energy consumption of the batch reactors and nutsche dryers. The data outputof the model is presented in Table A-10 to Table A-21. The different outcomes

will be discussed with the help of the figures provided in this chapter.

Figure 6-9 presents the modelled specific steam consumption of the reactors

and nutsche dryers. The different unit operations requiring steam, the stirrer and

the heat of reaction input and the losses are investigated. The modelling is per¬

formed for one week and one month, both based on PSP data as described

above. It can be seen that the different production mixes during the two periodsresult in different modelling results for the specific steam consumption. Differ¬

ent products require different amounts of steam during their production process.

This results in differences in the (overall) specific steam consumption for this

apparatus group. The model accounts for the differences in the production

processes (see Chapter 6.2.4). The different "edge effects" discussed above

may have an influence on the product specific energy consumption as well.

This will be discussed in detail in Chapter 6.2.4 as well.

105


Reflux Evaporation Heating of Heating of Heating of Losses Stirrer Input Heat of

Substances Apparatus H/C-System Reaction

Aspect

11 Week D1 Month

Figure 6-9: Modelled specific steam consumption of the reactors and nutsche

dryers according to Equation (3-25) (PSP data)

Reflux conditions are used more often during the investigated month than

during the investigated week.

The evaporation of solvents used a significantly higher amount of specificsteam during the investigated month than during the investigated week.

The higher specific losses for one month depict that longer heating periodsand higher process temperatures have an influence on the loss coefficient, the

heating of the heating/cooling-system, and on the heating of the apparatus. The

specific consumption of the heating of the apparatus and the heating/cooling-

system could both be improved by moving products from smaller apparatus to

bigger ones. This improves the relation between outside surface of an apparatusand its content. Since the weight of an apparatus (metal) is related to its surface

area (surface area times thickness of the metal times the density of the metal

equals the weight of a reactor), the specific energy consumption for heating the

apparatus is decreased by increasing the size of an apparatus. The heating of the

heating/cooling-system uses less specific steam during the month than duringthe week. This is a hint that different apparatus were operated during the two

periods (see Chapter 6.2.5 as well). An explanation could be that the apparatusused during the week were mostly reactors with thinner walls but with the same

water content of the heating/cooling-systems as the reactors with the thicker

walls (i.e., specifically larger water content). This would increase the specificsteam consumption for the heating/cooling-system and reduce the specific con¬

sumption for the heating of the apparatus (i.e., bigger influence on the heating of

the heating/cooling-system than on the heating of the apparatus).Stirrer input (i.e., friction heat introduced to the system by stirrers) may be

neglected. It is only about 2% of total and of specific steam consumption.Losses, nevertheless, are significant and responsible for about 50% of total

steam consumption for the reactors and nutsche dryers. About 50% of these

106


losses are, as mentioned in Chapters 5.1 and 5.2, caused by the losses throughthe steam traps. Heat transfer through the outside wall of the apparatus is re¬

sponsible for the other 50%. It is obvious that minimisation of the losses pro¬

vides the best possibility to optimise the steam consumption of the reactors and

nutsche dryers. An improvement of about 10% of the losses would result in a

reduction of about 5% of total steam consumption for these apparatus while an

optimisation of 10% for an improvement of the heating of the substances (i.e.,lower process temperatures or solvents with lower heat capacity) would onlyresult in about 1% of total steam consumption.

The heat input by reaction (i.e., heat of reaction) lies in the same order of

magnitude as the heat required for the heating of the apparatus material for the

reactions occurring in the investigated building. It may not be neglected but is

only a minor contributor to steam savings.Focus may be put on the steam consumption for reflux conditions. It is

questionable whether these conditions are always required for the production

process (e.g., drying of the solvent) or whether it is only a simple method for

keeping process temperature constant. With today's possibilities of controllingthe process temperature of an apparatus, reflux just for keeping a temperaturelevel would be useless. Detailed investigations of the different PSP could reveal

the actual use of the reflux conditions and lead to an optimisation of the steam

consumption of this specific unit operation.For the electricity consumption of the reactors and nutsche dryers, Figure

6-10 gives the modelling results for one week and one month. The different

equipments investigated are stirrers, vacuum pumps, circulation pumps, diverse

small motors of the vessels, the electric heating, and the APOVAC pumps.

Vacuum Pump Circulation Pump Div Motors Electric Heating APOVAC

Aspect

1 Week D1 Month

Figure 6-10: Modelled specific electricity consumption for the reactors and

nutsche dryers according to Equation (3-25) (PSP data)

107


During the investigated month, no electric heating occurred while during the

investigated week the reactor heated with electricity was operated. It is seen

that the electric heated high-temperature reactor is requiring a large specificamount of electricity if operating. Since electricity is a high-valued energy (see

e.g., (Spreng 1988)), it is disputable whether it should be used for heating pur¬

poses. Moreover, electricity is generated in the site by expansion of high-

pressure steam to low pressure steam. This high-pressure steam could be partlyused for directly heating the high-temperature reactor therefore eliminating the

conversion losses from steam to electricity and back to heat. This would in¬

crease total steam consumption but probably decrease total utility costs of the

investigated building.The differences between the modelled results of specific energy use of the

one month period and the one week period are mostly due to the different prod¬uct mix for the two periods (see Chapters 6.2.4 and 6.2.5 as well). The applica¬bility of the models for the changing production mix of multipurpose batch

plants and the high variability between the products is therefore shown.

The stirrer motors are the largest electricity consumers apart from the elec¬

tric heating. The motors possess high nominal power. Although the relation to

actual used power is low, their total consumption is significant and the largest of

all investigated equipment (motor) groups. The stirrer motors consume more

than 30%) of the electricity consumption of the reactors and nutsche dryers. In¬

stalling smaller motors (i.e., improving the efficiency of the motors) could de¬

crease this consumption since the motors have low efficiency at low power us¬

ages (see Table D-7). Because of the reasons discussed in Chapter 5.5, this may

not be practical for all the apparatus and motors but has to be considered if in¬

stalling new apparatus.

Only small (vacuum) pumps and other motors are sometimes installed for

specific apparatus - here, they are summarised as diverse motors. As seen from

the modelled data, these motors may be neglected without significantly affect¬

ing the modelling accuracy. Whether or not these motors are required for an

apparatus has to be considered during installation and will not be discussed

here.

Surprising is the large consumption of the circulation pump (about 25% of

total consumption for one month as shown in Table A-19). Circulation pumps

are mostly small pumps with nominal power of about 2 kW as presented in

Table D-7. They are running with high efficiency at almost nominal power.

Since they are operating as long as the reactor or nutsche dryer is operating (i.e.,the heating/cooling-system is active either with steam, water, or brine), they are

in use almost continuously. According to Equation (3-8), the power consump¬

tion is proportional to the operating time. This causes the high consumption. If

for smaller heating/cooling-systems, smaller circulation pumps would be usable,this could improve the power consumption of this apparatus group and contrib¬

ute to significant energy savings.The APOVAC pumps are not in use continuously but only during thermal

drying periods in the nutsche dryers. This limited time decreases the power

consumption although the high nominal power of this apparatus type.

108


The introduction of energy efficient motors to the plant (see e.g., (Greiner1999; Kordik 2001)) could decrease electricity consumption significantly. Pay¬back times, installation and running costs have to be considered carefully but

the possibility exists to save significant amounts of energy as discussed in sev¬

eral case studies mentioned for example in (de Almeida 1997; Andreas 1992;Bundesamt für Konjunkturfragen 1992; Francis Murray 1994; Lindegger 2002).

In Figure 6-11, the modelled specific brine consumption is depicted for the

different aspects of a production process requiring brine. The analysis is con¬

ducted for the cooling of the substances and the apparatus36, the losses, the stir¬

rer input, the brine used for the APOVAC pumps, and for the heat of reaction.

Again, the difference in production mix is accounted for in the model.

Therefore, the specific brine consumption of each aspect is different for the two

periods investigated.The largest single users, specific as well as absolute, are the APOVAC

pumps (about 33% of total brine consumption for one month). As mentioned in

Chapter 5.1, only few reaction vessels are using brine. Since the APOVAC

pumps are required for the thermal vacuum drying of the products in the nutsche

dryers, these four equipment units are operated frequently. The large specificconsumption of brine discussed in Chapter 5.4.2 results in the high consumptionof these unit operations. Whether or not the cooling to low temperatures is re¬

quired for these unit operations is therefore of major interest when trying to re¬

duce the brine consumption.

Cooling of Cooling of Losses Stirrer Input APOVAC Heat of Reaction

Substances Apparatus

Aspect

1 Week D1 Month]

Figure 6-11: Modelled specific brine consumption for the reactors and nutsche

dryers according to Equation (3-25) (PSP data)

Stirrer input is only of minor importance. During crystallisation processes,

stirring is sometimes halted or the stirrer is at least operated at low power level

'

Brine may only be used for cooling below 30 °C.

109


for preventing destruction of the crystals. Since stirrer input is proportional to

time as shown in Equation (3-9), the low stirrer input may be explained.As seen from the cooling of the substances and the apparatus, more and

lower cooling was performed for the products produced during the month than

during the week of modelling. This results in significantly higher specific brine

consumption. Changing from smaller to larger vessels (both filled completely)would reduce the specific usage of brine for cooling down the apparatus as men¬

tioned above since the relation between vessel and content decreases with in¬

creasing vessel size.

The losses are also significant for the total brine consumption of the reactors

and nutsche dryers (responsible for about 20% of total brine consumption of

these apparatus group). Lower cooling temperatures and longer cooling times

for the different products produced result in the higher specific brine consump¬

tion of the results of one month compared with one week. Losses for brine con¬

sumption are smaller than for steam consumption as discussed in Chapter 5.1.3

and only caused by heat input through the wall of the apparatus from the envi¬

ronment. Nevertheless, they are significant and should be minimised by better

insulation of the apparatus.The heat of reaction is of minor importance for the production processes oc¬

curring in the investigated building. Only one process required brine cooling in

the reaction phase. If more processes would require reaction at lower tempera¬

tures, the contribution of the heat of reaction to total brine consumption would

increase significantly. Therefore, this has to be considered and special care

needs to be taken for investigating whether reaction at low temperatures is re¬

quired or not.

6.2.4 The Differences between the Products

In the investigations of the TODOMO presented in Chapter 4, it was as¬

sumed that all the products of a multipurpose batch plant use about the same

amount of specific energy. With the help of the BOTUMO, this assumptionwill be investigated in this chapter. The modelled values are shown in Chap¬ter A.3.1 in the Appendix. Especially the percentages of the different energy

consumptions for each product are presented in Table A-22. For some products,the modelling was performed during the "one week" (W) period and for the

"one month" period (M).The findings presented in this chapter are according to Equation (3-20)

(i.e., Epm) presented in Chapter 3.2.3. The shading of the energy consumption

bars of the different products gives an even more detailed analysis by presentingdifferent aspects of the production processes (similar to the investigations pre¬

sented in Chapter 6.2.3).

110


Figure 6-12 presents the specific modelled steam consumption for the dif¬

ferent products and the number of synthesis steps involved in the production of

the different chemicals.

7500

7000

6500

tf,

A(M) B(W) C(W) D(M) E(M) F (W) G (M) G (W) H (M) I (M) J (M) J (W) K (M) N (W) O (W)

D Reflux

H Losses

1 Evaporation

1 Stirrer Input

E] Heating of Substances n Heating of Apparatus a Heating of H/C-System

Heat of Reaction o Synthesis Steps

Figure 6-12: Specific modelled steam consumption of the different products(A, B,..,N, O) according to Equation (3-20) and number of synthesis steps

(PSP data; modelling period: W = one week, M = one month)

The modelling of one month and one week for two products (G and J)shows the accuracy of the model for both periods. The aforementioned "edgeeffects" (see Chapter 6.2.1) influencing the specific energy consumption for the

two periods differently may be neglected. The model, nevertheless, accounts

for the differences in the production recipes of the different products. There¬

fore, the differences of total specific consumption for the two periods found in

the former chapters are caused by differences in production mix and not by in¬

accurate accounting of the produced amount of chemicals.

From Figure 6-12 it may be seen that the assumption of similar specific en¬

ergy consumption for all of the different products postulated in Chapter 4 is not

true for the specific steam consumption of the investigated building. The prod¬ucts vary widely in absolute steam consumption as well as in the specific con¬

sumption for the different unit operations and aspects involved in the production

process. The statement made in Chapter 4 that only production of one productor constant production mix of different products allows the top-down modellingof the whole production building is confirmed. The investigated building

(Building 1 in Chapter 4) shows no constant production mix and the productsare varying widely in specific steam consumption.

The different aspects of the production processes are discussed extensivelyin Chapter 6.2.3. It is seen, nevertheless, from Figure 6-12 that the losses are

the most significant specific steam consumers for all the products. The losses

ill


are varying with the number of different vessels used (i.e., the total surface

area), the batch times and the temperatures of the specific process.

Products with extensive reflux operation (i.e., long batch times), processes

with high temperatures and production in several different vessels are found to

be the major consumers of steam. Optimisation should therefore start with these

products and with those apparatus with the highest specific energy consump¬

tions (see Chapter 6.2.5).

Comparing the number of synthesis steps given in Figure 6-12 with the

modelled steam consumption shows a limited relationship. Three different lev¬

els of energy consumption (i.e., low, medium, high) and three different levels of

synthesis steps (i.e., few, some, many) may be differentiated. The number of

synthesis steps could therefore give a first indication of the order of magnitudeof the energy consumption of a product. Nevertheless, deviations from this rule

may be found as may be seen in Figure 6-12. Product I possesses only one syn¬

thesis step, but high specific steam consumption. This may be explained by the

fact that this product requires long distillation, reflux, and holding periods at

high temperatures. Losses and total consumption are therefore high. Another

exception is Product O. This PSP consists of no chemical synthesis steps but

only of physical transformations. These transformations require steam as well.

Therefore, the steam consumption is not zero. Nevertheless, the number of syn¬

thesis steps could provide a first and easy crosscheck for the modelled energy

consumption.

112


The specific modelled electricity consumption of the different products is

presented in Figure 6-13. First, the electricity consumption of Product B should

be highlighted. This is the chemical produced in the high-temperature reactor.

As mentioned above, this reactor uses electricity for heating. This is repre¬

sented by the high specific electricity consumption of this vessel. Without the

electric heating (and with a circulation pump similar to the ones of the other re¬

actors), the electricity consumption would be in the same order of magnitude as

for the other products.The products with the second largest electricity consumptions (i.e., Prod¬

ucts G, I and F) show the largest consumption for the stirrer motors. These are

products dried in the nutsche dryers. These apparatus have a significantly

higher nominal stirrer power and require the APOVAC pumps while the content

is dried thermally. This results in the high electricity consumptions. Moreover,these chemicals are produced within several reactors and have long batch times.

This results, according to Equation (3-8), in high electricity consumption as

well.

The products with only minor contributions to electricity consumption and

small specific electricity consumption (i.e., Products A, C, H, J, K, N, and O)are chemicals produced only in one or two reactors (see Chapter 6.2.5). The

batch times are small and the total electricity consumption, as well as the spe¬

cific one is limited therefore.

S 2'000 -

1u~

ë re

500

_H_

El

35

-=3i-

A(M) B(W) C(W) D(M) E(M) F (W) G (M) G (W) H (M) I (M) J (M) J (W) K (M) N (W) O (W)

Products (modelling period)

I Stirrer Vacuum Pump Circulation Pump mDiv Motors E] Electric Heating EBAPOVAC o Synthesis Steps |

Figure 6-13: Specific modelled electricity consumption of the different prod¬ucts (A, B,..,N, O) according to Equation (3-20) and number of synthesis steps


As found for the specific steam consumption, the specific electricity con¬

sumption shows also no mean value to be used for a TODOMO as presented in

Chapter 4. The products of a multiproduct batch plant with varying productionrecipes are as diverse that no mean specific electricity consumption is able to

113


model the varying production mixes. With the help of the BOTUMO, this is,

nevertheless, possible. The BOTUMO leads to the good description of the elec¬

tricity consumption of the complete building as mentioned in Chapter 6.2.1.

The minor influence of the "edge effects" and the reproducible modelling of

the specific energy consumption are represented by Products G and J, producedin both periods as discussed above.

As before, the synthesis steps are presented in Figure 6-13 as well. Again, a

relationship between synthesis steps and electricity consumption is found.

About two levels could be found: one synthesis step resulting in low consump¬

tion and several synthesis steps resulting in higher consumption. For Product O

the same explanation as above is valid: the physical transformations require

electricity, but are of course not represented in the number of synthesis steps.Product B requires the high-temperature reactor heated by electricity. There¬

fore, electricity consumption is much higher than for the other products. Ne¬

glecting the electricity consumption for electric heating would result in a similar

consumption according to the other products requiring a similar amount of syn¬

thesis steps. An explanation of the high electricity consumption of Product I

(compared to the other products with only one synthesis step) is, that this prod¬uct is dried for a long time in the nutsche dryers, resulting in high electricity

consumption for the stirrer and the APOVAC pumps.

Figure 6-14 presents the specific modelled brine consumption of the differ¬

ent products. Only products actually requiring brine during their production

process are presented in the figure. It can be concluded instantaneously, that no

mean specific brine consumption exists for the production processes of the in¬

vestigated building. This explains that the modelling of the investigated build¬

ing (i.e., Building 1 in Table 4-1) with the help of the TODOMO had to fail as

discussed in Chapter 4 for brine consumption as well.

D (M) E (M) F (W)

Products (modelling period)

Id Cooling of Substances aCooling of Apparatus »Losses Stirrer Input a APOVAC aHeat of Reaction o Synthesis Steps I

Figure 6-14: Specific modelled brine consumption of the different products

(A, B,..,G, I) according to Equation (3-20) and number of synthesis steps


114


The modelling of Product G for one month and one week showed once

again the minor influence of the "edge effects" (see above). The differences in

specific brine consumption mentioned in the preceding chapters are therefore

results of the modelling of the different production procedures of the different

products and not side effects of the modelling method. This product is the onlyone that requires brine during the reaction step, resulting in brine consumptionfor the heat of reaction. The brine consumption for the heat of reaction for

Product G is significant (i.e., about 17% of the total brine consumption of this

product). Since it is the only product requiring brine for removing the heat of

reaction, the influence of the heat of reaction on total brine consumption for all

products is only minor (see Figure 6-11). Nevertheless, low-temperature reac¬

tions should be investigated carefully since they are an important contributor to

specific brine consumption.The products with long cooling times (i.e., high losses according to Equa¬

tion (3-7)) or that are treated in the nutsche dryers (i.e., requiring the APOVAC

pumps) show the highest brine consumptions (i.e., Products G and I). The

APOVAC use increases the brine consumption of all the products using this ap¬

paratus significantly. Similarly to the findings for the other utilities, Figure6-14 shows for which products the processes should be investigated in more de¬

tail. For products like Product A, C, or E, the total saving potentials possible byoptimisation are small. The processes of the major consumers (i.e., Products G

and I) should be investigated intensively to optimise the total brine consumptionof the building.

Moreover, the numbers of synthesis steps for the different products usingbrine in their production processes are presented in Figure 6-14. Small and

large energy consumption is connected with few and medium number of synthe¬sis steps. The only exception from this rule is Product I. As mentioned above,this product requires long drying in the nutsche dryers. This results in largebrine consumption of the APOVAC pumps. As before, the number of synthesis

steps gives a first impression of the specific brine consumption.The investigations presented above show, that the BOTUMO is helpful in

analysing the specific energy consumption of the different products. The mod¬

els reflect the specifications of each product and the results show the user where

energy is used and where optimisation should start. Furthermore, the investiga¬tions show how this optimisation could be possible from a general point of view

(e.g., by better insulation or by changing the heating process).Investigations on a relationship between unit operations performed during a

production process or total production time and energy consumption were con¬

ducted similar to the investigations on synthesis steps. Nevertheless, no rela¬

tionship was found although the effort in gathering the respective data is higherthan for the synthesis steps. Therefore, investigations like this would be of no

use for crosschecking the modelled energy consumption and are therefore not

further discussed in this thesis. The number of synthesis steps gives a first im¬

pression on the order of magnitude of the energy consumption of the different

products. Because of the unique kind of each type of reaction, results are, nev¬

ertheless, not easily transferable to other production buildings. This relationshipshould only be used to compare production processes of the same building.

115


Programming errors could be revealed. Values contradicting the relationshipcould occur, as mentioned above, but need to be explained according to the PSP.

6.2.5 The Differences between the Apparatus

The modelling results of the different apparatus available in the building are

presented in Table A-23, Table A-24, and Table A-25 and will be discussed

briefly in this chapter.The modelling described in this chapter was performed, according to Epm as

depicted in Equation (3-16) (i.e., the consumption of each specific apparatus,

consuming one kind of energy, and producing one specific chemical).Figure 6-15 shows the specific steam consumption for each apparatus avail¬

able in the investigated building during the one month modelling period accord¬

ing to PSP data. It is seen that, although all the apparatus have the same loss co¬

efficient, the specific steam consumptions per ton of produced chemical are

quite different. This shows that the model accurately accounts for the different

specifications of the apparatus and the different process conditions. The higherthe process temperature, the longer the batch time, and the more solvent is

evaporated (i.e., distilled or hold at reflux conditions), the more steam is con¬

sumed according to Equation (5-1).

3'000 -, 1

E1 2'500

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Apparatus No.

| D Product A h Product D ra Product G ta Product H Product I H Product J Product K s Product E

Figure 6-15: Modelled specific steam consumption of the apparatus

(1, 2,..,26, 27) during one month according to Equation (3-16) (PSP data)

Furthermore it is seen, that some products (e.g., Product G) use several ap¬

paratus, while some products (e.g., Product K) use only one or two apparatus.This supports the findings of the last paragraph that the number of apparatusused by a production process has a significant influence on its energy consump¬

tion. Nevertheless, this is not the only or major influence since the large energy

consumption of Product I is only distributed over four apparatus (see Figure6-15).

116


Some production processes are performed in the same apparatus. Differing

energy consumption may be due to differences in process times, temperatures,

filling of the vessel, production processes, or physical properties of the chemi¬

cals. Investigations of performing the same production process in different ap¬

paratus were not performed but could be done with detailed production data ac¬

cording to the PSP.

Figure 6-16 depicts the modelled specific electricity consumption of the ap¬

paratus. The modelled period is one month and the data basis is provided byPSP data. The process requiring electric heating is not running during this pe¬

riod and therefore, the high temperature reactor consumes no electricity. The

influence of the different nominal powers of the motors may be seen by the high

consumptions of Apparatus 24 and 26 (i.e., two nutsche dryers). The motors of

these nutsche dryers have large nominal power consumptions and are therefore

major users of electricity (the process in nutsche dryer No. 27 is a very short one

and therefore, the high nominal power has only minor influence). The con¬

sumptions for the other vessels are in the same order of magnitude (dependingon the degree of filling of the apparatus since specific electricity consumption is

depicted).A higher degree of filling (if possible) could improve the specific energy

consumption for all the vessels. Another possibility for optimisation could be to

install smaller motors with higher efficiencies. Nevertheless, this results in the

drawbacks discussed in Chapter 5.5.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Apparatus No

D Product A a Product D a Product G a Product H Product I H Product J Product K s Product E

Figure 6-16: Modelled specific electricity consumption of the apparatus


Sensitivity Analysis of the BOTUMO

Figure 6-17 shows the modelling results for the specific brine consumption.

Investigated period is again one month. It may be seen, that only a few appara¬

tus actually require brine for cooling (the nutsche dryers No. 24 to 27 requirebrine for the APOVAC pumps). The consumptions differ because of the same

reasons as mentioned for the steam consumption (different production processes

etc.). No mean specific brine consumption can be found for the apparatus of the

investigated building.

-500 t

450 :-

$ -400 ;-

| -350 ;-

|300 ;

| -250 :-

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Apparatus No.

| D Product A H Product D B Product G Product H Product I H Product J m Product K B Product E |

Figure 6-17: Modelled specific brine consumption of the apparatus


6.3 Sensitivity Analysis of the BOTUMO

The sensitivity analysis was performed for the parameters given in Table

5-7. Because of the better applicability of the model for longer periods as dis¬

cussed in Chapter 6.2, the analysis was done for the period of one month. Each

parameter was investigated by setting the parameter value to 50%, 80%, 120%,and 150%) of his original value given in Table 5-7. The results are provided in

Table A-26, summarised in Table 6-5 and will be discussed for each parameter

separately in the following subchapters.The influence of the temperature of the reaction mixture and the mass of

chemicals provided to the batch were not considered in this sensitivity analysis.These parameters, given in the PSP, are critical for producing the right qualityof products. Discussions with industry experts showed, that the values given in

the PSP are strictly followed during daily production. Therefore, the values do

not fluctuate largely (verified by considering several PR) and no sensitivity

analysis was performed.The influence of the parameters on total utility consumption is investigated.

It is obvious that the changes in the parameters influence the consumption of the

apparatus group related to the parameters directly through the equations given in

118


Chapters 3.2 and 5. The influence on the total consumption is not easy to inves¬

tigate and depends on the context. Therefore, a sensitivity analysis was per¬

formed for the total consumption modelled with the help of the BOTUMO.

