Analysis and Modelling of the Energy Consumption
Transcript of 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
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
5.2.1 Description of the Equipment 54
5.2.2 Measurements 55
5.2.3 Model and Conclusions 55
vii
5.3 Heat-Chamber 58
5.3.1 Description of the Equipment 58
5.3.2 Measurements 58
5.3.3 Model and Conclusions 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
5.5.1 Description of the Equipment 66
5.5.2 Measurements 67
5.5.3 Model and Conclusions 71
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
5.7.1 Description of the Equipment 78
5.7.2 Measurements 78
5.7.3 Model and Conclusions 79
5.8 Batch Distillation Column 82
5.8.1 Description of the Equipment 82
5.8.2 Measurements 82
5.8.3 Model and Conclusions 83
5.9 Centrifuge 86
5.9.1 Description of the Equipment 86
5.9.2 Measurements 86
5.9.3 Model and Conclusions 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
nutsche dryers according to Equation (3-25) (PSP data) 107
Figure 6-11 : Modelled specific brine consumption for the reactors and
nutsche dryers according to Equation (3-25) (PSP data) 109
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
Equation (3-14) (one month; PSP data) 122
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
Equation (3-14) (one month; PSP data) 124
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
investigated plant (period: one month; PSP data; total
consumption: 315 MWh) 134
Figure 7-3: Analysis of the total modelled brine consumption of the
investigated plant (period: one month; PSP data; total
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
Structure of a Batch Plant
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
Structure of a Batch Plant
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
Two Approaches for Energy Modelling
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
Two Approaches for Energy Modelling
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
The Bottom-Up Approach
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
Two Approaches for Energy Modelling
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
The Bottom-Up Approach
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
Two Approaches for Energy Modelling
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
Top-Down Modelling of Production Plants (TODOMO)
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
Analysis of the Different Energy Carriers
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
Amount of Products PO [t/month]
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
Amount of Products PO [t/month]
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
Amount of Products PO [t/month]
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
Top-Down Modelling of Production Plants (TODOMO)
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
Analysis of the Different Energy Carriers
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
Top-Down Modelling of Production Plants (TODOMO)
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
Analysis of the Different Energy Carriers
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
Top-Down Modelling of Production Plants (TODOMO)
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
Analysis of the Different Energy Carriers
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
Top-Down Modelling of Production Plants (TODOMO)
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
Analysis of the Different Energy Carriers
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
Top-Down Modelling of Production Plants (TODOMO)
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
Amount of Products PO [t/month]
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
Amount of Products PO [t/month]
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
Top-Down Modelling of Production Plants (TODOMO)
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
Top-Down Modelling of Production Plants (TODOMO)
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
Modelling of Single Unit Operations
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
Modelling of Single Unit Operations
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
Modelling of Single Unit Operations
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
Modelling of Single Unit Operations
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
10 m2, stainless steel nutsche17'19
10 m2, stainless steel nutsche18'19
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
Modelling of Single Unit Operations
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
Modelling of Single Unit Operations
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
5.2.1 Description of the Equipment
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
Modelling of Single Unit Operations
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.
5.2.3 Model and Conclusions
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
Modelling of Single Unit Operations
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
5.3.1 Description of the Equipment
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.
5.3.3 Model and Conclusions
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
Modelling of Single Unit Operations
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
Modelling of Single Unit Operations
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)
Description of the Equipment
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
Modelling of Single Unit Operations
'
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.
Model and Conclusions
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
Modelling of Single Unit Operations
5.4.3 Steam-Jet Vacuum Pumps
Description of the Equipment
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.
Model and Conclusions
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
5.5.1 Description of the Equipment
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
Modelling of Single Unit Operations
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
Modelling of Single Unit Operations
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
Modelling of Single Unit Operations
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%.
5.5.3 Model and Conclusions
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
Description of the Equipment
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
Modelling of Single Unit Operations
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).
Model and Conclusions
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
Description of the Equipment
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
Continuous Equipment
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.
Model and Conclusions
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
Continuous Equipment
5.6.3 Falling-Film Evaporator
Description of the Equipment
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
Modelling of Single Unit Operations
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
Model and Conclusions
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
5.7.1 Description of the Equipment
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
Modelling of Single Unit Operations
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
5.7.3 Model and Conclusions
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
Modelling of Single Unit Operations
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
5.8.1 Description of the Equipment
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
Modelling of Single Unit Operations
5.8.3 Model and Conclusions
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
Modelling of Single Unit Operations
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
5.9.1 Description of the Equipment
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.