The influence of the few small motors of the heat-chambers was investi¬

gated as well as presented in Table A-26 and Table 6-5. The influence was, as

expected, very small according to the small nominal power of these few motors.

Therefore, this sensitivity analysis is not presented nor discussed in this section.

6.3.1 Time

The influence of changes in batch time on the specific consumptions of the

different utilities is presented in Figure 6-18. The batch times were changed by

changing the times of the different unit operations all similarly as discussed in

Chapter C.l in the Appendix. The times given in the PSP were considered as

the mean time for a specific unit operation.The good correlation between the mean values (i.e., base case) and the

measurements shown in Figure 6-18 support the assumption that the values

given in the PSP are equivalent to the mean values of the actual processes for

the investigated building. This, nevertheless, has to be tested for a new buildingwhere the BOTUMO should be applied without performing any direct meas¬

urements of single apparatus.

A change in the different times of all the processes of ±50% results in a

change of the different utility consumptions of about ±20%. Time is therefore

the most important parameter that influences all utility consumptions. This may

be explained by considering that in all of the equations given in Chapter 3.2,time is present as a variable and because losses are found to be significant.

Optimisations of process times are therefore not only influencing the capac¬

ity of a batch plant but also the energy consumption positively (assuming that

the temperature levels and motor powers required do not change).

LJ 1000

•& 500

Deviation of the

Parameter

Value from the

Base Case

a 50%

20%

Base Case

+20%

H +50%

— Measurement

Utility

Figure 6-18: Sensitivity analysis of the batch time t with regard to the specificutilities according to Equation (3-14) (one month; PSP data)

119


6.3.2 Steam Heat Transfer

Figure 6-19 gives the sensitivity analysis results of the dependence of total

steam consumption (5 and 15 bar) from the steam loss coefficient K accordingto Equation (3-7). This loss coefficient is required for the modelling of all appa¬

ratus utilising steam (i.e., reactors, nutsche dryers, distillation columns, and

heat-chambers). The influence on the outcome is more significant than for the

brine loss coefficient. A 50% increase or decrease in the loss coefficient results

in a 20%) (13%>) change for the 15 bar (5 bar) steam consumption modelled with

the BOTUMO. This shows the high influence of the loss coefficient to the

model as seen before (see Chapter 6.2.3 and (Dahinden 2003)). The losses are

high for the steam consumption and should be minimized. Because of the

higher temperature level and the more intense use of 15 bar steam (i.e., heavier

heating duty), the loss coefficient has a bigger influence on the 15 bar than on

the 5 bar steam consumption.The loss coefficient used for the modelling (see Table 5-7) gives more accu¬

rate results for the 15 bar than for the 5 bar steam consumption. Since the other

parameters and modelling equations are the same, this could be a hint that

shows that the loss coefficient for the 5 bar steam is higher than for the 15 bar

steam (e.g., steam traps are more efficient at this pressure level or the insulation

is thicker for 15 bar steam vessels). Such a result was not found from the meas¬

urements performed in this thesis but could be verified by making more intense

measurements in future research (see Chapter 7).

1 500q 1

15 bar 5 bar

Pressure Level

Figure 6-19: Sensitivity analysis of the steam loss coefficient Kst with regard to

the specific steam consumption according to Equation (3-14)

(one month; PSP data)

120


6.3.3 Brine Heat Transfer

In Figure 6-20, the sensitivity analysis for the loss coefficient of the brine-

cooling regime is presented. It can be seen that even for the high value of 150%

of the value of the loss coefficient found by the measurements presented in

Chapters 5.1 and 5.2, the modelled brine consumption lies significantly below

the measured value. Second, the figure shows the large influence on the

BOTUMO, the brine loss coefficient has. Nevertheless, the model is linear in

the loss coefficient. The higher the loss coefficient is, the higher is the brine

consumption.A 50%) decrease or increase in the loss coefficient results in a 7% increase or

decrease of total modelled specific brine consumption. The base consumptionand the consumption related to the chemicals are therefore significant as dis¬

cussed above.

Since the model results in smaller brine consumptions than measured, even

for the highest loss coefficient, the model for brine consumption has some defi¬

cits. First, the high inaccuracies mentioned in Chapter B.2 in the Appendix mayresult in major deviations of the model and the reality. For investing this devia¬

tion, detailed measurements with more accurate measuring equipment should be

performed. Second, the security heat exchangers of the reaction vessels (to pre¬

vent solvents from leaving the system through the vent) are operated with brine.

These apparatus could not be measured because of the physical installation (notlong enough straight pipe for the ultrasonic flow meter). The consumptionshould not be high according to industry experts, but even small consumptioncould result in big overall consumptions over time. Last, and probably the least

important, the assumption of neglecting the enthalpy of crystallisation in the

model may be doubted. Further investigations should be performed for reachingat an easy-to-use model that incorporates the (mostly unknown) enthalpy of

crystallisation.

Figure 6-20: Sensitivity analysis of the brine loss coefficient Kq0 with regard to

brine consumption according to Equation (3-14) (one month; PSP data)

121


6.3.4 Condensation Enthalpy of Steam

Figure 6-21 presents the results of the sensitivity analysis for the condensa¬

tion enthalpy of the steam Although this value is changing with temperature, a

constant value of about 0 65 kWh / kg steam was assumed for the model This

value accounts for heat of condensation as well as for a mean energy content of

the condensed water (assumed for a mean temperature difference between jacketof the reactor and condensation temperature) The same value was taken for

5 bar and for 15 bar steam for ease of modelling The influence of the parame¬

ter is bigger for the 5 bar than for the 15 bar steam a change of ±20% (beingequal of a temperature change from <0 °C to >220 °C according to data given in

(Lide 1995)) results in a change of ±1% for 15 bar and of ±3% for 5 bar steam

consumption This difference is due to the total amount of steam consumed,since the specific consumption of 5 bar steam is smaller than of 15 bar steam,

the relative influence of the condensation enthalpy is bigger for the 5 bar steam

Since the uncertainty in the value and the physical difference of the condensa¬

tion enthalpy between the two pressure levels is much smaller, the influence

may be considered as minimal The assumption of taking a mean value for the

energy content of the steam is therefore justified

Deviation of the

Parameter

Value from the

Base Case

H 50%

20%

DBase Case

E3 +20%

H +50%

— Measurement

Pressure Level

Figure 6-21: Sensitivity analysis of the steam condensation enthalpy AHv with

regard to steam consumption according to Equation (3-14)

(one month, PSP data)

122


6.3.5 Stirrer Input to the Reaction Vessels

As shown in Figure 6-22, the efficiency rj of the stirrer (i e,the amount of

electric energy that is actually transformed in mechanical energy inside the ves¬

sels and finally in thermal energy) has only minor influence on the total specific

consumption of steam and brine This is in agreement with the findings of

Chapter 6 3 6 Changes of ±50% of the efficiency of the stirrer (i e, heat

input rf), result in changes of about ±1 to 2% for the modelling results Heat

input by the stirring of the apparatus is therefore only of minor importance for

total consumption (see Chapter 6 3 6) Neglecting this parameter would onlyresult in minor inaccuracies and would probably make a future model easier to

program and use

1400

Deviation of the

Parameter

Value from the

Base Case

a 50%

0 20%

DBase Case

a +20%

El +50%

— Measurement

Utility

Figure 6-22: Sensitivity analysis of the stirrer input rj with regard to utility con¬

sumption according to Equation (3-14) (one month, PSP data)

123


6.3.6 Stirrer Electricity Consumption

The influence of the electricity consumption of the stirrer motors (and the

other small motors connected with the reaction vessel; i.e., P/Pn (y)) on the

total electricity consumption, the steam consumptions, and the brine consump¬

tion are presented in Figure 6-23. The influence on the electricity consumptionis the most significant of all four modelled utility consumptions. The more

electricity is consumed, the less steam is required and the more brine is con¬

sumed (although influence on these utilities is only minor). The electricity con¬

sumed is used for stirring the apparatus. As stated in Chapter 5.1, it first pro¬

vides mechanical energy to the apparatus that is secondly transformed to ther¬

mal energy (with an efficiency of rj, as discussed in Chapter 6.3.5). This ther¬

mal energy reduces the consumption of steam or increases the consumption of

brine as shown in Figure 6-23.

For electricity, a change of ±50% of electricity uptake efficiency (i.e.,P / Pn (y)) results in a change of total specific electricity consumption of about

±12%). It can be stated that base consumption has a significant influence on

electricity consumption, which decreases the influence of this parameter to total

electricity consumption. Nevertheless, the stirrers are responsible for a signifi¬cant part of total consumption. This is in agreement with the findings in Chap¬ter 6.2.3.

iooo —;

o

2?

Deviation of the

Parameter

Value from the

Base Case

a 50%

-20%

DBase Case

+20%

s +50%

— Measurement

Utility

Figure 6-23: Sensitivity analysis of the stirrer electricity consumption y with

regard to utility consumption according to Equation (3-14)(one month; PSP data)

124


6.3.7 Circulation Pump

Figure 6-24 presents the sensitivity analysis of the circulation pump effi¬

ciency (i.e., y) with regard to electricity consumption. Changes in the electricityuptake of the circulation pumps were considered. The influence on the steam

and brine consumption was not investigated since this influence was assumed

being minor and therefore not included in the model.

As found in the investigations of Chapter 6.2.3, the circulation pumps have a

significant influence on total electricity consumption despite their small nominal

powers. This is related to the activity times that are high for these apparatus.Since Equation (3-8) is directly dependent on the power uptake of the electric

equipment, the changes have a significant influence on the total electricity con¬

sumption. A decrease in power uptake of 50% results in a decrease of total

power consumption of about 6%>. This is not too significant for total consump¬

tion considering the accuracy of the BOTUMO but has to be regarded.

Deviation of the

Parameter

Value from the

Base Case

a 50%

-20%

DBase Case

+18%

— Measurement

Figure 6-24: Sensitivity analysis of the circulation pump efficiency y with re¬

gard to electricity consumption according to Equation (3-14)(one month; PSP data)

125


6.3.8 Vacuum Pumps

Figure 6-25 presents the results of the sensitivity analysis for the electricity

uptake of the vacuum pumps. A change of ±50% for the uptake efficiency y

(see Equation (3-8)) results in a change of ±5% for total specific electricity con¬

sumption. The change is therefore significant but not too important for the elec¬

tricity consumption of the whole building. Moreover, the fluctuations of the

parameters found for the measurements (see Chapter 5.4.1) are of no big impor¬tance to the total consumption. Therefore, the assumption of a constant con¬

sumption depicted in Equation (3-8) is valid for the vacuum pumps.

Deviation of the

Parameter

Value from the

Base Case

a 50%

20%

Base Case

D +20%

H +50%

— Measurement

Figure 6-25: Sensitivity analysis of the vacuum pump efficiency y with regardto electricity consumption according to Equation (3-14) (one month; PSP data)

126


6.3.9 APOVAC

Figure 6-26 depicts the sensitivity analysis of the electricity consumption of

the whole building when the parameter of the electricity uptake (y) of the

APOVAC pumps is changed according to Equation (3-8). It can be seen that,

although the parameter is changing from 50 to 150%) of its original value, the

consumption of the whole building is not changing significantly. The specific

consumption of the APOVAC apparatus group is changing (not shown) but the

total consumption stays approximately the same.

The slight fluctuation of the power ratio found in the measurements (seeChapter 5.4.2) has therefore no significant influence on the BOTUMO and the

assumption of a constant model according to Equation (3-8) was applicable.

£ 400

"

350

c

o= 300

£ziin

I 250

o

&g 200

El

•AX

•AX•axxX"XXXX

XXXX

XXXXXXXXXXXCxX

*xx

\ xx

*xx

ßßßßxxxßßß,ßxxxVXX

ß/pß/pß/pß/pß/pß/ßX#:^ ß ßXX

^ ß ßX#:^ ß ßXX

^ ß ßX#:^ ß ß

IIP

#11

Deviation of the

Parameter

Value from the

Base Case

a 50%

Q-20%

Base Case

ra +20%

s +50%

— Measurement

Figure 6-26: Sensitivity analysis of the APOVAC pumps efficiency y with re¬

gard to electricity consumption according to Equation (3-14)

(one month; PSP data)

127


6.3.10 Short Path Distillation

The modelling results of the sensitivity analysis for the power uptake of the

short path distillation motors are presented in Figure 6-27. Since this unit op¬

eration was performed the whole month, its consumption is significant despite a

low nominal power of the motors and the influence of the power consumptionon the total specific electricity consumption is significant too (see Chap¬ter 5.6.1). Since the efficiency was modelled according to Equation (3-8) with a

y of 96%) as the base case, no higher values were investigated.A 50%) decrease of the parameter value (i.e., y= 43%) results in a 6%> lower

output of the BOTUMO. The electricity consumption of this unit operation maytherefore not be neglected. This apparatus has a significant influence on the to¬

tal electricity consumption, but other parameters have a higher influence as

shown above.

500

450

400

"u 350

" 250

f"XX

'ßßßXXx'ßßß"XXx'ßßß"XXx'ßßß"XXx'ßßß"XXx

XXXXX

XXXXXXXXXXXXXXXXXXX ß

'

//d/

HPDeviation of the

Parameter

Value from the

Base Case

0-50%

0-20%

DBase Case

— Measurement

Figure 6-27: Sensitivity analysis of the short path distillation motors efficiency

y with regard to electricity consumption according to Equation (3-14)(one month; PSP data)

128


6.4 Conclusions

The modelling of a whole production building according to PR and PSP

with the help of the BOTUMO was performed. For such a BOTUMO the

SUOM developed in Chapter 5 were put together and summarised according to

the equations given in Chapter 3.2. The overall model was built according to

Equations (3-14), (3-15), and (3-16). Summation resulted in a model accordingto Equation (3-26) (or, in other words, to Equation (3-15)). This production de¬

pendent energy consumption is then inserted together with the infrastructure

consumption in Equation (3-14) to result in a model of the whole plant.The modelling according to the PR results in more accurate outcomes of the

model (in absolute terms). The results for short periods depict the problem that

it is not exactly known how many batches are produced during a short period.For modelling of longer periods than about two days, the PR showed to be a

data source too tedious to acquire. It was too time consuming and complicatedto extract the data. Therefore, a change to data extracted from the PSP was per¬

formed. For short periods (a few days), these models showed higher deviations

from the measurements than the PR-based ones. Nevertheless, they showed to

be significantly more accurate than the models used in daily production today.For longer periods (i.e., longer than about one week), the models built with the

PSP data showed good accuracy when compared to the measurements of the

whole building. The modelling according to PSP data is therefore possible and

the BOTUMO applicable for longer periods.With the help of the BOTUMO, it is possible to make detailed analyses of

the energy consumption of a whole production plant. Unlike the black-box

model presented in Chapter 4 (TODOMO), a breakdown of the total energy

consumption is possible. The energy consumption (the specific as well as the

absolute one) may be assigned to the different apparatus and products in the

plant. Analogous to Figure 3-2, it is possible to distinguish the energy con¬

sumption of the whole building for all of the different utilities, apparatus and

chemicals available in the plant.The analysis showed that the specific energy consumption is varying widely

for the different products. No mean specific consumption was found for the in¬

vestigated building. This explains why the TODOMO presented in Chapter 4

was not applicable for the investigated building (Building 1). In the TODOMO,mean specific energy consumption is postulated for all the different chemicals

available in a production plant. Modelling according to Equation (3-1) relies on

this mean specific energy consumption.As found in Chapter 5, losses are important for the steam and the brine con¬

sumption of a single apparatus. In this chapter, the losses showed to be impor¬tant for the total consumption of brine and steam as well. Overall, the apparatus

group reactors and nutsche dryers consumes the largest part of energy (apartfrom the electricity consumption of the infrastructure). These apparatus show

significant losses for brine and steam consumption. Focus should therefore be

put on optimising the losses of brine and steam operations. For brine opera¬

tions, the losses are smaller because only losses through heat transfer to the en¬

vironment have to be considered. For steam, losses may occur through heat

129

Conclusions

transfer through the wall, but also through suboptimally operating steam trapsand through badly sealing valves etc.

The model was able to show differences between the different apparatus in

terms of energy consumption as well. The model could therefore be used for

comparing the production of the same chemical in different apparatus.The results of the sensitivity analysis discussed in Chapter 6.3 are summa¬

rised in Table 6-5 and presented in Table A-26. The sensitivity analysis shows

that the influence of most of the parameters of the model is minor. Only the

loss coefficients K and the time t have significant influence on the outcome of

the calculations.

The loss coefficients K should therefore be investigated extensively in future

studies. More measurements would lead to more exact parameter values. With

the help of more exact parameter values, the model accuracy could be improvedand the model would become more reliable. This would improve transferabilityof the model as well.

Table 6-5: Summary of the sensitivity analysis of Chapter 6.3 showing the de¬

viation of the objective functions Em according to Equation (3-14) for changesin the parameter values of ±20%; modelling period: one month

Parameter Steam Electricity Brine

15 bar 5 bar

Stirrer efficiency ±0.3% ±0.3% ±4% ±1%

Stirrer input ±0.3% ±0.3% - ±1%

Circulation pump efficiency - - ±2% -

Vacuum pump efficiency - - ±2% -

Heat-chamber ventilator efficiency - - ±0.1% -

Short-path distillation efficiency - - -2% -

APOVAC efficiency - - ±1% -

Enthalpy of vaporisation (steam) ±0.1% ±3% - -

Loss coefficient (steam) ±9% ±6% - -

Loss coefficient (brine) - - - ±3%

Time ±9% ±6-8% ±6-7% ±9%

Time is the most influential parameter in the unit operation models as pre¬

sented in Table 6-5. Care should therefore be taken to acquire the most exact

values from the data given in the PSP. Sensitivity analyses should be performedagain for a new building to investigate the margin of deviation of the model out¬

puts according to this parameter.

130

Conclusions and Outlook

7 Conclusions and Outlook

Short conclusions and outlooks are already given separately at the end of

each chapter. Here these findings will be summarised and extended to give a

broad overview of the possibilities and future work that are related to the out¬

comes of this thesis.

7.1 Conclusions

The modelling of the energy consumption of batch production plants is pos¬

sible. Two different approaches were performed for energy modelling of whole

batch production plants: a top-down approach (TODOMO) and a bottom-up ap¬

proach (BOTUMO).The simpler of the two models, the TODOMO, has limited applicability and

several drawbacks. This model is only suitable for modelling the energy con¬

sumption of production plants where a mean specific energy consumption of the

products may be found. This is the case for monoproduct batch plants, for mul¬

tiproduct batch plants with similar products and for multiproduct batch plantswith constant production mix (on mass basis). For these buildings, it is possibleto extract the infrastructure consumption from the actual production dependentconsumption. This infrastructure consumption is responsible for a significant

part of total energy consumption especially for electricity. The production de¬

pendent energy consumption results in a specific energy consumption for all the

chemicals produced in the plant. The infrastructure consumption as well as the

specific product consumption of energy may then be compared to the consump¬

tions of other buildings. Furthermore, the production of the same chemical in

several different plants may be compared in terms of energy. This shows

whether or not the plants are comparable and which of the investigated ones is

the most efficient energy user. Focus may then be put either on infrastructure or

on production dependent consumption of energy. Optimisation potentials maybe found by challenging the plants against the most efficient ones.

The BOTUMO on the other hand requires more modelling effort than the

TODOMO but offers much more insight in the production processes and their

energy consumptions. It offers more insights in optimisation potentials than the

TODOMO is able to provide. The BOTUMO may be used for multipurposebatch production buildings with highly varying production mix and a large vari¬

ety of different production processes. The infrastructure consumption of the

building may either be measured or found by the help of a TODOMO. For sev¬

eral unit operations and apparatus, specific models were postulated in this thesis

and checked by measurements. These measurements led to single unit operationmodels (SUOM) programmed in Excel® worksheets. These SUOM were

checked on single apparatus basis (transferability and comparability) and found

to be accurate considering the uncertainties given (i.e., uncertain parameter val¬

ues and uncertain measurements). The summation of the different SUOM leads

to a BOTUMO of a whole plant. Programming was done in Excel®. The

SUOM (and therefore the BOTUMO as well) require only widely known pa¬

rameters for the apparatus and the chemicals and are built simple enough for

131

Conclusions

daily use in a production plant. With the help of these parameters, the generatedSUOM and the BOTUMO are transferable to other batch production plantswhere no (or a limited number of) measurements have to be taken to adapt the

model. The SUOM are fed with the production data originating from either PR

or PSP. The PR data was, nevertheless, much too tedious to extract for a period

longer than a few days. The BOTUMO based on the PSP data showed poorer

accuracy than the one based on the PR for short periods (e.g., one day). When

modelling longer periods (e.g., one month), the accuracy of the model based on

PSP data was good.The SUOM with the PSP data were then multiplied by the number of

batches produced during the investigated period and summarised according to

the general model provided in Equations (3-14), (3-15), and Equations (3-16) to

(3-26). The analysis of the energy consumption according to the modelled con¬

sumption showed several possibilities for energy savings (e.g., minimising the

losses or optimising the nominal power of the stirrer motors). The model also

showed, where the energy was consumed and which production processes

should be investigated in more detail. These highlighted production processes

offer the best possibility for large energy savings.A model for the heating steam of production plants was elaborated as well.

It was found that the heating steam consumption is only depending on air-

change of a production building, degree-days and a base consumption that

equals almost zero if no infrastructure equipment is connected with the heatingsteam system (see Equation (3-3)). This provides an easy-to-use tool for com¬

paring the heating efficiency of a building with a standard one. Moreover, it

shows how much heating steam could be saved by decreasing air-change or by

removing infrastructure consumption.The measurements of the steam, electricity and brine consumption were

conducted with measuring equipment available in the plant. Especially the

measurement of the brine showed to be difficult because of the high inaccura¬

cies of the three measured values (see Chapter B.2.2). The determination of the

brine consumption incorporated therefore more uncertainties than for steam and

electricity. This resulted in a worse correlation of the BOTUMO for the brine

consumption (compared to the other energies). Still, the deviation from the ac¬

tual value for the BOTUMO is much better than for the model used until now in

the investigated plant.The results summary of the modelling of the investigated plant during one

month with the help of the BOTUMO is shown in Figure 7-1, Figure 7-2, and

Figure 7-3 for steam, electricity, and brine consumption, respectively.

132


In Figure 7-1, the total modelled steam consumption of the investigated

building is analysed with the BOTUMO for the period of one month (with the

help of PSP data). The total modelled consumption for this month is

1,354 MWh. This is the actual, modelled consumption. Heat of reaction and

stirrer input reduce the modelled consumption for about 80 MWh and about

23 MWh, respectively. The hatched fields in Figure 7-1 represent the consump¬

tions not directly related to the chemistry of the process (i.e., base consumption,losses, etc.). This consumption is responsible for about 63%> of total steam con¬

sumption. Steam savings should therefore start not with the actual production

process but with the reduction of the base consumption, the losses, the heatingof the vessels, etc. It can be seen from Figure 7-1 as well that the apparatus

group reactors and nutsche dryers is responsible for the main part of the steam

consumption (i.e., about 90%; mainly because of the large losses). The more

detailed modelling of this apparatus group was therefore appropriate to help un¬

derstand its characteristics.

D Heat-Chambei

Steam Jet

HBase

Consumption

0 Reflux

Evaporation

Heating of

Substances

Ü Heating of

Apparatus

S Heating of

Heating-System

El Losses

Figure 7-1: Analysis of the total modelled steam consumption of the investi¬

gated plant (period: one month; PSP data; total consumption: 1,354 MWh; heat

of reaction: -80 MWh, stirrer input: -23 MWh)

133

Reactors &

Nutsche Dryer

89 4%

Conclusions

Figure 7-2 presents the modelled electricity consumption for the period of

one month The modelling was performed with the help of PSP data The total

consumption of the modelled month is about 315 MWh As the figure shows,about 50%) of total modelled consumption is caused by the building infrastruc¬

ture (base consumption) This finding corresponds with the findings of the

TODOMO of the different buildings in Chapter 4 3 2 As a rule of thumb, it can

be stated, that infrastructure equipment consumes about 50% of the electricity

consumption of a production building Optimisation and minimisation should

therefore start with the building infrastructure The apparatus group reactors

and nutsche dryers consumes about one third of the total electricity consump¬

tion and the vacuum pumps specific to processes consume about one sixth It is

therefore essential to switch-off the vacuum pumps if not in use Stirrer motors

are responsible for the largest part of the consumption of the reactors and

nutsche dryers (no electric heating is performed during the investigated month)If electric heating would be performed during the investigated month, the pic¬ture would be changed significantly as shown in Figure 6-10 and in Figure 6-13

Stirrer motors should be tried to optimise By reducing the nominal power of

the stirrer motors, efficiency of the motors would be improved and electricityconsumption would be reduced (see Chapter E in the Appendix as well)

Figure 7-2: Analysis of the total modelled electricity consumption of the inves¬

tigated plant (period one month, PSP data, total consumption 315 MWh)

134


The total modelled brine consumption for one month of operation of the in¬

vestigated building is presented in Figure 7-3. The total modelled consumptionis about 100 MWh for the investigated month. The model was based on PSP

data. It can be seen that the apparatus group reactors and nutsche dryers is re¬

sponsible for the largest consumption (about 80%> of total consumption or about

72%) of total consumption if enthalpy of reaction is excluded) of brine. As

above (see e.g., Figure 6-7 and Figure 6-8), heat of reaction is modelled and

listed separately from the apparatus group for reasons of transparency. The

hatched fields are once more the energy consumptions not related to or deter¬

mined by the chemistry. These consumptions (i.e., base consumption, coolingof apparatus, losses and stirrer input) are open for optimisation or minimisation.

Together, they are responsible for about 50% of total brine consumption.Therefore, significant reduction potentials in total brine consumption are re¬

vealed. The base consumption of the building (i.e., heat input from the main

circulation pumps and losses through the walls of the pipes) is responsible for

about one sixth of total consumption. This quite significant consumption may

be optimised as well. Another main consumer group are the APOVAC pumps.

Whether or not these systems really require the use of the low temperatures of

the brine or if cooling with water would be sufficient should be challenged in

further investigations. Significant savings would be achieved by the optimisa¬tion of this apparatus group.

D Short-Path

Distillation

EBase

Consumption

M Cooling of

Substance

Cooling of

Apparatus

ü Losses

M Stirrer Input

mAPOVAC

pumps

Heat of

Reaction

Figure 7-3: Analysis of the total modelled brine consumption of the investi¬

gated plant (period: one month; PSP data; total consumption: 100 MWh)

It can be seen from the abovementioned investigations on the total consump¬

tion of the different energy carriers, that a detailed analysis of an actual produc¬tion mix is possible with the help of the BOTUMO. This allows the user to

identify specific optimisation potentials. Focus may be put on the sensible unit

135

Reactors 8

Nutsche Dryers

79 5%

16 8%

Outlook

operations and apparatus groups and energy targets may be set according to the

possible savings found in a similar investigation as presented above.

7.2 Outlook

In future research, both the TODOMO and the BOTUMO should be tested

on data of further production plants. Although the TODOMO was applicable to

the plants investigated, its general applicability is not yet proven completely.Care should be taken to data acquisition, since the findings are only as good as

the underlying data. The BOTUMO on the other hand has to be transferred to

other production plants. In this thesis, the model was elaborated, built and

tested on the same production plant. The possibility of transferring the SUOM

to other plants exists. The models require only the most important product spe¬

cific data, standard data of the apparatus specifications and the base consump¬

tion of the building. The base consumption may be found by measuring the dif¬

ferent infrastructure consumers as it was done for this thesis or by performing a

TODOMO on the available data. Since, in this case, uncertainties would be

high for a multiproduct batch plant with high variability between products and

in production mix, the direct measurement should be the method of choice. If

apparatus are found in the further investigated production plants for which no

SUOM exists, these apparatus have first to be measured and new models, ac¬

cording to the existing ones should be developed. Then modelling with the helpof the PSP may be performed and the outcome analysed. Deviations from the

measured value could be discussed and analysed according to the analyses in

this thesis. If necessary, the models should be revised or adapted for the new

plant or new, generally applicable models may be found.

Since measurement possibilities were limited in the plant investigated (espe¬cially the measurements of the brine consumption), additional measurements of

unit operations requiring steam, electricity, and brine (with focus on brine)should be performed. These measurements would lead to a broader basis for the

parameter values used in the BOTUMO (see Table 5-7). Uncertainties would

decrease and modelling results would be more accurate and reliable. In addi¬

tion, transferability would be improved with these measurements, since variabil¬

ity between more different apparatus of the same kind would be accessed in

more detail.

If more accurate and detailed brine measurements would show a significantdeviation from the measured values in this thesis that could not be explained byuncertain parameters or random fluctuations in the outcomes of the model, sev¬

eral facts could be responsible for the deviations. First, the measured equipmentshould be extended by the safety heat exchangers of the reaction vessels. These

heat exchangers help to prevent the solvents from venting in the waste air sys¬

tem. These equipment units could not be measured so far. Their consumptionand losses remains unknown until other measurements are introduced. Other

minor consumers of brine (e.g., cooling down the washing solvents for filters)could also contribute to total brine consumption. Finally, and least importantly,the lack of simple models for incorporating the enthalpy of crystallisation could

be responsible for some part of the deviation between model and measurements.

136


Research should lead to simple, generally usable equations to incorporate the

heat of crystallisation in the SUOM.

Generally, uncertainty investigations (e.g., sensitivity analyses or Monte

Carlo simulations) could be performed for getting a better overview on how the

different parameters influence the model results. Since the parameters are inde¬

pendent, uncertainty investigations could be limited to sensitivity analyses as

done in this thesis.

With the set of apparatus models developed in this thesis and with future

additional apparatus models, investigations could be started how the energy

consumption changes for a process, performed in different apparatus. This

could mean, for example, to perform the drying of a product either in a nutsche

dryer, or a horizontal vacuum rotary dryer, or a spray dryer. The different appa¬

ratus would result in different energy consumptions and would reveal the most

energy-efficient apparatus for a specific product and unit operation. This would

lead to a further application of the model for optimisation of the energy con¬

sumption of a batch process or a complete batch production building. The infra¬

structure, a specific apparatus or a specific energy use aspect (e.g., energy

losses) could be checked for possibilities of optimisation. Optimisation in ener¬

getic means should be conducted together with optimisation or retrofit of the

plant. Energy consumption may not be considered without considering aspectsof product quality, plant usage, production schedule and the like. Therefore,

incorporating the apparatus models in other programs performing retrofit of

production plants could result in a further optimisation possibility and further

objective functions for these programs.

It was seen during the investigations performed in this thesis that most of the

times, the specific parameter values for the chemicals and products are not

available in literature or by company data. This drawback was overcome by

using the values of similar chemicals or standard values for organic compounds(e.g., a Cp of about 2 kJ / kg / K for generic organic compounds). For makingthe models usable for a broader range of products, it has to be considered to in¬

corporate models for predicting the required physical values for these products

(e.g., the group contribution theory discussed in (Daubert and Dannel 1985;Reid et al. 1987)). Another possibility would be to include data given in manu¬

als and databases (e.g., in (Daubert and Dannel 1984; Lide 1995; TRC 1998;VDI 1984)). This would make application in daily business easier since no

physical data would be required from literature.

Similar to the models for brine, electricity and steam, other utilities could be

implemented in the BOTUMO. This would increase the applicability of the

model, improve the knowledge of the model by daily application and providethe plant manager with a single, simple management tool for challenging the

utility consumption of his plant. Standard costs would be calculated more easilyand more accurate utility requirement planning could be performed.

It is seen by the use of the Excel® model that this form of the program is

useful for a first challenging of the model equations and the BOTUMO itself.

The advantage is, that the model may run on most computers available in indus¬

try and most people are familiar with the basic concepts of Excel® and may

therefore be able to use the program. Nevertheless, several drawbacks are re-

137

Outlook

lated to the use of a spreadsheet program. The equations are open to everybodyand links between spreadsheets are easily corrupted by transferring spreadsheetsfrom one computer to another. Calculation time is a problem as well, since the

many OLE-links between different spreadsheets, required for the modelling of

longer periods (PSP data) with the BOTUMO, slow down calculation times sig¬

nificantly. The input of data to the spreadsheet is similar to the written input to

the PR sheet. Therefore, data may be easily extracted from the PR data but

many data points are required for modelling even short periods. Input of PSP

data, on the other hand, requires more investigations and is not too intuitive.

Nevertheless, less data is required, since each PSP is only entered once (and af¬

terwards multiplied with the number of batches performed (n,) according to

Equation (3-26)). Since the basic equations of the model are all provided in

Chapter 3 of this thesis, an easier user interface could be programmed that al¬

lows a simple input of the data. This would improve acceptability and transfer¬

ability and would prevent that the base data (i.e., apparatus specifications and

base consumptions) would be corrupted by user manipulations. Many of the

analysis graphs, made by hand for this thesis, could be included in the program,

therefore making the standard analysis of a plant easier for the user.

The model could be used for modelling the energy consumption of proc¬

esses only known from laboratory results as well. Incorporation of time calcula¬

tions according to physical data (e.g., heating dynamics of the apparatus or reac¬

tion modelling as discussed in (Fogler 1999; Levenspiel 1999)) could be possi¬ble in early phases of reaction engineering. This would provide the Chemical

Engineer in the research phase of a project with a first decision tool how to op¬

timise energy consumption of his processes. He would see what effect his op¬

timisation efforts would have on the final process. This would lead to more ef¬

ficient optimisation and better design in terms of energy.

The measurements showed, that reaction vessels with 15 bar steam had

lower loss factors than with 5 bar steam. This could be because the steam traps

operate more efficient for higher pressures or because of better insulation or

other reasons. Detailed investigations should be performed to find the reason of

this fact or to show that it is a random finding related to the inaccuracy of the

measurements.

Although this thesis is mainly focusing on chemical industry, in principlethe model is not limited to the chemical industry. The basic concepts could be

applied in any multipurpose batch production as the concept of the top-downand bottom-up modelling of energy consumption is well known in industry (see

e.g., (Aebischer et al. 1988; Kubier 1986; Kuczmowski and Weyant 1990; Wer-

bos 1990)). The approach presented in this thesis with its easy to use model

equations requiring only the most important specifications of a process opens

doors for a generally accepted methodology for modelling, comparing, and ana¬

lysing the energy consumption in industry. With today's challenging energy

and environmental questions (see Chapter 1) this could lead to a better under¬

standing of the processes and a focusing on the most promising saving poten¬tials.

138

Appendix

Appendix

A The Model

A.l The Assumptions for the BOTUMO

A.1.1 The Assumptions for the Single Unit Operation Models

The following main assumptions were taken for the modelling of the differ¬

ent unit operations:Heat of vaporization and heat capacity of the substances are con¬

stant for the temperature range investigated as mentioned in

(Dahinden 2003)

Dryers and reactors consist entirely either of stainless steel or of

normal steel - coatings are neglectedThere exists no time-independent loss coefficient

Pressure inside the reaction vessels is measured; the pressure inside

the dryers is about 100 mbar (while at the end of the drying process,

the pressure falls to about 40 mbar37)Calculations of the gas-phase are possible with the help of the An-

toine-equation given in (Atkins 1990) or (Wedler 1987)No simultaneous heating and cooling is considered

No fluctuation of the temperature by heating and cooling occurs

The whole apparatus is heated to the temperature of the heating

jacketThe losses are partly compensated by energy input of the stirrer; the

whole energy consumed by the stirrer and not lost in the gear is

transformed first to mechanical energy of the reaction media and fi¬

nally to heat

The losses are proportional to time, to the surface and to the tem¬

perature difference between the ambient temperature and the tem¬

perature of the heating jacketSteam is condensing at 5 or 15 bar at a specific temperature and a

specific heat of vaporisation; the difference of this temperature and

the temperature of the heating jacket is considered as well accordingto Equation (3-5)If no heat capacity and heat of vaporisation was available, the pa¬

rameters of a similar substance were taken either from (Lide 1995)or from (http://webbook.nist.gov/chemistry 2003); if no similar sub¬

stance was found, standard values of 2 kJ / kg / K for Cp and of

500 kJ / kg for AHv were used

Internal temperature of the dryer could not be below 40 °C accord¬

ing to discussions with industry experts even if Antione-equationresults in a lower temperature

37

According to discussions with industry experts

I

APPENDIX

Internal temperature of the equipment is either measured directly

during the process or extracted as a mean value from the PR (or the

PSP)

Optimisation of the loss coefficient was performed on the basis of

the whole process, while the process was split in different parts for

the modellingReaction enthalpies were taken from the PSP (based on final mass

of the process step)No side-reactions or decomposition-reactions were considered

Reflux was taken as a constant energy consumption unit operation,independent from the solvent (assumption that only a certain

amount of energy may be transferred to the inside of an apparatus;

see Chapter C.2 in the Appendix)No mixing enthalpy is considered (sufficient for the attained accu¬

racy)

Enthalpies, heat capacities and the like of mixtures are proportionalto the mass fraction of the corresponding components (sufficient for

the attained accuracy)Melting enthalpies are neglected as stated in (Dahinden 2003)Turbulent behaviour of the reaction mass is guaranteed in the stirred

and baffled reaction vessels

In batch distillation, a solvent is cleaned from impurities - heat of

vaporisation and heat capacity of the mixture are assumed to be the

same as the main component

Heating of the insulation material is neglected

A.1.2 The Assumptions for the Plant Model

Besides the assumptions mentioned in Chapter A.1.1 in the Appendix, the

following assumptions are underlying the plant model:

Infrastructure consumption is constant

Time given in PSP is reflecting the mean batch time (see Chap¬ter C.l in the Appendix as well)The temperatures given in the PSP are the ones actually performedThe temperatures of the heating jackets are only insignificantlyhigher than the (final) internal temperatures

Energy consumption due to reflux is a time dependent constant

Batch times are varying in the same way for all unit operationsThe loss coefficients of all unmeasured batch reactors are the same

and equal to the mean loss coefficient

The brine consumption of the safety heat exchangers of the reactor

vessels can be neglected

II

The Model

A.2 The Excel® Model

A.2.1 The Sheets of the Program

The different worksheets of the Excel® program are presented in this chap¬ter. Since the sheets are quite self-explanatory, no further explanations will be

given. Explanations of the input fields may be found in Chapter A.2.2 in the

Appendix. No further information on the input fields will be provided here.

The different equations used are described in Chapter 3.2. The actual models

are presented in Chapter 5. Only the most important sheets (i.e., Daten,

Auswertung, Berechnung, Auflieizen, and Verdampfen) are presented here. Only

simple calculations are performed on the other sheets, which will not be shown

in detail. The general structure of the program and the different sheets is givenin Chapter 6.1.1. A detailed overview of the different sheets and their equationscan be found in (Dahinden 2003) as well.

U V W X Y Z AA AB AC AD AE AF

7 Materials

8

Ea

aj

tn

ro

5

o

H

tn

ro

5

o

H < <

E

H

-oc

ro

Bro

Q

3

O

I

tn

C

5 1 2 3 4

9 fbarl ftl fkql [h min] [min] fhl [min] [kg] [kg] [kg] [kg]10 0 0 0 0

11

Li."U_

0

ro ---.

P. in

-1^ Q

O <

O ">

J«

¥ CD

ooo

>II

<

m

g,5Dcon

CN

fc-Ä

s-sO) i

^ CD

f ^Il Li.

<<+oCD

NII

CD+

LL

m

ia:D

OIII

iLU

HD

Z

5II

m

<n

O -.

So

ü-5o <O) o"O) T-

G) CÛJl <^ +

Q ^85ll m

Il co"

ü<n

O Ätf> o

t.oo <O)" o"O) T-

" <^ +

Q ^©5LL. <»

Il co"

Q<II

O Ätf> o

t.Q0 <G)" Ö"O) T-

O) Q1 <^ +

Q ^©5LL. <»

Il co"

LU

<II

O —©ô

ü-Eo <O) o"O) T-

O) LU

<L<^z +

Q ^©5LL. <»

Il co"

Figure A-l: Equations on the input sheet Daten in the Excel model

III

APPENDIX

B C D E F G H 1 J K

4

Examination of

the Consumption

Reactors &

nutsche

dryers

Reactors &

nutsche

dryers period

total

period

Base

consumption

Reaction

enthalpy

Heat-

chamber Steam jet

Vacuum

ump

5 15 bar Steam [kg]

CT)CT)

O

g

0

cn

-C

o

0

0

CÛ

W OII o

O CO

m ,°

>_ ?» o>

ro 2> ® g

UT03- tim

3° S o

(0ID2loIl co CO o

CD

3

+

LU

CO

Q

0

0

EÇOCÖQ.

"0a

0

^0

C0CO

CM

*3-

0

0

sCÖ

0

6 5 bar Steam [kg]

CT)CT)

Q_

a!

Ocn

-C

o

0

0

CÛ

3_

« O

II o

o co

in,

o

n°-

in D_

=A. | X 9,

»SX |

>_g>» D)

0) ü c f=«!î8

u_ m. ü m

3u ^ o

« "3 3 CD

Il co « o

S"

CD

CD

3

+CD

LU

CO

Q

0

0

E

ÇOCÖ_Q.

"0a

0

^0

COCO

CM

_i

0

0

sCÖ

0

1^

z

5

CD

z

"5

0

Q

coQ

Q_

>

7 Electricity 500 V [kWh]

CT)CT)

>-

0cn

-C

o

0

0

CÛ

« O

II o

o co

m,

°

m >~ m >-

=A-|J{ t

» §>>-" 1>- §

« ?0? ^ ^ Z5

C J=» =

SES»Q 0 ^ aiÙTm. Q. CD

3° S o

« <° 3 CDIl co « o

1^

CD

3

+

l^

LU

CO

Q

0

0

E

ÇOCÖ_Q.

"0a

0

^0

m

*3-CM

0

CD

S

"5

0

Q

coQ

Q_

>

8 Electricity 400 V [kWh]

cÔ"

co

CD

3

+

co

LU

Q

0

0

E

ÇOCÖ_Q.

"0a

0

^0

CM1^

Il CO

9 Brine [kWh]

CT)CT)

<

<

0

cn

-C

o

0

0

CÛ

3_

« CMII o

=(SUMIF(Dateni$Y$Y,">=45

0

3

6

00",BerechnungeniAIAl)-

SUMIF(Dateni$Y$Y,">=55

03

06

00",BerechnungenIAIAI))

ctT

CT)

CD

3

+CT)

LU

Q

0

0

sCÖ

0

LY+

Z

0

0

sCÖ

0

LY

CD

g

"i/o0

Q

coQ

Q_

>

®.Figure A-2: Equations on the sheet Auswertung in the Excel model

IV

The Model

Reactor Product Electricity11 No No Batch No Steam [kg] [kWh] Brine [kWh]

0

tn w 0o< 1

ïzl ff^l <N < JCM >- c

LU g, CD LU S o

LU 0 0

N

N)

IF($

erechnun $E12)Be

Y)

IF($ -hnung 2)BereAl)

IF(

echnun E12)Be

>- 2 lu< 0 »

?£ 53 Q »

a a S1 5"5

C CO CM0^ *-

Ö) CM Q

s 5"^ » CM -i QS

12

Berec TE($C12 ATE($C1C12

Bere TE($C12 ATE($C12 77 sa

läeöCM LU Z

ana O < LU» < ffi

l$E$E NCATE ONCATLU Z

W

»LU £LU t O

»Q z Dateni$E$EP

CONCATESAO

CONCADaten PCO AOC P8u-»Q

u-< » u-

» O

s » o s Q-S3 < » 3 CL < 3 < S?

« » c « < »

2=0S

ateni$ Daten)))2=0 aten Date 2=0 eni$ aten

Reactor ($D1 F(D MIF( N))) 5 M Q~

» asr «£•== ItïïSD>:

($D1 F(D MIFAl

A

12 Consumptionu- S 3 Z

m 3 tn c TT =3<n S13

Product Steam 5 bar Steam 15 Electricity14 No Batch No [kg] bar[kg] 500 V [kWh] Brine [kWh]

O CO O co O CO O coo „ O „ O „ O „

z Ï? zï z £ zï?< Q < Q < Q < Q» » » » » » » »

Z lo Z LO Z LO Z LO ^

» O 0? »oô »uP »o<c sao. c sao c sa>- c Sâ<0 LU c

0 LU c0 LU c

0 LU c

05 I— 0 05 I— 0 05 l— 0 05 l— 0

Q < O) Q < O) Q < O) Q < O)

Li LU§ LiLU § Li LU

§ LiLU §

BatchS t 1 S t 1 S t 1 S t 1

15 Consumption»OSII Z 0

»ÜSIl Z 0

»ÜSIl Z 0

in o 2Il Z 0

Figure A-2 (continued): Equations on the sheet Auswertung in the Excel

model

Figure A-3: Equations on the sheet Berechnungen in the Excel model

V

APPENDIX

B C D E F G H

8 Cooling (Brine) Heating (Steam)

9 Date& [kJ] [kJ] Substance No

10 Time Sum Sum 1 2 3 4

11

c

Sro

Qil

=IF(Datenil12<15,IF((Dateni$H12- Dateni$H11)<0,IF(Datenil11>30,SUM(E11P11

)/(Dateni$H

12-Dateni$H11

)*(Dateni$H12-

30),SUM(E11P11)),0),0)

SX "

ca JIC CL

ro ^Q LU

uTS—- 3O CO

A 9CM A^

pasLL 'Sm Q

<uNC

S

"îflSi

CO

LU CNca <-

LU X

ça S?

S Ün os

sa

ca Ülu os "

srp-3 " X

^ x ca

O"*

c

O x üi ca m

m c Q

<uNC

ro

"wsi

3

CO

LU CN

ca <-

LLI Üça 2T

s Jn ni

£ Qtu ^

11ca S.il os "

srp-3 " X

5: x ca

O"*

c

O x üi ca ni

m c Q

NC

ro

"w.q

CO .

LU?!ca x

lu ca

ça c

a "Sc G-ni *

"w ^~

§5co <

o" c

<- <u

oSp3 X X

\£ ca ca

_i c ro

Il (D Q

(DNC

ro

"w.q

CO

silu x

ca *a

c c

0) (1)

n ro

S G-w *

Si^

«Sica Sx ro "

3 " X

5: x ca

O"*

c

O x üi ^ ni

il c Û

Figure A-4: Equations on the sheet Aufheizen in the Excel model

B D E F G H

8 Evaporation9 [kJ] Steam Substances

10 Date & Time Sum 1 2 3 4

11

>

a

CD+-<

03

QII

<m

LU

3

W

Wm

<

o

OlII

A

Ol

C

CD+-<

03

Q

LL

II

co ,-

-Q,-

»S

ce">

lu ro

oTP

ils.!2 "Sce Q

LU Î-.

c ce(D _

Ild) 03

> CO

Il -Q

i- 3

*- W

gws* ce

S LU

03 —

Q SLL N

^

Il 03 O

C0 i-

-Q 1-

«Io

c

^£U_ CD

oTP3 ^

o S

—'- (D

° "03

ce PLL Î-.

c ceCD _

Ild) 03

> CO

Il -Q

i- 3

*- W

gws* ce

S LU

03 —

Q SLL N

^

Il 03 O

C0 i-

ce">

«SCL *

3 ^

o»

—' CD

o "S

£p

c«*

Ä S»

E (D

03 N

T3 CSr. 03CD +-»

> w

^ w

O LUce ce

S= LUa> te

03 c

Q (D-_• NLL C ^^

ïiS°.

C0 i-

« ilo CT— (U

tûCL *

3 ^

O il9 =—1 0)

O CD

05- *

c^

E (D

CO N

T3 C

Sr. 03

> w

M "§^ W

O LUce ce

c LUa> te

03 c

Q (D--^ NLL C ^^

ïiS0.

Figure A-5: Equations on the sheet Verdampfen in the Excel model

VI

The Model

A.2.2 Description of the Required Input Data

The following table summarises and explains the required input data of the

program. Since the program was first programmed in German, the names of the

fields are mostly given in German. The explanations should allow, neverthe¬

less, even non-German speaking people to be able to use the program.

Table A-l: Required input data for the Excel models (sheet Daten)

Nomenclature in the Explanation Format/Units

Excel® sheet

Produkt Name of the substance -

Partie-Nr. Number of the batch -

Ziffer Put in 999 if a batch is finished com¬

pletely38999

Gerät Number of the apparatus -

Tag Date of the operation DD.MM.YY

Zeit Time of the operation hh:mm

Ti Inside temperature °C

Ta Outside temperature °C

Rührer Operation of the stirrer 0/1

Druck Normal pressure or vacuum (separate

modelling of large vacuum pumps)

0/1

Menge in Liter Volume of added/removed substance 1

Menge in kg Mass of added/removed substance kg

Abzug durch Ver¬ Whether or not substance was removed 0/1

dampfung by evaporationSubstanz Nr. Number of the substance according to

sheet "Substanzen"

Rückfluss Is reflux occurring 0/1

El Heizung Is the reactor heated with electricity 0/1

'

Needed for guaranteeing that the vessel is considered as empty in the model

VII

APPENDIX

Table A-2: Required input data for the Excel models (sheet Parameter)

Nomenclature in the Ex¬ Explanation Format/Units

cel® sheet

Wirkungsgrad el. Geräte Efficiency of the electric %

Reaktoren, DRN equipment (reactors and nut¬

sches)Wirkungsgrad: VP Efficiency of vacuum pumps %

Wirkungsgrad: WK Efficiency of the ventilator of

the heat-chambers

%

Wirkungsgrad: Dest Kol Efficiency of electric equipmentof the short path distillation

%

Enthalpie durch Dampf¬ Enthalpy of vaporisation of kJ/kgkondensation steam

K/m2 Loss coefficient for steam kJ/(min K m2)K/m2 Loss coefficient for brine kJ/(min K m2)Zeitverzug Time deviation for sensitivity -

Table A-3: Required input data for the Excel models (sheet WK)

Nomenclature in the Excel® Explanation Format/Units

sheet

NrWK No. of the heat-chamber -

Datum Ein Start date DD.MM.YY

Zeit Ein Start time hh:mm

Datum Aus End date DD.MM.YY

Zeit Aus End time hh:mm

Ti Inside temperature °C

Ta Outside temperature °C

Masse Mass of the heated substance kg

Cp Heat capacity of the heated sub¬

stance

kJ/(kgK)

Table A-4: Required input data for the Excel models (sheet VP-DS-Dest. Kol.)