5.9.3 Model and Conclusions
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
Modelling of Single Unit Operations
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
Modelling of Single Unit Operations
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Results of the BOTUMO
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
Steam Electricity Brine
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Results of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Results of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Results of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Results of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Results of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Results of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Results of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
(PSP data; modelling period: W = one week, M = one month)
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
Results of the BOTUMO
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
(PSP data; modelling period: W = one week, M = one month)
114
Bottom-Up Modelling of Multipurpose Batch Plants
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
Results of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
(1, 2,..,26, 27) during one month according to Equation (3-16) (PSP data)
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
(1, 2,..,26, 27) during one month according to Equation (3-16) (PSP data)
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Sensitivity Analysis of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Sensitivity Analysis of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Sensitivity Analysis of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Sensitivity Analysis of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Sensitivity Analysis of the BOTUMO
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
Bottom-Up Modelling of Multipurpose Batch Plants
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
Conclusions and Outlook
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
Conclusions and Outlook
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
Conclusions and Outlook
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
ȆSIl Z 0
ȆSIl 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
Steam Electricity Brine
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
nutsche dryers of the investigated building
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
Product Stirrer Vacuum Circulation Div. Electric APOVAC
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
nutsche dryers of the investigated building
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
and nutsche dryers of the investigated building
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
Decalcification 0 0 0 0 0 0
Cleaning 0 0 0 0 0 0
Preparation 0 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
nutsche dryers of the investigated building
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
and nutsche dryers of the investigated building
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
nutsche dryers of the investigated building
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
Product Stirrer Vacuum Circulation Div. Electric APOVAC
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
Decalcification 0 0 0 0 0 0
Cleaning 0 0 0 0 0 0
Preparation 0 0 0 0 0 0
Weighted Sum 98 4 43 4 0 24
Table A-20: Modelled brine consumption of one month of the reactors and
nutsche dryers of the investigated building
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
and nutsche dryers of the investigated building
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)MO^OOOslOiW|iWtOHo
oooOOOOOOOOOOOOOO
oooooooÜoo>
iSoooooooooooooooto
woO
OJOoooooÖ
ooojJ^OOO^OO^^^^^OO
oooooo^oooQ
ooooooooooooooooo
oo
-O°ooooooaa
00
ooooooooooooooo
ooooo
tooo
^o
^
^1
oHH
rH>
o
ooooooooooooooooo
ogoooooooo
<^
ooooooooooooooooooooooooo0ow
°^oooooooooooooooo
^o
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
oooooooooooooooooooooooooo
oOs
op
oooooooooooooooooooooooooo
o<^
oooooooooooooooooooooooooo
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
The Measuring Equipment
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
The Measuring Equipment
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-CnICnICnICnICnICOCOCOCOCO^I-^I-^I-^I-^I-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
Month Steam Electricity Brine Products
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
Table D-3: Measurements of Building 3
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
Table D-4: Measurements of Building 4
Month Steam Electricity Brine Products
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
Table D-5: Measurements of Building 5
Month Steam Electricity Brine Products
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
Table D-6: Measurements of Building 6
Month Steam Electricity Brine Products
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
^o
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
XXXXXXXXXXXXXXXXXXXXXXXXXXXX
<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
(^OoOoooOOOOOoOooooOoooOooO
S'XXXXXXXXXXXXXXXXXXXXXXXXXXXX
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
(^
(^oooooO
LfrO
LfrOOO
LfroooOoooooOoOO
S'XXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXX
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
S'XXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXX
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
(^OOo
(^
(^Oo
(^
(^oO
(^
(^^
Lfr
Lfr
Lfr
(^
(^
-P*
^1
(^
Lfr
(^
(^
Lfr
(^
Lfr
(^
^FinalReaction
-P*
(^
LO
Lfr
(^
LO
-p*
~-\
-P*
(^
Lfr
Jj.
Jj.
to
:iq
to
to^O
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*
^i
Ltl
-P*
-P*
-p*
-p*
00
Jj.
U)
Jj.
oooooooOoooooo
SEL
DistilledSolvent
-o
-o
^1
^1
^1
^1
^i
^1
^1
^1
^i
^1
^i
^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
toM^O
VO
00
^1
On
(^
4j
LO
toM
PS- o 3-
tO
KJ
to^^^
to^^Oo^Oo^OOOoO^^oOoO^^^^^
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
^i
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
(^
(^ooooOoO
Ltl
LtlOO
(^OO
(^O
(^oOoooOO
(^o
,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
4^O
^i
^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)
^iO
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