Nomenclature in the Excel® sheet Explanation Format/Units

Art der Apparatur Description of the apparatus

("1" = vacuum pump,

"0" = steam jet,"-" = distillation column)

Gerät No. of the apparatus -

Datum Ein Start date DD.MM.YY

Zeit Ein Start time hh:mm

Datum Aus End date DD.MM.YY

Zeit Aus End time hh:mm

VIII

The Model

Table A-5: Required input data for the Excel models (sheet Reaktionen)

Nomenclature in the Excel® sheet Explanation Format/Units

Endmasse Final reaction mass kgdHr Enthalpy of reaction kJ/kg

Cp Heat capacity kJ/(kgK)Anz. Ansätze Number of batches -

kg Dampf Which utility is affected -

Table A-6: Required input data for the Excel models (sheet Substanzen)

Nomenclature in the Excel® Explanation Format/Units

sheet

ProdNr Identification number -

Dichte Density of the substance kg/m3Molmasse Molar mass of the substance kg/kmol

Cp Heat capacity of the substance kJ/(kgK)dHv Enthalpy of vaporisation of the

substance

kJ/kg

Table A-7: Required input data for the Excel models (sheet Geräte)

Nomenclature in the Explanation Format/Units

Excel® sheet

Nr Identification number -

Masse Mass of the apparatus without engine kg

cp der Apparatur Heat capacity of the apparatus' material kJ/(kgK)Inhalt Wasser Amount of water in the circulation

system

kg

Wärmeübergang Loss coefficient (where known exactly) kJ/(h K)Rührereintrag Coefficient of the heat input through

the stirrer

kW

Motor x Nominal power of the electric equip¬ment

kW

el. Heizung mit Pum¬ Electric heating 1/0

pe

Dampf5 bar, 15 bar Used steam of the apparatus 1/0

Dampfverbrauch Fixed consumption (if applicable) kg/h

IX

APPENDIX

A.3 The Results of the Model

In this section, the supplementary tables of the modelling results are pre¬

sented. They are extensively discussed and explained in Chapter 6.2.

A.3.1 Modelling Results

Table A-8: Measurement and modelling of the utility consumption of the inves¬

tigated plant

Period Modelling Utilities

Method Steam Electricity Brine

TMWhl rkWh/tl TMWhl rkWh/tl TMWhl rkWh/tl

PSP 46.5 2,602 12.7 711 -3.0 -169

One PR 39.3 2,203 11.5 643 -2.8 -158

Day CPM 40.1 2,248 11.1 626 -3.4 -192

Measure 110.4 6,185 98.0 549 -1.1 -62

PSP 87.1 2,756 23.5 742 -6.1 -192

Two PR 77.6 2,454 22.4 708 -7.1 -226

Days CPM 80.3 2,540 22.4 707 -7.5 -238

Measure 198.7 6,285 18.2 574 -2.2 -71

One

Week

PSP 270.6 2,366 77.2 675 -17.7 -155

CPM

Measure

267.1

635.8

2,336

5,560

78.5

56.2

686

491

-24.4

-6.7

-213

-58

One

Month

PSP 1,353.6 2,190 315.4 510 -99.8 -161

CPM

Measure

1,431.0

3,511.4

2,315

5,680

299.2

328.5

484

531

-119.1

-58.3

-193

-94

X

The Model

Table A-9: Modelling results for the total utility consumption of the investi¬

gated building


5bar 15bar total

[MWh] [kWh/t] [MWh] [kWh/t] [MWh] [kWh/t] [MWh] [kWh/t] [MWh] [kWh/t]Total

Reactors

&113.7 994 127.3 1,113 241.0 2,108 36.8 322 -13.7 -120

Nutsches

>

Heat-

Chamber2.2 20 0.0 0 2.2 20 0.3 3 0.0 0

<D Steam Jet 1.1 9 0.0 0 1.1 9 0.0 0 0.0 0

O Vacuum

PumpsBase

0.0 0 0.0 0 0.0 0 9.8 86 -0.7 -6

Consump¬ 13.2 115 13.2 115 26.3 230 30.2 264 -3.4 -29

tion

Total

Reactors

&537.2 869 673.3 1,089 1,210.5 1,958 106.5 172 -79.3 -128

Nutsches

co

s

Heat-

Chamber4.8 8 0.0 0 4.8 8 1.9 3 0.0 0

Steam Jet 6.6 11 0.0 0 6.6 11 0.0 0 0.0 0

O Vacuum

PumpsBase

0.0 0 0.0 0 0.0 0 55.7 90 -3.7 -6

Consump¬ 65.8 106 65.8 106 131.6 213 151.2 245 -16.8 -27

tion

Table A-10: Modelled steam consumption of one week of the reactors and

nutsche dryers of the investigated building

Product Reflux

[kJ]

Evapo¬ration

[kJ]

Heating of

Substances

[kJ]

Heatingof

Apparatus

[kJ]

Heatingof

Jacket

[kJ]

Losses

[kJ]

Stirrer

Input

[kJ]

Heat of

Reaction

B

C

F

G

J

L

M

N

O

2.8E+05

2.2E+07

1.7E+07

5.7E+07

0

0

0

0

1.2E+07

1.6E+07

1.9E+07

1.8E+07

2.1E+07

1.8E+06

0

0

8.0E+06

1.2E+06

2.2E+07

1.5E+07

1.7E+07

1.7E+07

1.6E+07

0

0

1.2E+07

6.8E+06

5.2E+06

1.6E+07

9.2E+06

1.6E+07

9.8E+06

0

0

3.0E+06

2.3E+06

4.3E+07

1.2E+07

8.7E+06

1.3E+07

7.8E+06

0

0

3.5E+06

4.9E+06

3.9E+07

8.9E+07

1.0E+08

1.3E+08

4.3E+07

0

0

2.2E+07

4.5E+07

-1.6E+06

-4.2E+06

-2.6E+06

-5.5E+06

-1.4E+06

0

0

-1.5E+06

-9.9E+05

-1.4E+07

-6.5E+06

-4.3E+05

-6.8E+06

-5.0E+06

0

0

-3.7E+06

-4.1E+05

Brine con¬

centration

Brine distil¬

lation

Decalcifica¬

tion

Cleaning

Preparation

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

4.7E+06

0

0

0

0

3.9E+06

0

0

0

0

4.1E+06

0

0

0

0

0

0

0

0

0

0

0

0

Sum 1.1E+08 8.5E+07 1.1E+08 6.6E+07 9.7E+07 4.7E+08 -1.8E+07 -3.7E+07

XI

APPENDIX

Table A-11: Modelled specific steam consumption of one week of the reactors

and nutsche dryers of the investigated building

Product Reflux

[kJ/t]

Evapo¬ration

[kJ/t]

Heatingof Sub¬

stances

[kJ/t]

HeatingofAppa¬ratus

[kJ/t]

Heatingof

Jacket

[kJ/t]

Losses

[kJ/t]

Stirrer

Input

[kJ/t]

Heat of

Reaction

[kJ/t]B

C

F

G

J

L

M

N

O

3 7E+04

91E+05

1 7E+06

5 3E+06

0

0

0

0

7 9E+05

2 2E+06

7 6E+05

1 8E+06

1 9E+06

6 2E+04

0

0

4 6E+05

7 9E+04

2 9E+06

6 OE+05

1 7E+06

1 6E+06

5 6E+05

0

0

7 2E+05

4 5E+05

7 OE+05

6 4E+05

9 2E+05

1 5E+06

3 3E+05

0

0

1 7E+05

1 5E+05

5 8E+06

5 OE+05

8 7E+05

1 2E+06

2 6E+05

0

0

2 OE+05

3 2E+05

5 3E+06

3 6E+06

1 OE+07

1 2E+07

1 5E+06

0

0

1 3E+06

3 OE+06

-2 1E+05

-1 7E+05

-2 6E+05

-5 1E+05

-4 8E+04

0

0

-8 5E+04

-6 6E+04

-1 9E+06

-2 6E+05

-4 3E+04

-6 3E+05

-1 7E+05

0

0

-2 1E+05

-2 7E+04

Brine concen¬

tration

Brine distilla¬

tion

Decalci¬

fication

Cleaning

Preparation

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

4 1E+04

0

0

0

0

3 4E+04

0

0

0

0

3 6E+04

0

0

0

0

0

0

0

0

0

0

0

0

Weighted Sum 9 6E+05 7 4E+05 9 3E+05 5 7E+05 8 5E+05 41E+06 -1 5E+05 -3 2E+05

Table A-12: Modelled electricity consumption of one week of the reactors and


Product Stirrer Vacuum Circulation Div. Electric APOVAC

Pump Pump Motors Heating[MWh] [MWh] [MWh] [MWh] [MWh] [MWh]

B 1.88 0.09 2.59 0.01 12.47 0.50

C 2.50 0.04 1.22 0.01 0.00 0.67

F 2.35 0.05 0.62 0.70 0.00 0.90

G 4.52 0.07 1.61 0.02 0.00 1.06

J 0.59 0.16 0.44 0.00 0.00 0.00

L 0.00 0.00 0.00 0.00 0.00 0.00

M 0.00 0.00 0.00 0.00 0.00 0.00

N 0.75 0.00 0.24 0.00 0.00 0.00

0 0.51 0.00 0.17 0.00 0.00 0.00

Brine concen¬ 0.00 0.00 0.00 0.00 0.00 0.00

tration

Brine distilla¬ 0.00 0.00 0.00 0.00 0.00 0.00

tion

Decalcification 0.00 0.00 0.06 0.01 0.00 0.00

Cleaning 0.00 0.00 0.00 0.00 0.00 0.00

Preparation 0.00 0.00 0.00 0.00 0.00 0.00

Sum 13.10 0.41 6.95 0.75 12.47 3.14

XII

The Model

Table A-13: Modelled specific electricity consumption of one week of the reac¬

tors and nutsche dryers of the investigated building


Pump Pump Motors Heating

[kWh/t] [kWh/t] [kWh/t] [kWh/t] [kWh/t] [kWh/t]B 254 12 349 2 1690 68

C 102 2 50 1 0 27

F 235 5 62 70 0 90

G 419 6 149 2 0 98

J 20 5 15 0 0 0

L 0 0 0 0 0 0

M 0 0 0 0 0 0

N 44 0 14 0 0 0

0 34 0 12 0 0 0

Brine concentration 0 0 0 0 0 0

Brine distillation 0 0 0 0 0 0

Decalcification 0 0 0 0 0 0

Cleaning 0 0 0 0 0 0

Preparation 0 0 0 0 0 0

Weighted Sum 115 4 61 7 109 27

Table A-14: Modelled brine consumption of one week of the reactors and


Products Cooling of

Substances

[GJ]

Cooling of

Apparatus[GJ]

Losses

[GJ]

Stirrer

Input

[GJ]

APOVA

C

[GJ]

Heat of

Reaction

[GJ]B

C

F

G

J

L

M

N

O

-2.14

-1.93

-1.09

-4.16

0.00

0.00

0.00

0.00

0.00

-0.72

-0.60

-0.39

-3.34

0.00

0.00

0.00

0.00

0.00

-0.81

-0.45

-1.44

-4.82

0.00

0.00

0.00

0.00

0.00

-0.23

-0.14

-0.23

-2.07

0.00

0.00

0.00

0.00

0.00

-3.24

-4.32

-5.81

-6.80

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

-4.49

0.00

0.00

0.00

0.00

0.00

Brine

concentration

Brine distillation

Decalcification

Cleaning

Preparation

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

Sum -9.31 -5.06 -7.51 -2.66 -20.17 -4.49

XIII

APPENDIX

Table A-15: Modelled specific brine consumption of one week of the reactors


Products Cooling of Cooling of Losses Stirrer APOVAC Heat of

Substances Apparatus Input Reaction

[MJ/t] [MJ/t] [MJ/t] [MJ/t] [MJ/t] [MJ/t]B -289 -98 -110 -31 -438 0

C -78 -24 -18 -6 -176 0

F -109 -39 -144 -23 -581 0

G -386 -311 -447 -192 -632 -417

J 0 0 0 0 0 0

L 0 0 0 0 0 0

M 0 0 0 0 0 0

N 0 0 0 0 0 0

0 0 0 0 0 0 0

Brine

concentration0 0 0 0 0 0

Brine

distillation0 0 0 0 0 0




Weighted Sum -81 -44 -66 -23 -176 -39

Table A-16: Modelled steam consumption of one month of the reactors and


Product Reflux

[kJl

Evapo¬ration

[kJ]

Heatingof Sub¬

stances

[kJ]

Heatingof Appa¬ratus

[kJ]

Heatingof

Jacket

[kJ]

Losses

[kJ]

Stirrer

Input

[kJ]

Heat of

Reaction

A

D

G

H

I

J

K

E

3.6E+7

1.3E+7

3.7E+8

3.8E+7

1.3E+8

0.0E+0

4.9E+7

2.3E+6

2.9E+7

1.9E+7

1.3E+8

4.1E+7

6.2E+7

9.4E+6

1.5E+7

4.3E+6

8.0E+7

3.7E+7

1.1E+8

6.8E+7

6.7E+7

8.4E+7

3.8E+7

2.8E+6

3.2E+7

1.3E+7

9.8E+7

2.3E+7

7.7E+7

4.1E+7

1.6E+7

3.3E+6

3.2E+7

1.6E+7

8.1E+7

4.8E+7

5.7E+7

3.9E+7

3.3E+7

4.5E+6

1.4E+8

7.8E+7

7.9E+8

4.4E+8

4.3E+8

1.2E+8

1.0E+8

2.3E+7

-4.2E+6

-3.1E+6

-3.4E+7

-9.2E+6

-1.3E+7

-9.9E+6

-7.5E+6

-9.1E+5

-1.6E+8

-4.9E+7

-4.3E+7

9.7E+6

-1.8E+7

-2.5E+7

-1.1E+6

0

Brine concentra¬

tion

Brine distillation

Ethanol distilla¬

tion

Decalcification

Cleaning

Preparation

0

0

0

0

0

5.3E+6

0

0

3.8E+8

0

0

0

6.5E+6

0

7.0E+7

0

0

2.8E+6

5.7E+5

0

2.7E+5

2.4E+7

0

5.7E+6

1.0E+6

0

0

2.0E+7

0

5.3E+6

2.1E+6

0

4.3E+7

2.1E+7

0

1.1E+7

-2.3E+5

0

0

0

0

-2.5E+5

0

0

0

0

0

0

Sum 6.4E+8 6.9E+8 5.7E+8 3.3E+8 3.4E+8 2.2E+9 -8.3E+7 -2.9E+8

XIV

The Model

Table A-17: Modelled specific steam consumption of one month of the reactors


Product Reflux

[kJ/t]

Evaporation

[kJ/t]

Heatingof

Substances

[kJ/t]

Heatingof

Apparatus

[kJ/t]

Heatingof Jacket

[kJ/t]

Losses

[kJ/t]

Stirrer

Input

[kJ/t]

Heat of

Reaction

[kJ/t]A

D

G

H

I

J

K

E

4 35E+5

7 54E+5

5 53E+6

2 08E+5

4 13E+6

0

5 90E+5

4 34E+5

3 58E+5

1 10E+6

1 91E+6

2 27E+5

2 03E+6

6 23E+4

1 84E+5

7 98E+5

9 76E+5

210E+6

1 68E+6

3 75E+5

219E+6

5 57E+5

4 57E+5

516E+5

3 95E+5

7 53E+5

1 44E+6

1 27E+5

2 50E+6

2 74E+5

1 97E+5

6 12E+5

3 87E+5

9 34E+5

1 19E+6

2 66E+5

1 85E+6

2 61E+5

3 96E+5

8 39E+5

1 71E+6

4 44E+6

1 17E+7

2 41E+6

1 40E+7

8 27E+5

1 25E+6

4 26E+6

-5 09E+4

-1 77E+5

-5 06E+5

-5 07E+4

-4 23E+5

-6 55E+4

-9 14E+4

-1 71E+5

-1 95E+6

-2 79E+6

-6 31E+5

5 31E+4

-5 86E+5

-1 69E+5

-1 36E+4

0

Brine

concentration

Brine distillation

Ethanol

distillation

Decalcification

Cleaning

Preparation

0

0

0

0

0

8 55E+3

0

0

6 14E+5

0

0

0

1 05E+4

0

1 13E+5

0

0

4 56E+3

9 21E+2

0

4 41E+2

3 80E+4

0

915E+3

1 65E+3

0

0

319E+4

0

8 58E+3

3 36E+3

0

6 88E+4

3 35E+4

0

1 81E+4

-3 78E+2

0

0

0

0

-4 09E+2

0

0

0

0

0

0

Weighted Sum 1 04E+6 1 12E+6 9 21E+5 5 41E+5 5 44E+5 3 57E+6 -1 34E+5 -4 63E+5

Table A-18: Modelled electricity consumption of one month of the reactors and


Product Stirrer

[MWh]

Vacuum

Pump

[MWh]

Circulation

Pump

[MWh]

Div.

Motors

[MWh]

Electric

Heating

[MWh]

APOVAC

[MWh]A

D

G

H

I

J

K

E

3.51

4.23

28.31

4.94

9.63

4.55

3.96

1.00

0.00

1.09

0.41

0.00

0.48

0.00

0.37

0.05

2.80

1.67

10.15

1.83

3.12

2.51

1.09

0.35

0.00

0.03

0.12

0.00

2.07

0.00

0.00

0.01

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

6.65

0.00

7.93

0.00

0.00

0.00

Brine concen¬

tration

Brine

distillation

Ethanol

distillation

Decalcification

Cleaning

Preparation

0.17

0.00

0.00

0.00

0.00

0.14

0.00

0.00

0.00

0.00

0.00

0.01

0.19

0.00

2.64

0.29

0.00

0.09

0.00

0.00

0.00

0.03

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.05

Sum 60.46 2.41 26.74 2.27 0.00 14.63

XV

APPENDIX

Table A-19: Modelled specific electricity consumption of one month of

the reactors and nutsche dryers of the investigated building


Pump Pump Motors Heating

[kWh/t] [kWh/t] [kWh/t] [kWh/t] [kWh/t] [kWh/t]A 43 0 34 0 0 0

D 240 62 95 2 0 0

G 418 6 150 2 0 98

H 27 0 10 0 0 0

I 313 15 102 68 0 258

J 30 0 17 0 0 0

K 48 4 13 0 0 0

E 188 8 66 1 0 0

Brine concentra¬

tion0 0 0 0 0 0

Brine distillation 0 0 0 0 0 0

Ethanol distillation 0 0 4 0 0 0




Weighted Sum 98 4 43 4 0 24

Table A-20: Modelled brine consumption of one month of the reactors and


Products Cooling of

Substances

[GJ]

Coolingof

Apparatus

[GJ]

Losses

[GJ]

Stirrer

Input

[GJ1

APOVAC

[GJ]

Heat of Re¬

action

[GJ]A

D

G

H

I

J

K

E

-1.8

-3.1

-26.1

0.0

-18.9

0.0

0.0

-0.4

-10.4

-3.0

-21.0

0.0

-2.2

0.0

0.0

-0.3

-3.8

-8.2

-30.3

0.0

-14.8

0.0

0.0

-1.1

-1.0

-1.3

-13.0

0.0

-2.3

0.0

0.0

-0.2

0.0

0.0

-42.8

0.0

-51.0

0.0

0.0

0.0

0.0

0.0

-28.2

0.0

0.0

0.0

0.0

0.0

Brine

concentration

Brine

distillation

Ethanol

distillation

Decalcification

Cleaning

Preparation

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

-0.3

0.0

0.0

0.0

0.0

0.0

0.0

Sum -50.4 -36.9 -58.1 -17.8 -94.1 -28.2

XVI

The Model

Table A-21: Modelled specific brine consumption of one month of the reactors


Products Cooling of

Substances

[MJ/t]

Cooling of

Apparatus

[MJ/t]

Losses

[MJ/t]

Stirrer

Input

[MJ/t]

APOVAC

[MJ/t]

Heat of

Reaction

[MJ/t]A

D

G

H

I

J

K

E

-21.4

-178.4

-386.3

0.0

-616.3

0.0

0.0

-78.7

-127.0

-171.3

-310.7

0.0

-73.2

0.0

0.0

-47.4

-46.2

-463.3

-447.3

0.0

-480.9

0.0

0.0

-215.3

-12.4

-74.7

-191.9

0.0

-74.4

0.0

0.0

-39.3

0.0

0.0

-632.0

0.0

-1659.4

0.0

0.0

0.0

0.0

0.0

-417.1

0.0

0.0

0.0

0.0

0.0

Brine concentration

Brine distillation

Ethanol distillation

Decalcification

Cleaning

Preparation

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

-0.5

0.0

0.0

0.0

0.0

0.0

0.0

Weighted Sum -81.5 -59.7 -94.1 -28.8 -152.2 -45.7

Table A-22: Percentage of utility consumption of the produced chemicals

Product

MeasuringPeriod39

A

M

B

W

C

w

D

M

E

M

F

W

G

M

G

W

H

M

I

M

J

M

J

W

K

M

N

W

O

W

1

Reflux

Evaporation

Heating of

Substances

Heating of

Apparatus

Heating of

Jacket

Losses

Stirrer InputHeat of

Reaction

[%]

[%]

[%]

[%]

[%]

[%]

[%]

19

16

43

18

17

76

-2

87

0

15

20

5

39

36

-1

13

14

11

9

10

8

55

-3

-4

11

15

30

11

13

62

-2

39

6

11

7

8

12

58

-2

0

10

11

10

5

5

60

-2

0

25

9

8

6

5

52

-2

-3

24

9

7

7

6

53

-2

-3

6

6

10

4

7

67

-1

1

16

8

9

10

7

54

-2

-2

0

4

32

16

15

47

-4

-10

0

3

23

14

11

59

-2

-7

20

6

15

7

13

42

-3

0

0

18

29

7

8

50

-3

-8

17

2

10

3

7

64

-1

-1

Stirrer

Vacuum

PumpCirculation

PumpDiv Motors

Electric

HeatingAPOVAC

[%]

[%]

[%]

[%]

[%]

[%]

56

0

44

0

0

0

11

1

15

0

71

3

56

1

27

0

0

15

60

16

24

0

0

0

71

3

25

0

0

0

51

1

13

15

0

19

62

1

22

0

0

15

62

1

22

0

0

15

73

0

27

0

0

0

41

2

13

9

0

34

64

0

36

0

0

0

50

13

37

0

0

0

73

7

20

0

0

0

76

0

24

0

0

0

75

0

25

0

0

0

Ö

m

Cooling of

Substances

Cooling of

ApparatusLosses

Stirrer InputAPOVAC

Heat of

Reaction

[%]

[%]

[%]

[%]

[%]

10

61

22

6

0

0

30

10

11

3

45

0

26

8

6

2

58

0

20

19

52

8

0

0

21

12

57

10

0

0

12

4

16

3

65

0

16

13

19

8

26

17

16

13

19

8

26

17

N/A

N/A

N/A

N/A

N/A

N/A

21

3

17

3

57

0

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

M: Month, W: Week

N/A: Not Applicable (i.e., does not consume any brine)

XVII

XwPhPh

<

a

00

t«

!-H

eu

-M

<+HOeo3£eoeu

eo

o

'oeu

t«

T3eu

"S

T3Oeu

3s«

H

Wa

oo

-gp

Ü

B2

Oo

++

oooooooooWWoooooooooooooooo

^hm

oooooooooooooooooo

oooooooo

oooooooo

ooooooooooooo

oooo

mes

oo

++

ww

oo

vqo

mes

o+wOSoooooooooooooooooooo

o+wm

es

m

oo

++

oooooWoWooooooooooooooooooo

t-

^

o+w^o

*'

oooooooo

mmmm

(N

m

oooo

Oo

++

++

++

WWWWOOOWOooW

oOS

OS

(N

^H

00

'n

'n

^o

<N

«I

^H

oooooo

es

mm

Ooo

++

+WooWWoooooooooooooooo

O"**

eso+WOS

ooooooooooooooooo

es

es

oo

++

ww

00

^h

00

.-!

ooooo

HNmitifiiOh0!!O\Oh

(Sn*rtHHHH(S|(S|(N(S|(N(N(NM

X

to

to

-O

Os

to

ooo

os

Lti

4^

OJ

tO

> fa

Vi p

^WK)MOÔOOslOiW|iWtOHo

oooOOOOOOOOOOOOOO

oooooooÜoo>

iSoooooooooooooooto

woO

OJOoooooÖ

ooojJÔOOÔO^^^^ÔO

ooooooôooQ

ooooooooooooooooo

oo

-O°ooooooaa

00

ooooooooooooooo

ooooo

tooo

ô

^

^1

oHH

rH>

o

ooooooooooooooooo

ogoooooooo

<^

ooooooooooooooooooooooooo0ow

°ôooooooooooooooo

ô

Os°o

£oooooow

H AT

4-

O

X X

5:o

oS'

32.

S° o Vi

O >-+)

tr

fa

fa

Vi

H g* o en

X X

slO\WJiW(OHOlOOOslO\WJiUK)HO

> fa [1 (/I

p

oooootoooooooooooooooooootoo

00

vo

o>

OOOOOOOOOOOOOOOOo000^00000

o

Ö

OOOOOOOOOOOOOOOOOOOOOOOOOO

oo

os^

oo

o

o

o

oooooooooooooooooooooooooo

oKg


oOs

op


o<^


ow

OOOOOOOOOOOOOOOO0000000000

Os

o

w

H AT

O &.

(/I

<T>O o a*

3.

O3

n>o

3§

Ovi

3c

E?3 5' 3 Vi O >-+)

r+

3-

fa

fa

Vi

> W X

S 3- S5

The Model

A.3.2 Sensitivity Analysis

Table A-26: Results of the different sensitivity analysis for one month'

Parameter

O

1

151

"oco

[kg]

>ar Steam

<D

&

[kg/t]

o

>

Q

[%]

5b

"oco

[kg]

ar Steam

<D

&

[kg/t]

o

>

Q

[%1

El

"oco

[kWh]

ectncity

<D

&

[kWh/t]

o

>

Q

[%1

I

"oco

[kWh]

înne

<D

&

[kWh/t]

o

>

Q

[%]

Efficiencyof stirrer

(P/Pn) &

Efficiencyof vacuum

pump

(reactor)& Effi¬

ciency of

div mo¬

tors

05

08

1

12

1 5

1 14E+6

1 14E+6

1 13E+6

1 13E+6

1 12E+6

1 85E+3

1 84E+3

1 83E+3

1 83E+3

1 82E+3

0

0

-1

-1

-2

9 49E+5

9 45E+5

9 41E+5

9 38E+5

9 33E+5

1 54E+3

1 53E+3

1 52E+3

1 52E+3

1 51E+3

-10

-10

-11

-11

-11

2 59E+5

2 78E+5

2 90E+5

3 03E+5

3 22E+5

419E+2

4 49E+2

4 70E+2

4 90E+2

5 20E+2

-4

3

8

12

19

-9 73E+4

-9 88E+4

-9 98E+4

-1 01E+5

-1 02E+5

-1 57E+2

-1 60E+2

-1 62E+2

-1 63E+2

-1 66E+2

-18

-17

-16

-15

-14

Efficiencyof stirrer

(energy

input)

05

08

1

12

1 5

1 14E+6

1 14E+6

1 13E+6

1 13E+6

1 12E+6

1 85E+3

1 84E+3

1 83E+3

1 83E+3

1 82E+3

0

0

-1

-1

-2

9 49E+5

9 44E+5

9 41E+5

9 38E+5

9 34E+5

1 54E+3

1 53E+3

1 52E+3

1 52E+3

1 51E+3

-10

-10

-11

-11

-11

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

-9 74E+4

-9 88E+4

-9 98E+4

-1 01E+5

-1 02E+5

-1 58E+2

-1 60E+2

-1 62E+2

-1 63E+2

-1 65E+2

-18

-17

-16

-15

-14

Efficiencyof circula¬

tion pump

& pump

05

08

1

1 12

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

2 76E+5

2 84E+5

2 90E+5

2 95E+5

4 46E+2

4 60E+2

4 70E+2

4 78E+2

2

6

8

10

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

Efficiencyof vacuum

pumps

(equip¬

ment)

05

08

1

12

1 5

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

2 79E+5

2 85E+5

2 90E+5

2 95E+5

3 02E+5

4 51E+2

4 62E+2

4 70E+2

4 77E+2

4 88E+2

3

6

8

9

12

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

Efficiencyof ventila¬

tor of

heat-

chamber

05

08

1

12

1 5

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

2 89E+5

2 90E+5

2 90E+5

2 91E+5

2 91E+5

4 68E+2

4 69E+2

4 70E+2

4 70E+2

4 71E+2

7

8

8

8

8

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

Efficiencyof motors

for short

path dis¬

tillation

column

05

08

1

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

2 74E+5

2 84E+5

2 90E+5

4 43E+2

4 59E+2

4 70E+2

2

5

8

NI

NI

NI

NI

NI

NI

NI

NI

NI

Efficiencyof

APOVAC

pumps

05

08

1

12

1 5

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

2 83E+5

2 87E+5

2 90E+5

2 93E+5

2 97E+5

4 58E+2

4 65E+2

4 70E+2

4 74E+2

4 81E+2

5

7

8

9

10

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI: No /nfluence of the parameter to the outcome

From measured value

XXI

APPENDIX

Table A-26 (continued): Results of the different sensitivity analysis for one

month41

Parameter 15 bar Steam 5 bar Steam Electricity Bnne

O

1"o

lyO

o

>

Q

"olyO

o

>

Q

"olyO

o

>

Q

"olyO

o

>

Q

[kg] [kg/t] [%] [kg] [kg/t] [%] [kg] [kg/t] [%] [kg] [kg/t] [%]

Enthalpyof vapori¬

sation

(steam)

05

08

1

12

1 5

1 11E+6

1 13E+6

1 13E+6

1 14E+6

1 15E+6

1 80E+3

1 83E+3

1 83E+3

1 85E+3

1 86E+3

-3

-1

-1

0

1

8 37E+5

9 13E+5

9 41E+5

9 63E+5

9 88E+5

1 35E+3

1 48E+3

1 52E+3

1 56E+3

1 60E+3

-20

-13

-11

-8

-6

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

Loss coef¬

ficient for

heating

(steam)

05

08

1

12

1 5

8 97E+5

1 04E+6

1 13E+6

1 23E+6

1 37E+6

1 45E+3

1 69E+3

1 83E+3

1 98E+3

2 21E+3

-21

-8

-1

8

20

7 98E+5

8 88E+5

9 41E+5

9 99E+5

1 08E+6

129E+3

1 44E+3

1 52E+3

1 62E+3

1 76E+3

-24

-16

-11

-5

3

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

Loss coef¬

ficient for

cooling

(brine)

05

08

1

12

1 5

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

NI

-9 17E+4

-9 66E+4

-9 98E+4

-1 03E+5

-1 08E+5

-1 48E+2

-1 56E+2

-1 62E+2

-1 67E+2

-1 75E+2

-23

-19

-16

-14

-9

Time

05

08

1

12

1 5

8 61E+5

1 03E+6

1 13E+6

1 24E+6

1 40E+6

1 39E+3

1 67E+3

1 83E+3

2 01E+3

2 27E+3

-25

-10

-1

9

23

7 59E+5

8 69E+5

9 41E+5

9 97E+5

1 11E+6

1 23E+3

1 41E+3

1 52E+3

1 61E+3

1 79E+3

-28

-17

-11

-5

5

2 38E+5

2 70E+5

2 90E+5

3 06E+5

3 37E+5

3 85E+2

4 36E+2

4 70E+2

4 96E+2

5 46E+2

-12

0

8

14

25

-7 61E+4

-9 04E+4

-9 98E+4

-1 09E+5

-1 23E+5

-1 23E+2

-1 46E+2

-1 62E+2

-1 77E+2

-2 00E+2

-36

-24

-16

-8

4

XXII

The Measuring Equipment

B The Measuring Equipment

The measurements of the energy consumption of the whole building were

made with the help of the main measuring tools installed in the buildings. For

electricity measurements, a kWh-counter was installed. Eddy current flow me¬

ters as well as temperature and pressure probes were used for the measurement

of the steam consumption (pressure and temperature corrected steam measure¬

ments) and two temperature probes (PT100) as well as a conductive flow meter

were used to measure the consumption of cooling energy (brine).The measurements of the single unit operations were performed with trans¬

portable equipment. These equipment units will be described in detail in the

following chapters.

B.l Steam Measurements

B.l.l The Equipment

The steam measurements were conducted with a transportable steam meas¬

urement equipment depicted in Figure B-1. The steam entering the equipmentfrom the main pipe (5 or 15 bar) is first dried in a small cyclone. This ensures

that only saturated steam is measured in the device. The temperature and pres¬

sure of the steam are then measured in Tl and PI respectively. Afterwards, the

steam flows though the small pipe and through the eddy current flow meter Fl

into the apparatus. Depending on the flow though Fl, the valve opens or closes

the entrance to the bigger pipe that leads the steam through eddy current flow

meter F2. Then, the steam flows directly into the apparatus where the investi¬

gated process is conducted (e.g., a reaction vessel). The computer determines

the temperature- and pressure-corrected steam consumption in kg and plots it on

a sheet of paper.

Steam Inflow

from the pipe

1

Computer

tThpî

F2

F1Steam Outflow

to the Apparatus

Condensate

Figure B-1: Measuring principle for the steam measurements

Since measurement of an apparatus with this device required a direct inline

measurement, considerable labour was required to install the device. Because

of this, attention was paid which apparatus should be measured. The device was

explosion safe and could therefore be used with no limitation inside the produc¬tion plant.

XXIII

APPENDIX

B.1.2 The Accuracy

Industry experts considered the accuracy of eddy current flow meters as

high (within a few percent). Since the device was quite old (about 20 years),and was not used for a long time, preliminary experiments were conducted to

determine the accuracy.

For this purpose, the device was attached to a heated filter with direct steam.

After the filter, the condensate was collected in a 200 1 barrel and the mass of

the condensate was measured with the help of a balance. The results of these

measurements are shown in Table B-1. The deviation is in the order of magni¬tude of 10% and therefore ok for the purpose of this thesis. It can be assumed

that at the higher steam amounts usually consumed by production vessels, the

accuracy of the measuring device is even higher.

Table B-1: Test of the steam measurement device

ExperimentNo.

Measured

Steam

[kg]

Measured

Condensate

[kg]

Deviation

[%]1

2

51.7

116.8

47.1

102

-9.8

-14.5

3 82.6 77.2 -7.0

B.2 Measurement of Brine

B.2.1 The Equipment

The brine measurements were conducted with an ultrasonic flow meter (Por-taflow X from Fuji Electric43, depicted in Figure B-2) named Fl in Figure B-3

and two temperature probes (double-poled PT100) named Tl and T2 in FigureB-3.

Figure B-2: Scheme of the Portaflow X ultrasonic flow meter from Fuji Elec-

See http://ttiglobal.com/products/manuals/flcslx.pdf or

http://ttiglobal.com/Product.asp?Paraml=FLCSl

XXIV


With the help of the known heat capacity and density of the brine, the cool¬

ing energy consumption could be calculated according to Equation (3-5). The

temperature probes and the flow meter could be applied to a brine pipe from the

outside (after removing the insulation) and therefore, installation was quite easy.

A drawback was the fact, that the device was not explosion safe and, therefore,a special allowance was needed each time a measurement was conducted and

special care had to be taken during operation. The principle of the measure¬

ments is depicted in Figure B-3.

Brine

InletT1 - F1 T2

Brine

Outlet

Plotter

Figure B-3: Principle of the brine measurement

B.2.2 The Accuracy

The flow meter and the temperatures probes have different accuracies,which will be shortly discussed here. The general difficulties of brine meas¬

urements are described in (Churchman and Krueger 2003).The measurements of the brine consumption turned out to be difficult be¬

cause of the sensitivity of the measuring equipment. Only half filled pipes, dirt

inside the pipes or even noise outside in the plant could influence the measure¬

ments or make them impossible. Care has to be taken to obtain the exact pa¬

rameter values for the flow measurements and to isolate well the temperature

measurements so measuring errors could be minimised.

The ultrasonic flow meter is in principle a measurement instrument with a

high accuracy (less than 1% error, according to company information). Never¬

theless, the measurements are only as accurate as the input parameters. The ul¬

trasonic flow meter needs different input data for the measurements. Some, like

the dimensions of the pipe or the pipe material, are well known and accurate.

Others, like the sound velocity within the brine or the dynamic viscosity of the

brine, vary greatly with the composition and temperature of the brine. More¬

over, the composition and the temperature of the brine are not stable and vary

over time. The accurate values of these parameters at a mean composition are

not known from measurements but from literature data (e.g., (Lide 1995)). Two

different kinds of brines were in use: CaCb/Water and Ethylene-Glycol/Water

XXV

APPENDIX

mixtures. Data for the composition of the brines is provided by a contracting

company. The parameter values are shown in Table B-2. Moreover, noise in

the pipe and in the building (of motors, pumps, etc.), is also negatively influenc¬

ing a flow measurement based on ultrasonic sound. Because of these uncertain¬

ties, the accuracy of the flow measurements is not better than 10%.

Table B-2: Parameters for the flow measurements of the two kinds of brine44

Parameter CaCl2/Water Ethylene-Glycol/Water

Cp [kJ / kg / K] 2.67 2.81

p[kg/m3] 1299 1102

v[m2/s] 2.705-10"6 5.865-10"5

C [m / s] 1630 1661

The temperature measurements require only relative accuracy, since the dif¬

ference of the two temperature probes is essential for the calculation (see Equa¬tion (3-5)). Since inlet and outlet of the brine (see Figure B-3) are sometimes

localised far away from each other, the wire to one of the probes needed to be

prolonged. Because the probes had only two wires (and not three as usual todayfor PT100 probes), this resulted in a deviation of the measured temperature as

shown in Table B-3. It can be clearly seen that there is a difference between the

temperatures. This difference is more or less constant. A value of 4.4 °C was

taken as the mean deviation for correcting the measurements of the second tem¬

perature probe. Since the temperature difference on an apparatus lies between 2

and 4 °C, an inaccuracy of only 0.1 °C signifies a deviation of 2.5 to 5%. Such

an inaccuracy may occur at both probes and the deviation because of the longerwire is in the same order of magnitude resulting in an inaccuracy of about ±15%

only because of temperature measurements.

Information on the parameter values may be found in Bohne, D., Fischer, S., and Obermeier,E. (1984). "Thermal Conductivity, Density, Viscosity, and Prandtl-Numbers of Ethylene Glycol-Water Mixtures." Ber. Bunsenges. Phys. Chem., 88, 739-742.

Burton, C. J. (1948). "A Study of Ultrasonic Velocity and Absorption in Liquid Mixtures. "J

AcoustSocAm, 20(2), 186-199.

Corradini, F., Marchetti, A., Tagliazucchi, M., Tassi, L., and Tosi, G. (1995). "Thermodynamicsof Viscous Flow in Ethane-l,2-diol+Water Binary Mixtures." Aust. J. Chem., 48, 186-199.

Curme, G. O., and Johnston, F. (1953). "Glycols.", Reinhold Publishing Corp., New York.

Litowitz, T. A., Higgs, R., and Meister, R. (1954). "Ultrasonic Propagation and its Relation to

Molecular Structure in the Diols." J Chem Phys, 22(8), 1281-1283.

Reddy, V. K., Reddy, K. S., and Krishnaiah, A. (1994). "Excess Volumes, Speeds of Sound, and

Viscosities for Mixtures of 1,2-Ethanediol and Alkoxy Alcohols with Water at 308.15 K." J.

Chem. Eng. Data, 39, 615-617.

Schaaffs, W. (1941). "Über Schallgeschwindigkeit und Konstitution in flüssigen organischen

Verbindungen."^««Phys, 40(6), 27-404.

VDI. (1984). VDI-Warmeatlas, VDI-Verlag GmbH, Düsseldorf.

Weissler, A. (1948). "Ultrasonic Investigations of Molecular Properties of Liquids. II. The Al¬

cohols." JAm Chem Soc, 70, 1634-1640.

Willard, G. W. (1947). "Temperature Coefficient of Ultrasonic Velocity in Solutions." JAcoust

Soc Am, 19(1), 235-241.

XXVI


Table B-3: Temperature comparison of the two temperature probes

Tl T2 Difference

[°d [°q [°C]16.6 21 4.4

16.6 21 4.4

-15.9 -11.6 4.3

-16.1 -11.8 4.3

-17.3 -13 4.3

-17.7 -13.4 4.3

-18.6 -14.3 4.3

-18.2 -13.9 4.3

-17.9 -13.5 4.4

-17.2 -12.9 4.3

-16.7 -12.3 4.4

-16.2 -11.8 4.4

-16.7 -12.3 4.4

-17.2 -12.8 4.4

-18.1 -13.7 4.4

-18.5 -14.1 4.4

-18.0 -13.6 4.4

-17.4 -13.0 4.4

-16.9 -12.5 4.4

-16.4 -12 4.4

-16.1 -11.7 4.4

-16.5 -12.1 4.4

-16.6 -12.2 4.4

-17 -12.6 4.4

-17.5 -13.1 4.4

-18 -13.6 4.4

-18.3 -13.9 4.4

14.5 18.2 3.7

14.6 18.3 3.7

14.5 18.2 3.7

14.4 18.1 3.7

16.2 20.3 4.1

17.6 22 4.4

17 21.3 4.3

16.6 20.9 4.3

18.1 22.6 4.5

19.3 23.9 4.6

19.9 24.5 4.6

20.4 25 4.6

20.2 24.9 4.7

20 24.8 4.8

48.3 52.4 4.1

35.4 39.7 4.3

18.3 22.3 4

0.7 4.7 4

XXVII

APPENDIX

Combining temperature and flow measurements in a linear way according to

Equation (3-5), leads therefore to a deviation of about ±25%. This shows that

brine consumption is the most inaccurate one of the performed measurements.

In addition, the ultrasonic sensor requires a certain length of linear pipe for

its measurements (i.e., about 15 • diameter). Only this guarantees a linear flow

inside the pipe that is essential for accurate measurements of the velocity of the

fluid. Since this length was not available at all the apparatus using brine, meas¬

urements were limited to the apparatus with sufficient pipe length.

B.3 Measurement of Electricity Consumption

B.3.1 The Equipment

For the measurements of the electricity consumption, a LEM

Memobox 60S45 was used. Figure B-4 shows a picture of the successor model

of the used one, for giving an impression of how the measurements were con¬

ducted. The Memobox 603 measures the current and the tension voltage of an

electricity consumer directly. Out of these measurements, it calculates the

power consumption and other values for the given period. This device was

plugged in the wire at an external electricity distribution room. Therefore, ex¬

plosion risk was of no concern.

Figure B-4: Picture of a LEM Memobox 800'

B.3.2 The Accuracy

The measurements show a high accuracy because of the well-established

measuring method and the usage of a standard measuring device. Industry ex¬

perts and the manufacturer agreed in an accuracy of the measurements of maxi¬

mum ±2%). The electricity measurements were, therefore, the most accurate

ones performed in this thesis.

See http://www.lem.comSource: http://www.lem.com.cn/product/instrument/elecqual/memobox800.htm

XXVIII

Miscellaneous

C Miscellaneous

C.I Distributions of the Times Given in the PSP

For the sensitivity analysis, performed in Chapter 6.3, the relation between

the standard times given in the PSP and the actual times needs to be investi¬

gated. This was done with the help of detailed measurements of the production

processes for Products A, B, and G and with the measurements, mentioned in

Chapter 5 and presented in Chapter D.2 in the Appendix. As an example of the

company internal measurements, the results for Product A are presented in

Figure C-1. About ten batches are considered for each product. The "technical

times" are times slightly differing from the PSP values but considered as practi¬cal. It is seen that the actual times are close to the technical ones for the investi¬

gated building. This is not always the case according to industry experts.Whether or not the times given in the PSP are equivalent to the technical times

of a mean time of operation has to be investigated in each building from new.

-

d\,

1

1111 In] Il

1 ! tf jrpipll Iff I InJtl II

t-COLOI^G5t-COLOI^G5t-COLOI^G5t-COLOI^G5t-COLOI^G5t-t-t-t-t-t-CnICnICnICnICnICOCOCOCOCOÎ-Î-Î-Î-Î-LO

Production Step No.

Average B Standard Deviation D Technical

Figure C-1: Detailed investigations on production time distribution for the pro¬

duction process (Steps 1, 2,.., 50, 51) of Product A in two 4 m3 and one 10 m3

glass lined reaction vessels47

Production step means operations like transfer, filling, reaction, heating, etc.

XXIX

APPENDIX

15 45 16 00 16 15 16 30 16 45 17 00 17 15 17 30 17 45 18 00 18 15 18 30 18 45 19 00 19 15 19 30 19 45 20 00

Reaction Time [h:min]

D Measured Time PSP Time

Figure C-2: Time measurements of Product J in a 6.3 m glass-lined reactor

Figure C-2 and Figure C-3 present measurements as taken during the inves¬

tigations of this thesis (see Chapter D.2 in the Appendix). The two examplesshow that the times given in the PSP may have completely different correlations

with the measured values for different production steps and products (see also

the investigations presented in Chapter C.2 in the Appendix). Sometimes theyare about equal to the mean value and sometimes they are below or above the

measurements. For all the investigated unit operations and products, the distri¬

bution of the relation between the given time in the PSP and the average meas¬

ured time for the specific unit operation was varying widely. Table C-1 summa¬

rises the findings of the time investigations. It may be seen from this table that

the mean value of the relation of the times given in the PSP and the average

measured value are about one. Therefore, the sensitivity analysis, taking the

PSP times as base case and investigating behaviour of deviations from this base-

case value is appropriate.

15 30 16 00 16 30 17 00 17 30 18 00 18 30 19 00 19 30 20 00 20 30 2100 2130 22 00 22 30 23 00 23 30

Brine Cooling Time [h:min]

Measured Time PSP Time

Figure C-3: Time measurements of Product G in a 4 m glass-lined reactor

xxx

Miscellaneous

Table C-1: Time investigations of PSP and measurements

Unit Operation Average

PSP / Average measured time48

Standard Relative Min.

Deviation Standard Dev.

Max.

Fill

HeatingCoolingHold

Reaction

Stir

Distil

Transfer

Split Phases

Own Measurements49

124%

111%

99%

82%

92%

89%

98%

116%

112%

93%

55%

27%

24%

22%

8%

34%

11%

37%

41%

31%

44%

24%

25%

27%

9%

38%

11%

32%

37%

33%

40%

93%

57%

57%

81%

37%

90%

69%

83%

40%

190%

159%

141%

122%

101%

148%

113%

192%

159%

171%

Total50 102% 42% 41% 37% 267%

C.2 Reflux Conditions

Reflux conditions are a special sort of distillation conditions. In the PSP or

in the PR it is usually not mentioned, how much reflux has to be performed. It

is only written "... four hours of reflux..."

or "... hold the reaction mixture under

strong reflux conditions for half an hour to dry the solvent...". These unclear

statements prevented a detailed modelling of this unit operation.For investigating the reflux conditions, measurement of the steam consump¬

tion under "strong reflux" of a common solvent (1-butanol) was performed dur¬

ing normal operation (recovering of the reflux in a separate vessel). The results

of these measurements are presented in Table C-2. The measurements showed,that an average of about 12.4 1 /min of 1-butanol is evaporated under "strongreflux" conditions. This lead to the assumption (see Chapter A.l in the Appen¬

dix) that under reflux conditions, about 5,900 kJ/min (i.e., about 100 kW) of

steam are required for the evaporation of the solvent (not accounting for the

losses occurring during this period according to Equation (3-7))51. This value

was assumed the standard consumption for reflux conditions in the modelling of

the different unit operations of the investigated building.

Extreme values of more than 200% standard deviation were neglected49

See Chapter D.250The average is not weighted with the absolute time of the different processes. Processes as

Transfer or Fill are usually short processes, while processes as Reaction or Heating could be

much longer. The values show, nevertheless, that the assumption mentioned in Chapter A. 1 that

the times given in the PSP are equal to the average time of the process is fulfilled - at least for

the accuracy of this modelling approach51

Taking into account the enthalpy of vaporization of 1-butanol of about 584 kJ / kg and its

density of about 810 kg / m3 according to Lide, D. R. (1995). "Handbook of Chemistry and

Physics.", CRC Press, London.

XXXI

APPENDIX

Table C-2: Time measurements for distillation of 300 1 of 1-butanol in a 10 m3

stainless steel vessel

Batch No. Time according to

Production Protocol

[min]1 15

2 25

3 35

4 20

5 30

6 15

7 20

8 25

9 20

10 25

11 25

12 25

13 35

14 30

15 20

16 25

17 30

18 20

19 20

20 25

21 30

22 20

23 25

24 25

25 25

26 25

27 20

Average 24.3

Standard Deviation 5.1

XXXII

Miscellaneous

As shown in Figure C-4, the measured values given in Table C-2 are about

normally distributed for the unit operation "reflux". This stays in contrast to the

other unit operations where no distribution type could clearly be assigned to the

measurements (see Chapter C.l in the Appendix). Since the time for the unit

operation "reflux" is exactly given in the PSP, a normal distribution is expectedfor the actual times (random deviations from one given value).

The measurements given in Table C-2 showed that the standard deviation

for this unit operation is small. Therefore, a model according to Equation (3-13)is postulated for this unit operation, with a value for the steam consumption C

of the above-mentioned 98 kW.

Time [mm]

Figure C-4: Frequency of the measured times of reflux condition (i.e., 30 min

of reflux) in a 10 m3 stainless steel vessel given in Table C-2

C.3 Investigations on the Cleaning of Vessels

For investigating the influence of a clean and dirty heating jacket to the

steam consumption of a reaction vessel, the following program was performedin a 6.3 m3 glass lined vessel:

Before Experiment 1 fill the vessel with 6 m3 of water

Start the distillation system and the cooling water

Heat with an outside temperature of 130 °C to reflux

Hold at reflux for one hour

Distil 0.5 m3 of water

Cool to room temperatureRefill the vessel to a content of 6 m3 of water

This program was performed and measured three times for a dirty reaction

vessel (about one year without cleaning) and three times for the same vessel af¬

ter intense cleaning. The measurements are depicted in Figure C-5 and Figure

XXXIII

APPENDIX

C-6 and the values are given in Table C-3. It can be seen that the clean vessel

uses about 10% less steam and about 4% less batch time than the dirty vessel.

This may be explained by the better heat transfer to the inside of the vessel and

the lower batch time, which reduces directly the losses according to Equa¬tion (3-7).

The changes on the outside heat transfer have to be investigated by model¬

ling and will be discussed below.

15 07 2003 15 07 2003 15 07 2003 16 07 2003 16 07 2003 16 07 2003 16 07 2003 17 07 2003

06 00 12 00 18 00 00 00 06 00 12 00 18 00 00 00

| IT OT Steam Consumption

Figure C-5: Measurements for a dirty 6.3 m3 glass-lined reactor

Date & Time

-OT Steam Consumption

Figure C-6: Measurements for the same clean 6.3 m glass-lined reactor

XXXIV

Miscellaneous

Table C-3: Steam measurements for the cleaning investigations for a 6.3 m3

glass lined reactor

Experiment Clean

or

Dirty

Steam Consumption [kg] Experiment Time [hh:mm]No. Measure¬

ments

Average Standard

Deviation

Time Average Standard

Deviation

1

2

3

Dirty

2331.5

2085.9

2076.2

2164.5 144.7

07:35

07:15

07:05

07:18 00:15

4

5

6

Clean

1831.9

2040.5

1962.7

1945 105.4

07:18

06:45

06:55

06:59 00:16

The modelling of the steam consumption of the vessel was performed ac¬

cording to the equation discussed in Chapter 5.1. The results of the modellingare presented in Figure C-7 and Figure C-8. It can be seen, that the model is

applicable to both campaigns. The clean vessel has significant lower steam

consumption than the dirty one. Nevertheless, the losses are significant for both

vessels. The smaller steam consumption is mainly based on the lower batch

time (i.e., faster heating up of the reaction mass because of better heat transfer

to the inside).Considering the loss coefficient found in the modelling of the experiments

(i.e., K in Equation (3-7)), it was found, that the dirty vessel had a loss coeffi¬

cient of about 4.2-10"2 kW/m2/K and the cleaned vessel had one of about

3.7-10"2 kW / m2 / K. This difference is considered as being mainly due to the

cleaning of steam trap and the steam entrance. By cleaning these two equip¬ment parts, the waste of steam through the pipes is minimized, resulting in a

smaller heat transfer coefficient.

Measured Steam Consumption EplRvst [kg/batch]

O Dirty D Clean]

Figure C-7: Modelling and measured values of the dirty and clean 6.3 m3 glass-lined reactor

xxxv

APPENDIX

2 500

1500

t

m500

XX XgH

08 00

-- 07 00

03

0500 a

04 00 i

^ M j||j_;

03 00

5 6

Clean

Batch No

[Heating up of Water Heating up of Apparatus a Evaporation & Reflux DLosses xBatchTrne —Steam Measurements |

Figure C-8: Modelling results of the dirty and clean 6.3 m3 glass-lined reactor

(in comparison with measured values and experiment duration)

It can be seen from the modelling that the cleaning of the vessels may sig¬nificantly improve heat transfer and decrease batch times. This could result in

significant savings not only in steam consumption but also in batch time result¬

ing in increased capacity of the whole plant. Therefore, the periodic cleaning of

the vessels is useful. An introduction to all vessels available in the productionplant is considered as being favourable.

xxxvi

Measurements

D Measurements

D.l Measurements for the TODOMO

The performed measurements for the different buildings discussed in Chap¬ter 4 are summarised in the tables presented in this chapter.

Table D-l: Measurements of Building 1

Month Steam Electricity Brine Products

Production Heating

[t/month] [t/month] [MWh/month] [MWh/month] [t/month]Jan 1998 559 320 153 17 264

Feb 1998 999 441 218 50 188

Mar 1998 1040 354 232 48 349

Apr 1998 1211 255 256 72 326

May 1998 830 68 181 36 228

Jun 1998 1089 4 209 61 324

Jul 1998 1007 0 209 84 438

Aug 1998 911 0 181 44 261

Sep 1998 1004 11 248 58 229

Oct 1998 1199 103 254 71 282

Nov 1998 1002 232 268 58 464

Dec 1998 1383 425 286 59 297

Jan 1999 712 411 166 27 206

Feb 1999 874 464 239 70 369

Mar 1999 1515 334 278 56 429

Apr 1999 1331 232 280 56 355

May 1999 677 48 178 24 259

Jun 1999 920 11 219 31 391

Jul 1999 1076 0 268 63 343

Aug 1999 964 0 245 56 484

Sep 1999 1141 0 286 57 523

Oct 1999 935 82 248 39 400

Nov 1999 1112 180 282 60 388

Dec 1999 1285 329 229 54 296

Jan 2000 726 471 204 35 261

Feb 2000 1323 302 250 56 326

Mar 2000 1612 261 303 57 382

Apr 2000 1297 171 233 47 486

May 2000 1234 56 266 55 502

Jun 2000 903 28 203 29 230

Jul 2000 663 0 213 31 282

Aug 2000 724 0 197 48 524

Sep 2000 1264 0 211 39 384

Oct 2000 1303 42 269 42 306

Nov 2000 1316 142 291 40 309

Dec 2000 1060 140 186 28 290

Jan 2001 1368 350 243 37 412

Feb 2001 1349 256 212 39 407

Mar 2001 1262 281 223 31 459

Apr 2001 1086 133 181 23 455

May 2001 1339 85 244 32 224

Jun 2001 768 1 173 22 160

Jul 2001 625 0 170 17 330

Aug 2001 765 0 195 35 371

Sep 2001 1155 11 206 56 293

Oct 2001 2007 21 281 63 416

Nov 2001 1383 159 284 49 283

Dec 2001 772 241 174 52 107

Zero Production 86 - 114 14 0

XXXVII

APPENDIX

Table D-2: Measurements of Building 2


Production Heating[t/month] [t/month] [MWh/month] [MWh/month] [t/month]

Jan 1998 997 608 128 49 391

Feb 1998 1245 645 183 87 391

Mar 1998 1354 569 210 116 391

Apr 1998 1127 543 186 98 391

May 1998 1235 291 171 49 391

Jun 1998 1431 192 194 84 391

Jul 1998 1512 332 248 89 307

Aug 1998 874 201 143 37 294

Sep 1998 1908 298 238 63 394

Oct 1998 1940 158 243 82 393

Nov 1998 1609 298 236 51 368

Dec 1998 1653 584 267 86 431

Jan 1999 1153 517 201 71 422

Feb 1999 1555 670 263 94 428

Mar 1999 1590 556 270 102 548

Apr 1999 1313 501 233 62 417

May 1999 1077 260 198 59 339

Jun 1999 1540 262 263 101 616

Jul 1999 1529 253 263 92 433

Aug 1999 1208 196 205 111 503

Sep 1999 1676 121 291 123 610

Oct 1999 1606 387 256 99 593

Nov 1999 1729 479 280 95 590

Dec 1999 1550 572 265 88 418

Jan 2000 988 810 218 67 370

Feb 2000 1394 663 268 74 487

Mar 2000 1609 629 296 92 565

Apr 2000 1479 448 243 68 365

May 2000 1355 385 249 65 501

Jun 2000 1152 238 186 64 447

Jul 2000 1550 125 239 87 478

Aug 2000 916 119 175 81 365

Sep 2000 1396 129 322 82 400

Oct 2000 1568 183 275 82 555

Nov 2000 1524 402 283 83 538

Dec 2000 993 317 196 59 372

Jan 2001 1277 665 273 38 525

Feb 2001 1354 498 266 40 545

Mar 2001 1532 457 292 128 525

Apr 2001 1143 374 227 90 433

May 2001 1311 358 266 94 404

Jun 2001 1029 322 228 74 436

Jul 2001 1217 123 258 107 424

Zero Production 83 - - 0

XXXVIII

Measurements


Month Steam Electricity Ice Products

Production Heating Electricity

[t/month] [t/month] [MWh/month] [MWh/month] [t/month]Jan 1995 1647 - - - 736

Feb 1995 1254 - - - 949

Mar 1995 1349 - - - 1084

Apr 1995 1150 - - - 509

May 1995 732 - - - 745

Jun 1995 1168 - - - 809

Jul 1995 872 - - - 774

Aug 1995 793 - - - 379

Sep 1995 679 - - - 371

Oct 1995 816 - - - 505

Nov 1995 1336 - - - 746

Dec 1995 1105 - - - 540

Jan 1996 1221 - - - 856

Feb 1996 1689 - - - 933

Mar 1996 1823 - - - 1106

Apr 1996 1245 - - - 1102

May 1996 1480 - - - 1069

Jun 1996 1130 - - - 999

Jul 1996 1105 - - - 1089

Aug 1996 1276 - - - 1151

Sep 1996 1136 - - - 1097

Oct 1996 1460 - - - 1285

Nov 1996 1344 - - - 864

Dec 1996 1244 - - - 518

Jan 1997 2189 - - - 1120

Feb 1997 1411 - - - 1306

Mar 1997 1294 - - - 1028

Apr 1997 1206 - - - 1224

May 1997 1492 - - - 1378

Jun 1997 1122 - - - 1206

Jul 1997 1221 - - - 1291

Aug 1997 1403 - - - 1434

Sep 1997 1202 - - - 1232

Oct 1997 1433 - - - 1239

Nov 1997 1484 - - - 1294

Dec 1997 1017 - - - 410

Jan 1998 1923 - 547 164 1097

Feb 1998 1172 - 704 160 426

Mar 1998 1408 - 516 103 861

Apr 1998 1447 - 663 187 854

May 1998 1069 - 520 198 1298

Jun 1998 1131 - 823 273 992

Jul 1998 1423 - 827 316 1181

Aug 1998 1214 - 746 284 1193

Sep 1998 1235 - 685 264 1104

Oct 1998 1774 - 781 293 1166

Nov 1998 1579 - 666 210 911

Dec 1998 1073 - 616 169 402

Zero Production 706 - 277 0 0

XXXIX

APPENDIX



Production Heating

[t/month] [t/month] [MWh/month] [MWh/month] [t/month]Jan 1998 170 346 55 - 124

Feb 1998 251 623 80 - 151

Mar 1998 298 340 92 - 250

Apr 1998 233 397 74 - 166

May 1998 199 346 72 - 112

Jun 1998 144 319 77 - 158

Jul 1998 230 343 123 - 326

Aug 1998 186 252 83 - 194

Sep 1998 199 274 87 - 306

Oct 1998 178 260 92 - 238

Nov 1998 167 297 87 - 158

Dec 1998 139 417 87 - 172

Jan 1999 71 359 64 - 64

Feb 1999 120 406 79 - 136

Mar 1999 110 365 88 - 282

Apr 1999 73 230 73 - 149

May 1999 86 136 74 - 295

Jun 1999 154 116 95 - 207

Jul 1999 139 110 98 - 280

Aug 1999 102 31 82 - 157

Sep 1999 196 46 101 - 269

Oct 1999 126 172 86 - 197

Nov 1999 78 319 76 - 137

Dec 1999 80 391 51 - 115

Jan 2000 51 436 69 - 138

Feb 2000 124 357 76 - 192

Mar 2000 124 326 84 - 194

Apr 2000 145 207 74 - 318

May 2000 128 84 71 - 183

Jun 2000 23 26 54 - 56

Jul 2000 62 7 52 - 94

Aug 2000 61 7 50 - 65

Sep 2000 81 19 53 - 70

Oct 2000 121 78 66 - 152

Nov 2000 120 222 68 - 148

Dec 2000 60 220 40 - 145

Jan 2001 53 402 57 - 108

Feb 2001 86 306 58 - 105

Mar 2001 101 330 80 - 184

Apr 2001 52 173 54 - 78

May 2001 48 147 62 - 95

Jun 2001 31 37 58 - 49

Jul 2001 61 48 77 - 95

Aug 2001 53 15 66 - 105

Sep 2001 35 44 65 - 50

Oct 2001 61 85 66 - 194

Nov 2001 59 219 67 - 58

Dec 2001 111 306 80 - 133

XL

Measurements



Production Heating

[t/month] [t/month] [MWh/month] [MWh/month] [t/month]Jan 2000 5717 - 357 102 2379

Feb 2000 3940 - 393 117 2525

Mar 2000 6720 - 441 138 3383

Apr 2000 6908 - 446 146 3425

May 2000 5722 - 414 129 2908

Jun 2000 5022 - 395 116 2989

Jul 2000 5742 - 403 130 3266

Aug 2000 6523 - 440 171 3490

Sep 2000 4499 - 287 100 2305

Oct 2000 6898 - 399 152 3445

Nov 2000 5518 - 317 109 3164

Dec 2000 2569 - 154 49 1434

Jan 2001 5035 - 334 107 2726

Feb 2001 5920 - 370 121 2717

Mar 2001 6888 - 396 119 2895

Apr 2001 4433 - 278 82 2323

May 2001 6730 - 420 133 3583

Jun 2001 6383 - 434 148 2973

Jul 2001 5894 - 412 155 3472

Aug 2001 5777 - 381 134 2948

Sep 2001 5644 - 457 115 2993

Oct 2001 3465 - 347 85 2249

Nov 2001 5511 - 429 128 2937

Dec 2001 1367 - 175 24 1061

XLI

APPENDIX



Production Heating

[t/month] [t/month] [MWh/month] [MWh/month] [t/month]Jan 2000 829 - 212 23 1678

Feb 2000 857 - 215 23 1609

Mar 2000 887 - 219 25 1960

Apr 2000 901 - 225 25 1947

May 2000 873 - 223 28 2076

Jun 2000 938 - 227 29 1888

Jul 2000 1004 - 231 29 1926

Aug 2000 1117 - 233 31 2158

Sep 2000 761 - 187 20 1307

Oct 2000 1181 - 241 30 1868

Nov 2000 1069 - 235 28 1948

Dec 2000 702 - 175 15 993

Jan 2001 1045 - 239 28 1908

Feb 2001 968 - 219 25 1614

Mar 2001 930 - 219 25 1694

Apr 2001 1000 - 196 23 1632

May 2001 1062 - 247 31 2348

Jun 2001 1023 - 226 28 1614

Jul 2001 1043 - 231 30 1908

Aug 2001 1047 - 240 32 2379

Sep 2001 971 - 214 28 1863

Oct 2001 596 - 169 17 1147

Nov 2001 762 - 185 21 1525

Dec 2001 653 - 164 12 860

XLII

Measurements

D.2 Measurements for the BOTUMO

The performed measurements in the investigated building for the BOTUMO

are summarized in the tables provided in this chapter.

;

^~^d'^<rd' s

_.«—^ß- ''

y

^^"^/'' '

_s

y

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

Generic Number

Nominal Power 75% Nominal Power - - - 50% Nominal Power — 25% Nominal Power

Figure D-l: Efficiency of standard motors at different levels of power con¬

sumption (BBC 1976)

XLIII

X r <

2 o 3vo

ô

VO

VO

VO

VO

VO

00

00

00

00

00

00

00

00

00

00

^1

^1

^1

^1

^1

^1

^1

^1

^1

^1

On

ON

(^

-P*.

LO

to

i—'o

VO

00

^1

On

Ltl

-P*

LO

to

i—'O

VO

00

^1

On

Ltl

-P*

LO

to

1—'o

VO3

OOOooOoooOOOOOoOooOoOoooOooO

XXXXXXXXXXXXXXXXXXXXXXXXXXXX

O 3


<T>

!-|

^1 X X 3

vo

vo

VO

VO

VO

VO

VO

00

00

00

00

00

00

00

00

00

00

^1

^1

^1

^1

^1

^1

^1

^1

^1

^1

Ono

On

(^

LO

LO

to

i—'o

VO

00

^1

On

Ltl

-P*

LO

to

i—'o

VO

00

^1

On

Ltl

-P*

LO

to

1—'o

VO3

OO

(ÔoOoooOOOOOoOooooOoooOooO

S'XXXXXXXXXXXXXXXXXXXXXXXXXXXX


TfO 3 <T>

!-|

(^

O X X 3vo

vo

VO

VO

VO

00

00

00

00

00

00

00

00

00

00

00

^1

^1

^1

^1

^1

^1

^1

^1

On

On

On

Ono

4^

LO

to

1—'o

VO

^1

On

Ltl

Ltl

-P*.

-P*

LO

i—'oo

VO

00

On

Ltl

-P*

LO

to

1—'

VO

00

^1

On3

(^

(ôooooO

LfrO

LfrOOO

LfroooOoooooOoOO



TfO 3 <T>

!-|

to X X 3

\o

oo

00

00

00

00

00

^1

^1

^1

^1

^1

^1

^1

^1

^1

^1

On

On

On

On

On

On

On

On

Lfr

Lfr

Lfro

ooo

On

Ltl

-P*.

toO

VO

00

00

^1

^1

On

-P*

LO

toO

VO

VO

00

^1

On

-P*

lo

i—'

VO

00O

3ooOOOooO

Lfro

LfroOoooo

LfrOoOOOOOOoO



TfO 3 <T>

!-|

H » AT G W o_ fl>'3 o_ n>'

Vi

o

3C^

Q3

wg-

°s

On

fcr:

Ctl

3 Vi

O O O o 3

W X

Measurements

Table D-8: Measurements of the steam consumption of the reaction vessels

Reactor No. Batch No.

00e

•43 CD

CO CCD "

[h: min]

e0

'-^0cöCD

—-H t/i

cö t/i

s ^Ph 2

[kg]

-m

0

3T3O!-H

Ch

[kg]

13&o

-aCD

« c

Q S

[kg] £

Inside1—1

TemoeratureB %CD B

.2 S

F-* Î-H

0 g

[°C] ^

Measured^

Steam3- —'

Consumption

3

1

2

3

4

5

6

7

8

20:25

26:50

21:50

25:40

29:10

28:00

31:35

21:00

12102

13247

12184

10875

11529

13492

11611

12633

1268

1251

1304

1282

1201

1254

1160

1336

4008

2863

3926

5235

4581

2618

4499

3477

89

90

90

90

90

80

91

90

110

110

110

120

110

110

120

110

1880

1764

2251

2873

3090

1338

3039

2718

5

1

2

3

4

5

6

7

8

9

10

11

12

35:30

31:40

33:21

29:50

30:20

30:45

26:20

29:15

30:30

36:50

31:05

39:05

8901

9328

9479

9436

9562

9675

9783

9668

9654

9523

9630

13259

1578

2081

1944

1979

1720

2165

1985

1999

1993

1959

1972

1860

10565

10135

9480

9057

9369

10733

9227

8879

9223

10854

10084

9929

107

107

107

105

109

104

106

105

105

107

105

106

145

145

140

145

145

150

150

150

145

145

140

145

9986

9432

13790

11809

8472

10431

8937

9678

9639

9936

10762

10412

7

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

14:45

21:25

22:35

19:40

13:05

12:50

14:45

14:35

14:50

14:50

16:25

21:15

14:00

14:35

14:50

14:40

13:00

4068

4182

4032

4116

4254

4107

4022

3960

4032

4094

4450

4264

4188

4155

4005

4156

4438

4068

4182

4032

4116

4254

4107

4022

3960

4032

4094

4450

4264

4188

4155

4005

4156

4438

600

720

940

600

720

760

780

750

750

720

570

710

830

750

700

710

810

140

140

140

140

140

140

140

140

140

140

141

142

140

141

140

140

140

145

145

145

145

145

145

145

145

145

145

145

145

145

145

145

145

145

1373

1720

2135

1462

1436

1473

1375

1204

1346

1548

1313

1831

1267

1529

2062

1371

1492

APPENDIX

Table D-8 (continued): Measurements of the steam consumption of the reac¬

tion vessels

Reactor No. Batch No.

Ù0e

•43 CD

cö CCD <h

[h: min]

e0

'-^0cöCD

ri—H t/i

cö t/i

C cö

[kg]

-m

0

3T3O!-H

Ch

[kg] f?

DistilledSol¬

era '-'

vent

,1

*ëCD

^~

H ÇDCD 3

[°C] fi

Outside1—1

Temoerature^

Measured^

Steam3- —'

Consumption

g52

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

12:05

14:25

14:25

14:30

14:05

14:10

14:30

14:30

21:30

13:40

22:00

13:30

14:05

14:50

13:20

14:30

13:35

15:05

13:20

15:25

15:20

16:30

15:00

14:55

14:55

5458

4875

4633

4874

4863

4906

4873

4881

4886

4877

4570

4554

4900

4780

4682

4744

4636

4545

5082

4589

4657

4540

4622

4422

4654

3400

2860

2808

2900

2808

2840

2808

2808

2808

2840

2840

2800

3000

3000

2840

3000

2808

2800

3250

2840

2808

2840

2808

2840

2840

2058

2015

1825

1974

2055

2066

2065

2073

2078

2037

1730

1754

1900

1780

1842

1744

1828

1745

1832

1749

1849

1700

1814

1582

1814

165

165

165

165

165

165

165

165

165

165

165

165

165

165

165

165

165

165

165

165

165

165

165

165

165

178

178

178

178

178

178

178

178

178

178

178

178

178

178

178

178

178

178

178

178

178

178

178

178

178

6261

6847

7413

7316

7267

7778

8013

7540

9116

8144

11066

8066

9474

8789

8392

8011

8975

9016

7881

8735

7221

9255

8219

8068

8780

g53

1

2

3

4

5

6

08:45

07:45

07:50

07:30

09:00

08:30

4842

4921

5937

5038

5115

5115

3000

3000

4000

3200

3200

3200

1842

1921

1937

1838

1915

1915

165

165

165

165

165

165

178

178

178

178

178

178

2251

1576

1457

1545

2027

2380

With simultaneous heating and coolingWithout simultaneous heating and cooling

XLVI

X r 3

Reactor

O

No.

Batch

-P*

LO

toMo

VO

00

^1

On

(^

-P*

LO

to

No.

3-

Heat

ing

LO

LO

LO

LO

LO

LO

LO

to

LO

LO

LO

LO

LO

LO

3LO

tO

-P*

-P*

to

LO

-P*

-P*

(^

KJO

-P*

-P*

-P*

Time

(ÔOo

(^

(Ôo

(^

(ôO

(^

(^^

Lfr

Lfr

Lfr

(^

(^

-P*

^1

(^

Lfr

(^

(^

Lfr

(^

Lfr

(^

^FinalReaction

-P*

(^

LO

Lfr

(^

LO

-p*

~-\

-P*

(^

Lfr

Jj.

Jj.

to

:iq

to

toÔ

to

00

Ltl

On

-P*

Ltl

•-i

Jj.

to

00

Mass

-P*

-P*

-P*

-P*

-P*

-P*

-P*

-P*

-P*

-P*

-P*

-P*

-P*

-P*

Product

OoOo

toO

toOOO

KJo

KJo

?r

VO

VO

VO

VO

too

(^

VO

VO

VO

00

VO

00

VO

:iq

-P*

-P*

-p*

-p*

î

Ltl

-P*

-P*

-p*

-p*

00

Jj.

U)

Jj.

oooooooOoooooo

SEL

DistilledSolvent

-o

-o

^1

^1

^1

^1

î

^1

^1

^1

î

^1

î

^1

o

OInsideTem-

LO

LO

1>J

1>J

(>J

1>J

i>j

1>J

1>J

1>J

i*j

1>J

i>j

1>J

oerature(max")

VO

VO

VO

VO

VO

VO

VO

VO

VO

VO

VO

VO

VO

VO

o

oOutside

(^

(^

(/l

(/l

(/l

(/l

(/l

(/l

(/l

(/l

(/l

(/l

(/l

(/l

Temoerature

IMeasured

LO

tO

1—1

1—1

LO

LO

-P*

LO

-P*

Lfr

-P*

LO

LO

LO

Steam

-OO

(^

VO

00

VOo

00

1—1

^1

Ltl

^1

VO

^1

00o

(^

(^

to

-p*

-p*

^1

VO

LO

to

^1^r

Consumption

H » tt G 00

n © S 5' c n a pa

3§

<B

c»n

K-

r?

vi

3*

pa 3 o o 3 Vi

C 3 K 5' 3 O r+

3-

<T>

paO

X r KH

^JÖ

o^

^1

oLO

to

•faO r+

O

^td

-O

On

(^

4j

LO

toM

00

^1

On

Lfr

4j

LO

toM

Lfr

4j

LO

toMÔ

VO

00

^1

On

(^

4j

LO

toM

PS- o 3-

tO

KJ

to^^^

to^ÔoÔoÔOOoO^ôOoO^^^^^

3-

HO

5'o

3O

LOOO

VO

00

00o

to

1—'

00

(^

i—'

Ltl

00

to

to

U)

to

U)

U)

4j

U)

VO

^1

î

00

1—'

i—'

i—'o

On

3.tO

4j

4jo

too

11(^

11LO

LOO

Lfro

4j

LO

(^

4j

LOoOoo

LOo

too

to

LOo

to

<T>

31

OOO

(^

(ôoooOoO

Ltl

LtlOO

(ÔO

(Ô

(ôOoooOO

(ô

,3,

3 OQ

W<t

?d^n

to

to

to

to

to

to

to

Lfr

(^

(^

(^

Lfr

Lfr

Lfr

4j

4j

4j

4j

(^

4j

ON

ON

ON

ON

ON

ON

ON

ON

ON

ON

ON

DQ

faS

5LO

LO

LO

LO

LO

LO

LO

KJ

LO

LO

KJ

LO

to

LO

to

LO

4jO

4j

VOoOooooooOoo

Vi

SEL

00

00

00

00

00

00

00

KJ

~J

00

U)

Lti

KJ

00

00

VO

~J

KJO

KJooooooooooo

Vi

ooooooo

(^

Ltl

00

4j

Ltl

00

00

4j

On

4j

to

to

^1

to

to

to

to

to

to

to

to

to

to

to

o 3

to

to

to

to

00

00

00

VO

<\

00

00O

to

to

LO

LO

to

LO

LO

LO

to

LO

to

LO

LO

O

OJJ

U)

00

U)

00

4j

U)

U)

to

VO

to

Ltl

~J

KJ

KJ

to

U)

On

to

VO

U)

to

00

VO

Lti

On

VO

U)

On

KJ

c oVO

(^

00

Ltl

^1

00

On

4j

^1

too

4j

Ltl

^1O

00

U)

^1

VO

4j

U)

Ltio

VOo

-i

toH

U)

4jo

okJ

hh

Lti

On

(^

On

On

On

-!j

(^

On

KJ

OdOn

On

On

00

Lfrpp

Lfr

-!j

to

to

LO

to

',

to

to

to

',

',

to

Oside

empe00

4j.

VO

00

00

VOo

to

^1

to

LO

to

LO

4j

4jO

i>J

4Ô

î

^1

OOn

00

LO

00

On

Lfr

On

a [1 a

IO o 3Mea

BrinLO

LO

LO

LO

LO

to

LO

4j

4j

4j^

4j

to

to

LO

>L.

to^^^

to

to

to

to^

>L.

LO

to

>L.^

to

3-C

nvi

LO

tO

to

too

~J

KJo

VO

U)

KJ

4j

Ltl

VOO

4j

to

KJ

4j

KJ

On

(^

On

(^

00

VO

KJ

U)

00

~J

00

3 o'3

a

LOO

U)

U)

VO

^1

^1oo

Ono

4j

(^

00

^1O

4j

4j

U)O

to

U)

Lfr

U)

îO

O &.

H AT Ö Oo c«

c 3 3 pa

O >-+)

r+

3-

2.3 pa

c»

C a 3

W X

Measurements

Table D-10: Batch times for the electric heating in a 4 m3 stainless steel reac¬

tion vessel (high temperature)

Batch No. Heating Time

[h:min]1 12:25

2 12:40

3 12:40

4 12:55

5 12:25

6 13:30

7 13:00

8 13:00

9 11:50

10 13:15

11 13:25

12 12:40

13 12:25

14 13:30

Average 12:50

Standard Deviation 00:29

Table D-11: Measurements with simultaneous heating and cooling in 10 m2

nutsche dryers

Nutsche 25; Product P Nutsche 26; Product C

No. Steam

Consump¬tion

Batch

Time

DryProduct

Solvent Steam

Consump¬tion

Batch

Time

DryProduct

Solvent

[kg] [h:min] [kg] [kg] [kg] [h:min] [kg] [kg]1 743.6 11:00 2750 710.1 2444.5 12:00 1264 684.9

2 762.5 10:50 2332 394.5 1645.9 08:10 1132 684.9

3 669.8 09:30 2463 710.1 2270.9 11:00 1122 322.3

4 835 13:00 1755 867.9 1071.5 08:20 1330 684.9

5 686.4 10:10 2207 946.8 - - - -

XLIX

APPENDIX

Table D-12: Measurements without simultaneous heating and cooling in 10 m2

nutsche dryers

Nutsche 25; Product P Nutsche 26; Product C

No. Steam

Consump¬tion

Batch

Time

DryProduct

Solvent Steam

Consump¬tion

Batch

Time

DryProduct

Solvent

[kg] [h:min] [kg] [kg] [kg] [h:min] [kg] [kg]1 682.4 10:20 2274 631.2 568.5 09:00 995 1000

2 - - - - 339.3 09:25 1324 900

3 - - - - 287.1 08:30 1169 900

4 - - - - 305.1 06:45 1337 400

5 - - - - 252.6 07:50 1223 500

6 - - - - 245.9 07:45 1152 500

Table D-13: Measured power consumption of different vacuum pumps

Vacuum Pump ID P/Pn

1

2

3

4

5

6

51%

52%

43%

47%

59%

59%

AverageStandard deviation

rel. stand, deviation

52%

6%

12%

Measurements

Table D-14: Steam and cooling water consumption of different steam-jet vac¬

uum pumps (four stages) according to (GEA.b )

Norm

Pressure

[mbar abs.]

Pressure of

the steam54

[bar]

Steam

consumption[kg/h]

Cooling Water

Consumption[m3/h]

Suction

connection

[DN]55

1

3

5

8

60

55

3.5

3.2 80

2.5

3

5

8

46

42

2.6

2.5 80

5

3

5

8

40

36

33

2.2

2.1

2.0

50

10

3

5

8

35

30

29

2.0

1.8

1.7

50

20

3

5

8

29

26

26

1.7

1.5

1.5

25

Table D-15: Summary of the Brine Measurements for the APOVAC pumps

Batch ID Total Hourly Operating

Consumption Consumption Time

[kWh] [kW] [h]1 267 33.4 8

2 117 19.5 6

3 144 24.0 6

4 227 46.0 5

5 227 37.9 6

6 211 35.2 6

7 221 31.5 7

Average 202 32 -

Standard deviation 53 9 -

rel. stand, deviation 26% 27% -

Above ambient pressure

Norm-width of the connection

LI

APPENDIX

Table D-16: Infrastructure Measurements of the investigated building

Power Consumption[kW]

Date of measurement taken 02.11.2001 13.11.2001 Oct. 2003

Air Conditioning Electricity Rooms

WorkshopBoiler

Forklift Accumulators

Ventilation & HeatingDiverse

Brine Pumps

81.4

0.0

49.0

10.1

13.5

79.9

0.0

39.1

8.4

13.6

0.1

0.6

Total 154.0 141.1 0.7

Table D-17: Measurements of the steam consumption of a batch distillation

column

Batch Batch Measured Distilled Distilled Solvent /

ID Time Steam Solvent Measured Steam

[h:min] [kgl [kgl [kg/kgl1 30:00 10273 10170 0.99

2 45:00 14304 11400 0.80

3 43:20 14525 13002 0.90

4 39:50 15463 12181 0.79

5 33:40 14414 11594 0.80

6 38:40 13803 10600 0.77

7 34:45 14107 11640 0.83

8 35:30 15242 12760 0.84

9 37:35 13715 9920 0.72

10 43:00 14729 12140 0.82

ll56 29:40 18596 9796 0.53

Distillation of another product

LII

Measurements

Table D-18: Steam measurements (15bar) for the high temperature reactor

(4 m3 stainless steel reaction vessel)

Batch no. Steam Consumption[kgl

1 505

2 521

3 528

4 514

5 520

6 542

7 518

8 531

9 512

10 545

11 554

12 520

13 528

14 519

15 526

16 526

17 574

18 526

19 525

Average 528

Standard deviation 16

LIII

Improvement Potentials for the Investigated Plant

E Improvement Potentials for the Investigated Plant

During the work in the investigated plant, several findings were made not

directly related to this thesis. Nevertheless, these findings should be stated here

for providing the industrial reader with a small checklist of different improve¬ment potentials or critical operations that have to be kept in mind during dailybusiness. The checklist is first presented shortly for giving an overview of the

different possibilities. The paragraphs following the checklist are explaining in

a more detailed way the findings mentioned in the checklist.

Challenge the infrastructure consumption on similar buildings and in¬

dustry standards

o Minimise air change rate

o Optimise heating of the building (temperature and period)o Increase shutdown periods to optimise workload

Optimise the heating/cooling-systemo Improve insulation

o Control steam traps

o Preferably use thermodynamic instead of swimming ball steam

traps

o Optimise design of heating/cooling-systemo Clean the vessels and the heating/cooling-systems periodicallyo Prevent simultaneous heating and cooling by optimising pro¬

grams and continuous controls of valve set-points

Optimise cooling of pumps to decrease water consumptionMinimise nominal power of the stirrer motors

Shutdown unused equipment (e.g., heat exchangers, vacuum pumps, se¬

lective shutdown of infrastructure equipment)

Optimise batch time

Prevent waiting times

Fill the apparatus completelyReduce solvent - process production in more concentrated solutions

Minimise the time used for each step (e.g., do not dry the product too

long in the dryers)Check all possibilities for heating to high temperatures if needed - elec¬

tricity may not be the most preferable heating media

The measurements and the model results showed that infrastructure contrib¬

uted significantly to total consumption. The infrastructure may therefore not be

taken as unchangeable and challenged from time to time. Optimisation poten¬tials could be found everywhere. Whether the high air change rate, the heatingto high temperatures while leaving windows and doors open during winter times

or the continuous operation of all waste air treatment facilities are required for

operation may be questioned from time to time. The BOTUMO delivers a pos¬

sibility to compare different plants with each other on the same basis. Gener¬

ally, it can be stated that low capacity usage of a plant is a big drawback for the

specific energy consumption since most infrastructure equipment needs to run

LV

APPENDIX

anyway. Considerations of longer shutdown periods of a plant could therefore

be advantageous for capacity usage and specific energy consumption (as stated

by (Bieler et al. 2003)).The losses of the heating/cooling-systems and of the batch apparatus are sig¬

nificant. These losses cannot be avoided completely, but the measurements

show that significant variations between the loss coefficients exist. The loss

coefficients of steam and brine are dependent on the transfer of heat to and from

the environment. These losses could be minimised by better insulation of the

equipment. Especially, the top of the batch vessels should be insulated as well,

despite flexibility aspects. Moreover, the steam system looses steam throughthe steam traps. First, these steam traps could be defect. Through defect steam

traps, the systems loose continuously steam to the condensate system (or the

waste water system). Without periodic measurements, defect steam traps are

seldom found during normal operation of the plant. A periodic check of all

steam traps is therefore preferably (see e.g., (Chari 2001; Viola and Holt 2001)).Second, each kg of steam that enters the system, requires another kg of hot wa¬

ter to leave the system, as mentioned in Chapter 5.1. To prevent cavitation, the

steam is introduced to the system at the top, flows with the hot water through an

expansion vessel and is pumped at the bottom of the vessel into the heating

jacket by the circulation pump as depicted in Figure 5-1. Therefore, a signifi¬cant part of the hot water leaves the system at the hottest (highest) point without

ever reaching the heating jacket. Investigations should be performed, whether it

would be possible to introduce the steam to the system right after the circulation

pump (before the circulating water is entering the heating jacket). Temperaturecontrol would probably turn out to be more complicated, but cavitation should

not occur. This would enable the steam to provide his full heat content to the

reaction mass. Moreover, the switch to systems where the steam is condensingin the heating jacket is preferably from an energetic point of view. Material

stress is, nevertheless, greater in such a system with known drawbacks on

equipment lifetime.

As may be seen from the measurements presented in Chapter C.3 in the Ap¬

pendix, the cleaning of an apparatus has a significant positive influence on its

energy consumption. Therefore, periodic cleaning of the equipment should be

performed. With the help of the constant conditions for heat transfer, the con¬

trol of the process is easier as well and production time and product qualitycould be improved.

Many of the pumps operated in a chemical plant are cooled by cooling water

to prevent overheating of the equipment. Discussions with experts revealed,

nevertheless, that cooling is only required if process temperature is above about

100 °C (depending on the kind of pump). Since such a pumps requires about

several hundred litres of cooling water per hour, significant amounts of waste¬

water could be saved by optimising the cooling of pumps.

As seen in Chapter 5.5, the power consumption of the stirrer motors is most

of the time far below the nominal power. In this region of operation, the motors

have poor efficiency and waste a lot of energy. Large motors are installed, to

avoid the possibility that the reaction mass is too heavy for the stirrer to stir and

the stirrer would be stuck therefore in the vessel. If operation with difficult

LVI

Improvement Potentials for the Investigated Plant

products with extremely high viscosity could be excluded (only few products

possess such properties), smaller motors with higher efficiencies could be used

to save energy.

The heating/cooling-systems are considered reacting slowly to changes in

temperatures. Nevertheless, the production control programs sometimes try to

prevent temperature swings at the beginning of operation or try to keep the ex¬

act temperature over a longer period. This often requires fast and frequent

changes between steam and water input. This simply 'destroys' energy since

the steam introduced to the system is cooled down immediately by new coolingwater introduced to the system shortly afterwards and vice versa. Longer lag-times in the programs and control sequences (PID-controllers) could improve or

eliminate this systematic loss of energy.

A similar finding is the possibility offered by some heating/cooling-systemto simultaneously heat and cool. This is possible when a system is equippedwith an indirect heat exchanger (for cooling at high temperatures). If the valves

are adjusted wrongly (detected at one reaction vessel during the investigationsof this thesis), the steam consumption is increased significantly. Since the op¬

erator has no direct possibility to check whether the valves are working cor¬

rectly, this state of operation could continue for quite a long period. A periodiccontrol of the programs and the valves (sealing and set-points) could reveal such

an operation and offer the possibility to improve steam and water consumption

significantly.Heat exchangers should be checked for correct operation often. It often

happens that a heat exchanger is not closed or shut down completely after an

apparatus was used. Cooling water or brine is therefore continuously flowing

through the heat exchanger. This continuous flow, although probably not large,could contribute significantly to the consumption of brine or cooling water.

As seen from the sensitivity analyses in Chapter 6.3.1, batch time has a sig¬nificant influence on utility consumption. Batch time should therefore be mini¬

mised. This would not only improve the specific energy consumption of a sin¬

gle batch but also improve the specific infrastructure consumption since the

same infrastructure consumption is divided by more production output. If wait¬

ing time for another apparatus is inevitable, the reactor should be operated at

temperature as close as possible to room temperature without any vacuum or

other equipment running and without stirring (always under the assumption that

this is not affecting the safety of the process step).For energy reasons, the apparatus should be filled as completely as possible.

This would improve the specific infrastructure consumption and the product de¬

pendent consumption (e.g., the apparatus has to be heated once, unimportanthow much product it contains).

Generally, unused equipment should be shut down. For some equipment,this could nevertheless not be favourably, since shutdown periods affect equip¬ment lifetimes of some apparatus negatively. Therefore, this has to be investi¬

gated for each apparatus separately.Care should be taken not to overheat the production plant. An overheated

plant is uncomfortably for the workmen. Therefore, doors and windows will be

opened to bring the temperature down and energy is lost to the environment.

LVII

APPENDIX

The time required for a production step should be minimised not only for

energy optimisation but for optimisation of the production schedule and capac¬

ity reasons as well. This means, for drying processes as example that care

should be taken not to dry a product too much. If this does not negatively affect

product quality, it is probably not considered as important but it increases en¬

ergy consumption significantly and reduces capacity of the drying plant.If processes with high-temperature operations and secondary heat¬

ing/cooling-system are considered, all possibilities should be checked. It is not

only possible to heat such equipment with electricity; high-pressure steam and

natural gas burner are just two other possibilities that could be economically fa¬

vourable and better from an energetic point of view. Moreover, the insulation

of an apparatus should be adapted according to its usual temperature level. The

higher the temperature difference to the environment is in usual operation mode,the better should the insulation be.

LVIII

Glossary

F Glossary

A small glossary of the most important terms in batch production is pre¬

sented in the following table.

Table F-1: Definitions of the ISA-S88.01-1995 standard for batch production

(ISA 1995)

Nomenclature ExplanationArea

Batch

Batch process

Batch schedule

Control recipe

Enterprise

Equipment entity

Equipment mod¬

ule

Equipment op¬

eration

Equipment phase

Equipment pro¬

cedure

Equipment unit

procedureFormula

General recipe

Master recipe

Mode

Operation

Phase

Procedure

Process

Process action

A component of a batch manufacturing site that is identified by physical,

geographical, or logical segmentation within the site

1. The material that is being produced or that has been produced by a

single execution of a batch process

2. An entity that represents the production of a material at any point in

the process

A process that leads to the production of finite quantities of material by

subjecting quantities of input materials to an ordered set of processing ac¬

tivities over a finite period of time using one or more pieces of equipmentA list of batches to be produced in a specific process cell

A type of recipe which, through its execution, defines the manufacture of a

single specific productAn organisation that co-ordinates the operation of one or more sites

A collection of physical processing and control equipment and equipmentcontrol grouped together to perform a certain control function or set of con¬

trol functions

A functional group of equipment that can carry out a finite number of spe¬

cific minor processing activities

An operation that is part of equipment control

A phase that is part of equipment control

A procedure that is part of equipment control

A unit procedure that is part of equipment control

A category of recipe information that includes process inputs, process pa¬

rameters, and process outputsA type of recipe that expresses equipment and site independent processing

requirementsA type of recipe that accounts for equipment capabilities and may include

process cell-specific information

The manner in which the transition of sequential functions are carried out

within a procedural element or the accessibility for manipulating the states

of equipment entities manually or by other types of control

A procedural element defining an independent processing activity consist¬

ing of the algorithm necessary for the initiation, organization, and control of

phasesThe lowest level of procedural element in the procedural control model

The strategy for carrying out a process

A sequence of chemical, physical, or biological activities for the conver¬

sion, transport, or storage of material or energy

Minor processing activities that are combined to make up a process opera¬

tion

LIX

APPENDIX

Table F-1 (continued): Definitions of the ISA-S88.01-1995 standard for batch

production

Nomenclature ExplanationProcess cell

Process input

Process opera¬

tion

Process output

Process parame¬

ter

Process stage

Recipe

Recipe opera¬

tion

Recipe phase

Recipe proce¬

dure

Recipe unit pro¬

cedure

Shared-use re¬

source

Site

Site recipeState

Train; Line

Unit

Unit procedure

Unit recipe

A logical grouping of equipment that includes the equipment required for

production of one or more batches. It defines the span of logical control of

one set of process equipment within an area

The identification and quantity of a raw material or other resource requiredto make a productA major processing activity that usually results in a chemical or physical

change in the material being processed and that is defined without considera¬

tion of the actual target equipment configurationAn identification and quantity of material or energy expected to result from

one execution of a control recipeInformation that is needed to manufacture a material but does not fall into

the classification of process input or outputA part of a process that usually operates independently from other process

stages and that usually results in a planned sequence of chemical or physical

changes in the material beingThe necessary set of information that uniquely defines the production re¬

quirements for a specific productAn operation that is part of a recipe procedure in a master or control recipe

A phase that is part of a recipe procedure in a master or control recipeThe part of a recipe that defines the strategy for producing a batch

A unit procedure that is part of a recipe procedure in a master or control rec¬

ipeA common resource that can be used by more than one user at a time

A component of a batch manufacturing enterprise that is identified by physi¬cal, geographical, or logical segmentation within the enterpriseA type of recipe that is site specificThe condition of an equipment entity or of a procedural element at a giventime

A collection of one or more units and associated lower level equipment

groupings that has the ability to be used to make a batch of material

A collection of associated control modules and/or equipment modules and

other process equipment in which one or more major processing activities

can be conducted

A strategy for carrying out a contiguous process within a unit. It consists of

contiguous operations and the algorithm necessary for the initiation,

organization, and control of those operationsThe part of a control recipe that uniquely defines the contiguous production

requirements for a unit

LX

Literature

Literature

ACS. (1996). "Technology Vision 2020 - The U.S. Chemical Industry.", Ameri¬

can Chemical Society, American Institute of Chemical Engineers, Chemical

Manufacturers Association, Council for Chemical Research, Synthetic Or¬

ganic Chemical Manufacturers Association, Washington.Aebischer, B., Kanala, R., Pain, D., and Giovannini, B. (1988). "Perspecitves de

la demande d'énergie en Suisse, 1985-2025." 805.818, Centre Universitaire

d'étude des Problèmes de l'Energie (CUEPE), Université de Genève,Genève.

Aggarwal, S. (2002). "Boost Energy Efficiency in Plant Utilities." Chem Eng,

109(4), 70-73.

Allen, D. T., and Shonnard, D. R. (2002). Green Engineering: EnvironmentallyConscious Design ofChemical Processes, Prentice Hall PTR, Upper Saddle

River.

de Almeida, A. T. (1997). "Energy Efficient Technologies: Electric Motor Sys¬tems Efficiency." CRC Handbook of Energy Efficiency, F. Kreith and R. E.

West, eds., CRC Press, New York, 379-397.

Anastas, P. T., and Warner, J. C. (1998). Green Chemistry: Theory and Prac¬

tice, Oxford University Press, Oxford.

Andreas, J. C. (1992). Energy-Efficient Electric Motors, Marcel Dekker, Inc.,New York.

Anlauf, H. (2003). "Zentrifugen - Achema-Berichte." Chem Ing Tech, 75(10),1545-1551.

Anonymous. (1986a). "Batch Reactors can also be Integrated." Energy Man¬

agementFocus(9), 24.

Anonymous. (1986b). "Where Energy is used." Energy Management Focus(9),4-6.

Anonymous. (2001). "Facts & Figures for the Chemical Industry." Chem. Eng.News, 79(26).

Arellano-Garcia, H., Martini, W., Wendt, M., Li, P., and Wozny, G. "Improvingthe Efficiency of Batch Distillation by a New Operation Mode." European

Symposium on Computer AidedProcess Engineering - 12, 619-624.

Ashton, G. (1993). "Design of Energy Efficient Batch Processes." Energy Effi¬

ciency in Process Technology, P. A. Pilavachi, ed., Elsevier Applied Sci¬

ence, London, 1050-1062.

Atkins, P. W. (1990). Physikalische Chemie, VCH, Weinheim.

Aziz, N., and Mujtaba, I. M. (2002). "Optimal Operation Policies in Batch Re¬

actors." Chem. Eng. J., 85, 313-325.

Baier, H. (1989). "Vakuumerzeugung durch Dampfstrahl-Vakuumpumpen."Vakuum in der Praxis(2), 127-135.

Barolo, M. (2000). "Batch Distillation." Encyclopedia of Separation Science, I.

D. Wilson and M. Cooke, eds., Academic Press, San Diego, 995-1004.

BBC. (1976). "Auswahl und Einsatz von Elektromotoren - Planungsunterlage."D EM 60001D, Brown Boveri Company, Mannheim.

LXI

APPENDIX

Benz, G. T. (2003). "Optimize Power Consumption in Aerobic Fermenters."

Chem Eng Prog, 99(5), 32-35.

Bertrand, J., Couderc, J. P., and Angelino, H. (1980). "Power Consumption,

Pumping Capacity and Turbulence Intensity in Baffled Stirred Tanks: Com¬

parison Between Several Turbines." Chem Eng Sei, 35, 2157-2163.

BFE. (2001a). "Anhang zur Richtlinie über freiwillige Massnahmen zur Reduk¬

tion von Energieverbrauch und C02-Emissionen." 003693825, Bundesamt

für Energie, Bern.

BFE. (2001b). "Richtlinie über freiwillige Massnahmen zur Reduktion von En¬

ergieverbrauch und C02-Emissionen.", Bundesamt für Energie, Bern.

BFE. (2002). "Energie- und C02-Management schafft Unternehmenswerte."

805.008 d, Bundesamt für Energie, Bern.

Bhatt, M. S. (2000a). "Energy Audit Case Studies I - Steam Systems." Appl.Therm. Eng, 20,285-296.

Bhatt, M. S. (2000b). "Energy Audit Case Studies II - Air Conditioning (Cool¬

ing) Systems." Appl. Therm. Eng., 20, 297-307.

Bieler, P. S., Fischer, U., and Hungerbühler, K. (2003). "Modelling the Energy

Consumption of Chemical Batch Plants - Top-Down Approach." Ind. Eng.Chem. Res, 42, 6135-6144.

Blickenstorfer, C. (1999). "Analyse des Energieverbrauchs eines Mehrprodukte-Batch-Betriebes," Ph.D. dissertation, No. 13411, Zurich, ETH,

http://e-collection.ethbib.ethz.ch/cgi-bin/show.pl?type=diss&nr=13411.Bohne, D., Fischer, S., and Obermeier, E. (1984). "Thermal Conductivity, Den¬

sity, Viscosity, and Prandtl-Numbers of Ethylene Glycol-Water Mixtures."

Ber. Bunsenges. Phys. Chem., 88, 739-742.

Bouhenchir, H., Cabassud, M., Le-Lann, M. V., and Casamatta, G. (2001). "A

General Simulation Model and a Heating/Cooling Strategy to Improve Con¬

trollability of Batch Reactions." Trans IChemEPartA, 79, 641-654.

Bundesamt für Konjunkturfragen, B. (1992). "Strom rationell nutzen.", VDF,Zürich.

Burkhardt, P. G. (2002). "Windenergie, C02-Abgabe und Erfolgskontrolle."NZZ, Zurich, 17.

Burton, C. J. (1948). "A Study of Ultrasonic Velocity and Absorption in LiquidMixtures." JAcoust Soc Am, 20(2), 186-199.

Calderôn, Z., Grau, R., Espuria, A., and Puigjaner, L. (2000). "Scheduling of

Multiproduct Plants Considering Energy and Waste Costs." Hung. J. Ind.

Chem., 28, 31-35.

Carpenter, K. J. (2001). "Chemical Reaction Engineering Aspects of Fine

Chemical Manufacture." Chem Eng Sei, 56(2), 305-322.

Charalambides, M. S., Shah, N., and Pantelides, C. C. (1995). "Synthesis of

Batch Reaction / Distillation Processes using Detailed Dynamic Models."

Comput. Chem. Eng, 19, S167-S174.

Chari, S. R. (2001). "Take Another Look at Steam Traps." Chem Eng Prog,

97(6), 62-63.

Churchman, C, and Krueger, R. (2003). "Evaluating Chiller Performance."

Chem Eng Prog, 99(5), 54-61.

LXII

Literature

Corradini, F., Marchetti, A., Tagliazucchi, M., Tassi, L., and Tosi, G. (1995).

"Thermodynamics of Viscous Flow in Ethane-1,2-diol+Water Binary Mix¬

tures." ita. J. Chem., 48, 186-199.

Courtois, F., Lebert, A., Lasseran, J. C, and Bimbenet, J. J. "Simulation of In¬

dustrial Dryers: Solving Numerical and Computer Problems." Drying'92,Montreal, Quebec, Canada, 1268-1277.

Curme, G. O., and Johnston, F. (1953). "Glycols.", Reinhold Publishing Corp.,New York.

Dahinden, D. (2003). "Modellierung des Energieverbrauches eines Chemischen

Mehrzweckbetriebes," Master Thesis, ETH, Zurich, 3-906734-34-X.

Daubert, T. E., and Dannel, R. P. (1984). "Data Compilation Tables of Proper¬ties of Pure Components.", AIChE Design Institute for Physical Property

Data, New York.

Daubert, T. E., and Dannel, R. P. (1985). "Manual for Predicting Chemical

Process Design Data.", AIChE Design Institute for Physical Property Data,New York.

Diwekar, U. M. (1996). Batch Distillation, Taylor & Francis, Washington.DOE. (2000). "Energy and Environmental Profile of the U.S. Chemical Indus¬

try.", Energetics Inc. & Department of Energy (DOE), Columbia.

Eidgenossenschaft. (1999). "Bundesgesetz über die Reduktion der CO2-

Emissionen." 641.71, Bern.

Eissen, M., Metzger, J. O., Schmidt, E., and Schneidewind, U. (2002)."10 Years after Rio - Concepts on the Contribution of Chemistry to a Sus¬

tainable Development." Angew. Chem. Int. Ed., 41, 414-436.

El-Dessouky, H., Ettouney, H., Alatiqi, I., and Al-Nuwaibit, G. (2002). "Evalua¬

tion of Steam Jet Ejectors." Chem. Eng. Process., 41, 551-561.

Eilermann, K. (1991). "Leistungsberechnung von Rührorganen." CD/S8/91,Stelzer Rührtechnik GmbH, Warburg.

Ewers, H.-J. (2000). "Alternativen zum Konzept der Bundesregierung." Gaia,

9(1), 61-62.

Fogler, S. H. (1999). Elements of Chemical Reaction Engineering, Prentice

Hall, Upper Saddle River (NJ).Francis Murray, J. "Electric Motor and other Equipment Standards in the United

States." Industrial Energy Efficiency: Policies and Programmes, Washing¬ton, DC, 267-270.

Furman, K. C, and Sahinidis, N. V. (2002). "A Critical Review and Annotated

Bibliography for Heat Exchanger Network Synthesis in the 20th Century."Ind. Eng. Chem. Res, 41, 2335-2370.

Galindez, H., and Friedenslund, A. A. (1988). "Simulation of MulticomponentBatch Distillation Processes." Comput. Chem. Eng., 12(4), 281-288.

Ganji, A. R. (1999). "Conducting an Energy Audit." Chemical Processing,

62(9), 64-70.

GEA.a. "Dampfstrahl-Vakuumpumpen.", sales brochure, GEA-Wiegand, Et¬

tlingen.GEA.b. "Porzellan/Graphit-Dampfstrahl-Vakuumpumpen.", sales brochure,

GEA-Wiegand, Ettlingen.

LXIII

APPENDIX

GEA.c. "Aus der Theorie der Dampfstrahlpumpen.", sales brochure, GEA-

Wiegand, Ettlingen.Gehrmann, D. (2003). "Trocknungstechnik." Chem Ing Tech, 75(10), 1507-

1514.

Gränicher, W. H. H. (1997). "Nationaler Energie Forschungs-Fonds NEFF

1997-97.", vdf Hochschulverlag AG, Zurich.

Grant, C. D. (1996). "Energy Management in Chemical Industry." Kirk-Othmer:

Encyclopedia of Chemical Technology, J. I. Kroschwitz, ed., John Wiley &

Sons, New York, 12-1-12-16.

Grassmann, P., Widmer, F., and Sinn, H. (1998). Einführung in die thermische

Verfahrenstechnik, Walter de Gruyter, Berlin.

Greiner, H. (1999). "Energie sparen mit der richtigen Auswahl von Antriebs-

Komponenten." IE 0699, Danfoss Bauer GmbH.

Gundersen, T. "Achievements and Future Challenges in Industrial Design Ap¬

plications of Process Systems Engineering." 4th International Symposium on

Process Systems Engineering, Montebello, Quebec, Canada, 1.1.1-1.1.32.

Gundersen, T., and Naess, L. (1988). "The Synthesis of Cost Optimal Heat Ex¬

changer Networks: An Industrial Review of the State of the Art." Comput.Chem. Eng, 12(6), 503-530.

Haman, W. G. (2000). "Total Assessment Audits (TAA) in Iowa." Resources,

Conservation andRecycling, 28(3-4), 185-198.

Hessel, G, Kryk, H., Schmitt, W., Seiler, T., Weiß, F.-P., Hilpert, R., Roth, M.,and Deerberg, G. (2002). "Monitoring-System mit adaptiven Wärmebilan¬

zen für Batch-Reaktoren." Chem Ing Tech, 74(12), 1692-1698.

Heuser, A. (2003). "Gea Wiegand: Erfolgreicher Start mit Verdampfer-Patent."Chem Ing Tech, 75(10), 1393-1395.

Hinrichs, F. (1991). "Aufbau, Betriebsverhalten und Regelbarkeit von

Dampfstrahl-Vakuumpumpen." Vakuum in der Praxis(2), 102-109.

Höller, V., and Renken, A. (2000). "Neue Konzepte für chemische Prozesse."

CAV{\), 62ff

Hoshide, R. K. (1995). "Effective Energy Audits." Energy Engineering, 92(6),6-17.

http://unfccc.int/resource/docs/convkp/kpeng.html. "Kyoto Protocol.", United

Nations Framework Convention on Climate Change.http://webbook.nist.gov/chemistry. (2003). "MST Chemistry Webbook.", MST.

http://www2.din.de/index.php?lang=en. "Homepage des Deutsches Institut für

Normung e. V.", Deutsches Institut für Normung e. V. DIN Standards.

http://www.aie.org.au/melb/material/resource/cop.htm., Australian Institute of

Energy.

http://www.eia.doe.gov/neic/infosheets/degreedays.htm. "Energy Information

Sheets - Degree Days.", National Energy Information Centre (NEIC).ISA. (1995). "Batch Control Part 1: Models and Terminology." ISA-S88.01-

1995, ISA, Raleigh, NC.

Jezowski, J. (1994a). "Heat Exchanger Network Grassroot and Retrofit Design.The Review of the State-of-the Art: Part I." Hung. J. Ind. Chem., 22, 279-

294.

LXIV

Literature

Jezowski, J. (1994b). "Heat Exchanger Network Grassroot and Retrofit Design.The Review of the State-of-the Art: Part II." Hung. J. Ind. Chem., 22, 295-

308.

Jiménez-Gonzâlez, C, and Overcash, M. R. (2000). "Energy Optimization dur¬

ing Early Drug Development and the Relationship with Environmental Bur¬

dens." J Chem Technol Biotechnol, 75, 983-990.

Kemp, I. C, and Macdonald, E. K. (1988). "Application of Pinch Technology to

Separation, Reaction and Batch Processes." Understanding Process Integra¬tion II, R. Smith, ed., The Institution of Chemical Engineers, Rugby, UK,239-259.

Kneubühl, F. K. (1994). Repetitorium der Physik, B. G. Teubner, Stuttgart.

Kordik, T. (2001). "Erprobtes Energiesparkonzept für Schraubenkompressoren -

bis zu 35 % Strom sparen." CAV(5), 46ff

Krummenacher, P. (1997). "Energetische Integration von Batch-Prozessen."

SI+A, 38(9), 757-763.

Krummenacher, P., and Favrat, D. (2001). "Indirect and Mixed Direct-Indirect

Heat Integration of Batch Processes Based on Pinch Analysis." Int. J. Ap¬

plied Thermodynamics, 4(3), 135-143.

Krummenacher, P., Zogg, M., and Geissmann, M. (2002). "Diskontinuierlich

Energie sparen." ENET-News(53), 35-36.

Kubier, K. (1986). "Grundlagen, Erfahrungen und Perspektiven der Energiemo¬

dellanwendung in der Bundesrepublik Deutschland.", ETH, Zürich.

Kuczmowski, T., and Weyant, J. (1990). "Energy-Economy Modeling: A Sur¬

vey." Energy, 15(3/4), 387-394.

Kürsten, M. (1996). "World Energy - A Changing Scene.",E. Schweizerbart'sche Verlagsbuchhandlung, Stuttgart.

Levenspiel, O. (1999). Chemical Reaction Engineering, John Wiley & Sons,New York.

Lide, D. R. (1995). "Handbook of Chemistry and Physics.", CRC Press, Lon¬

don.

Liepe, F., Sperling, R., and Jembere, S. (1998). Rührwerke, Theoretische

Grundlagen, Auslegung und Bewertung, Eigenverlag Fachhochschule

Köthen, Köthen.

Lindegger, M. (2002). "Energiesparmotor." 85 458, BFE, Bern.

Linnhoff, B. (1993). "Pinch Analysis - a State-of-the-Art Overview." Trans

IChemEPartA, 71(6), 503-522.

Linnhoff, B., Ashton, G. J., and Obeng, E. D. A. (1988). "Process Integration of

Batch Processes." Understanding Process Integration II, P. R. Crump, D. V.

Greenwood, and R. Smith, eds., Institution of Chemical Engineers, Rugby,221-237.

Linnhoff, B., Townsend, D. W., Boland, D., Hewitt, G. F., Thomas, B. E. A.,

Guy, A. R., and Marsland, R. H. (1982). A User Guide on Process Integra¬tion for the Efficient Use ofEnergy, The Institution of Chemical Engineers,Rugby, UK.

Litowitz, T. A., Higgs, R., and Meister, R. (1954). "Ultrasonic Propagation and

its Relation to Molecular Structure in the Diols." J Chem Phys, 22(8), 1281-

1283.

LXV

APPENDIX

Martin, H. (2003). "Wärme- und Stoffübertragung." Chem Ing Tech, 75(10),1472-1475.

Mersmann, A., Einenkel, W.-D., and Käppel, M. (1975). "Auslegung und Mass-

stabsvergrösserung von Rührapparaten." Chem Ing Tech, 47(23), 953-964.

Mujumdar, A. S. (1995). Handbook ofIndustrial Drying, Marcel Dekker, Inc.,New York.

Nakaiwa, M., Huang, K., Endo, A., Ohmori, T., Akiya, T., and Takamatsu, T.

(2003). "Internally Heat-Integrated Distillation Columns: A Review." Trans

IChemEPartA, 81, 162-177.

Ng, K., and Yianneskis, M. (2000). "Observations on the Distribution of Energy

Dissipation in Stirred Vessels." Trans IChemE Part A, 78(3), 334-341.

Oppenheimer, O., and Sorensen, E. (1997). "Comparative Energy Consumptionin Batch and Continuous Distillation." Comput. Chem. Eng., 21(Suppl.),529-534.

Papageorgiou, L. G., Shah, N., and Pantelides, C. C. (1994). "Optimal Schedul¬

ing of Heat-Integrated Multipurpose Plants." Ind. Eng. Chem. Res., 33(12),3168-3186.

Parakrama, R. (1985). "Improving Batch Chemical Processes." The Chemical

Engineer, 417, 24-25.

Parti, M., and Palâncz, B. (1974). "Study of Batch Drying." Int. J. Heat Mass

Transfer, 17, 669-679.

Perlmutter, B. A. "Principles of Pressure Nutsche Filter-Dryer Technology."Drying'92, Montreal, Quebec, Canada, 1321-1329.

Perry, R. H., Green, D. W., and Maloney, J. O. (1997). "Perry's Chemical Engi¬neers' Handbook.", McGraw-Hill, New York.

Phillips, C. H., Lauschke, G., and Peerhoussaini, H. (1997). "Intensification of

Batch Chemical Processes by Using Integrated Chemical Reactor Heat Ex¬

changers. "Appl. Therm. Eng., 17(8), 809-824.

Râsonyi, P. (2002). "Aufgeschobene amerikanische Klimapolitik." NZZ, Zu¬

rich, 21.

Reddy, V. K., Reddy, K. S., and Krishnaiah, A. (1994). "Excess Volumes,

Speeds of Sound, and Viscosities for Mixtures of 1,2-Ethanediol and

Alkoxy Alcohols with Water at 308.15 K." J. Chem. Eng. Data, 39, 615-

617.

Reid, R. C, Prausnitz, J. M., and Poling, B. E. (1987). The Properties of Gases

and Liquids, McGraw Hill Inc., New York.

Reklaitis, G. V., Pekny, J., and Jogelekar, G. S. (1997). "Scheduling and Simu¬

lation of Batch Processes." Handbook of Batch Process Design, P. N. Shar-

ratt, ed., Chapman & Hall, London, 24-60.

Rippin, D. W. T. (1992). "Batch Process Systems Engineering: A Retrospectiveand Prospective Review." Comput. Chem. Eng, 17(Supplement), S1-S13.

Robert, H., and Markus, W. (1994). Organisation und Energiemanagement,Bundesamt für Konjunkturfragen, Bern.

Rumazo, C. A., Pryor, A., Mendoza, F. O., Villareal, J. C, Robledo, J. M., and

Mercado, E. R. (2000). "Cleaner Production in the Chemical Industry." Wa¬

ter Science and Technology, 42(5-6), 1-7.

LXVI

Literature

Sahinidis, N. V., Grossmann, I. E., Fornari, R. E., and Chathrathi, M. (1989).

"Optimization Model for Long Range Planning in the Chemical Industry."Comput. Chem. Eng, 13(9), 1049-1063.

Schaaffs, W. (1941). "Über Schallgeschwindigkeit und Konstitution in flüssigenorganischen Verbindungen." Ann Phys, 40(6), 27-404.

Schalcher, M., Battaglia, U., and Bush, E. (2003a). "Energieverbrauch von Pro¬

zesssteuerungen (SPS)." 46 963, Bundesamt für Energie (BFE), Bern.

Schalcher, M., Battaglia, U., and Bush, E. (2003b). "Projekt Energieverbrauchvon Prozesssteuerungen." 46 963/87 079, BFE, Chur.

Seader, J. D., Siirola, J. J., and Barnicki, S. D. (1997). "Distillation." Perry'sChemical Engineers' Handbook, R. H. Perry, D. W. Green, and J. O. Ma-

loney, eds., McGraw-Hill, New York, 13.96-13.103.

Severson, D. S. "Benchmarking Process Energy Performance from Historical

Data." Eighteenth National Industrial Energy Technology Conference,Huston, Texas, 87-91.

SIA. (1992). "Kühlleistungsbedarf von Gebäuden - SIA 382/2." SIA 382/2, SIA,Zurich.

SIA. (1995). "Elektrische Energie im Hochbau - SIA 380/4." SIA 380/4, SIA,Zurich.

SIA. (1997). "Energie im Hochbau - SIA 380/1." SIA 380/1, SIA, Zurich.

Silva, J. M., Knoechelmann, A., Meirelles, A. J. A., Wolf-Maciel, M. R., and

Lopes, C. E. (2003). "On the Dynamics of Nonequilibrium Simple Batch

Distillation Processes." Chem. Eng. Process., 42, 475-485.

Spreng, D. T. (1988). Net-Energy Analysis and the Energy Requirements ofEn¬

ergy Systems, Praeger, New York.

Stahel, W. A. (1995). Statistische Datenanalyse; Eine Einführungfür Naturwis¬

senschaftler, Friedrich Vieweg & Sohn, Wiesbaden a.o.

Stankeiwicz, A. I., and Moulijn, J. A. (2000). "Process Intensification: Trans¬

forming Chemical Engineering." Chem Eng Prog, 96(1), 22-34.

Stitt, E. H. (2002). "Alternative Multiphase Reactors for Fine Chemicals - A

World Beyond Stirred Tanks?" Chem. Eng J., 90, 47-60.

Stoltze, S., Mikkelsen, J., Lorentzen, B., Petersen, P. M., and Qvale, B. (1995)."Waste-Heat Recovery in Batch Processes Using Heat Storage." Journal of

Energy Resources Technology; Transactions ofthe ASME, 117(2), 142-149.

Strumillo, C, Jones, P. L., and Zylla, R. (1995). "Energy Aspects in Drying."Handbook of Industrial Drying, A. S. Mujumdar, ed., Marcel Dekker Inc.,New York, 1241-1276.

Suhami, I, and Mah, R. S. H. (1982). "Optimal Design of Multipurpose Batch

Plants." Ind. Eng. Chem. Process Des. Dev., 21(1), 94-100.

Sulzer, M. (2003). "Gute Gebäudetechnik - oder die unendliche Uneinigkeit der

Ingenieure!", BFE.

Taca, C. D., and Paunescu, M. (2001). "Power Input in Closed Stirred Vessels."

Chem Eng Sei, 56(14), 4445-4450.

Tapp, M., Kauchali, S., Hausberger, B., McGregor, C, Hilderandt, D., and

Glasser, D. (2003). "An Experimental Simulation of Distillation Column

Concentration Profiles Using a Batch Apparatus." Chem Eng Sei, 38, 479-

486.

LXVII

APPENDIX

Thöne, E., and Fahl, U. (1998). "Energiewirtschaftliche Gesamtsituation."

Brennstoff-Wärme-Kraft, 50(4), 26-38.

Thumann, A. (1983). Handbook ofEnergy Audits, The Fairmont Press, Inc., At¬

lanta, Georgia.TRC. (1998). "TRC Thermodynamic Tables.", Thermodynamics Research Cen¬

tre, College Station, Texas.

Tsotsas, E. "Measurement and Modelling of Intraparticle Drying Kinetics: A

Review." Drying'92, Montreal, Quebec, Canada, 17-41.

Turner, W. C. (1982). "Energy Management Handbook.", John Wiley & Sons,New York.

Uhlemann, J., Garcia, V., Cabassud, M., and Casamatta, G. (1996). "Optimie¬

rungsstrategien für Batchreaktoren." Chem Ing Tech, 68(8), 917-926.

Vaklieva-Bancheva, N., Ivanov, B. B., Shah, N., and Pantelides, C. C. (1996)."Heat Exchanger Network Design for Multipurpose Batch Plants." Comput.Chem. Eng, 20(8), 989-1001.

Vaselenak, J. A., Grossmann, I. E., and Westerberg, A. W. (1986). "Heat Inte¬

gration in Batch Processing." Ind. Eng. Chem. Process Des. Dev., 25(2),357-366.

Vaselenak, J. A., Grossmann, I. E., and Westerberg, A. W. (1987). "An Embed¬

ding Formulation for the Optimal Scheduling and Design of MultipurposeBatch Plants." Ind. Eng. Chem. Res., 26(1), 139-148.

VDI. (1984). VDI-Wärmeatlas, VDI-Verlag GmbH, Düsseldorf.

Venkateswarlu, C, and Avantika, S. (2001). "Optimal State Estimation of Mul-

ticomponent Batch Distillation." Chem Eng Sei, 56, 5771-5786.

Verwater-Lukszo, Z. (1996). "A Practical Approach to Recipe Improvement and

Optimization in the Batch Processing Industry," Ph.D. dissertation, Techni¬

cal University of Eindhoven, Eindhoven.

Villermaux, J. (1995). "Future Challenges in Chemical Engineering Research."

Trans IChemE Part A, 73, 105-109.

Vin, J. P., and Ierapetritou, M. G. (2000). "A New Approach for Efficient Re¬

scheduling of Multiproduct Batch Plants." Ind. Eng. Chem. Res., 39(11),4228-4238.

Viola, E., and Holt, R. (2001). "Don't Leave Steam-Trap Maintenance to

Chance." Chem Eng, 108(11), 88-91.

Vose, D. (1996). Quantitative Risk Analysis: A Guide to Monte Carlo Simula¬

tion Modelling, John Wiley & Sons, New York.

Wang, S. K. (2000). "Handbook of Air Conditioning and Refrigeration.",McGraw-Hill, New York.

Wang, Y., and Feng, X. (2000). "Exergy Analysis Involving Resource Utiliza¬

tion and Environmental Influence." Comput. Chem. Eng., 24, 1243-1246.

Wardle, A. P., Alpbaz, M., Mehta, A., Vennor-Morris, D., and Halksworth, D.

"Control of the heat input to batch reactor." Innovation in Process Energy

Utilisation, Bath, 89-100.

Wedler, G. (1987). Lehrbuch der Physikalischen Chemie, VCH, Weinheim.

Weissler, A. (1948). "Ultrasonic Investigations of Molecular Properties of Liq¬uids. II. The Alcohols." JAm Chem Soc, 70, 1634-1640.

LXVIII

Literature

Werbos, P. J. (1990). "Econometric Techniques: Theory Versus Practice." En¬

ergy, 15(3/4), 213-236.

Willard, G. W. (1947). "Temperature Coefficient of Ultrasonic Velocity in

Solutions." JAcoust Soc Am, 19(1), 235-241.

Worrell, E., Phylipsen, D., Einstein, D., and Martin, N. (2000). "Energy Use and

Energy Intensity of the U.S. Chemical Industry." LBNL-44314, Universityof California, Berkeley.

Worrell, E., Price, L., and Ruth, M. (2001). "Policy Modeling for Energy Effi¬

ciency Improvement in US Industry." Annu. Rev. Energy Environ., 26, 117-

143.

Würsten, F. (2003). "Es braucht grosse Anstrengungen." ETHLife.Zalba, B., Marin, J. M., Cabeza, L. F., and Mehling, H. (2003). "Review on

Thermal Energy Storage with Phase Change: Materials, Heat Transfer

Analysis and Applications." Appl. Therm. Eng., 23, 251-283.

Zehner, P. (2003). "Mischen und Rühren." Chem Ing Tech, 75(12), 1876-1880.

Zhao, X. G, O'Neill, B. K., Roach, J. R., and Wood, R. M. (1998). "Heat Inte¬

gration for Batch Processes: Part 2: Heat Exchanger Network Design."Trans IChemE Part A, 76, 700-710.

LXIX

Resume

Resume

Patrie S. Bieler

Born 28 July, 1975 in Männedorf (ZH), Switzerland

Citizen of Luzern (LU) and Giswil (OW)

Education

June 2000 -

April 2004

October 2000 -

November 2002

October 1995

May 2000

April 1988-

October 1995

Ph.D. in Chemical Engineering, ETH Zurich: "Energy

Modelling of a Multipurpose Chemical Batch Plant" in

close collaboration with IndustryAdvisor: Prof. Dr. K. Hungerbühler, ETH Zurich

Postgraduate Diploma in Industrial Management,ETH Zurich, Diploma thesis „Market Analysis, Positioningand Marketing Strategy for a Producer of UV-Absorbers"

in close collaboration with industryAdvisor: Prof. Dr. A. Seiler, ETH Zurich

Diploma in Chemical Engineering, ETH Zurich

Diploma thesis "Multi Objective Decision-Making under

Uncertainty: A Tool for Automated Screening of Process

Alternatives" at MIT (Boston, USA)Advisors: Prof. Dr. G. J. McRae (MIT) and

Prof. Dr. K. Hungerbühler (ETH Zurich)

Matura Type C (mathematics & natural science), KZO

Wetzikon

Professional Knowledge

2000-2004

2003

March and

April 1999

July 1998 -

October 1998

June 1995 -

September 1995

January 1995

October 1994-

December 1994

Supervising Chemical Engineering Students in their Case

Studies

Supervising a Diploma Thesis in Chemical Engineering

Internship at "The Boston Consulting Group GmbH &Partner" in Munich and Düsseldorf (Germany)

Internship at „Novartis Crop Protection AG" in

Munchwilen (Switzerland)

Internship at „Eidg. Forschungsanstalt für Obst-, Wein-

und Gartenbau" in Wädenswil (Switzerland)

Internship at „CU Chemie Uetikon" in Uetikon (Switzer¬

land)

Internship at „Kantonales Labor" in Zurich (Switzerland)

LXXI

Analysis and Modelling of the Energy Consumption

Documents

Transcript of Analysis and Modelling of the Energy Consumption