Dynamics and control of a heat pump assisted … · Dynamics and control of a heat pump assisted...
Transcript of Dynamics and control of a heat pump assisted … · Dynamics and control of a heat pump assisted...
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
1
Dynamics and control of a heat pump assisted extractive dividing-1
wall column for bioethanol dehydration 2
3
Iulian Patraşcu,1 Costin Sorin Bildea,1 Anton A. Kiss,2.3* 4 1 University “Politehnica” of Bucharest, Polizu 1-7, 011061 Bucharest, Romania 5 2 AkzoNobel Research, Development & Innovation, Process Technology SRG, Zutphenseweg 6
10, 7418 AJ Deventer, The Netherlands. E-mail: [email protected] 7 3 Sustainable Process Technology Group, Faculty of Science and Technology, University of 8
Twente, PO Box 217, 7500 AE Enschede, The Netherlands 9 * Corresponding author: [email protected], Tel: +31 26 366 9420 10
11
Keywords 12
Bioethanol dehydration; process dynamics; process control; integrated design; heat pumps 13
14
Highlights 15
• Highly integrated design leading to challenges in process dynamics and control 16 • Efficient control structure ensuring stable operation of VRC assisted E-DWC 17 • Disturbances in feed flowrate and composition can be effectively rejected 18
19
Abstract 20
Recently, a novel heat-pump-assisted extractive distillation process taking place in a dividing-21
wall column was proposed for bioethanol dehydration. This integrated design combines three 22
distillation columns into a single unit that allows over 40% energy savings and low specific 23
energy requirements of 1.24 kWh/kg ethanol. However, these economic benefits are possible 24
only if this highly integrated system is also controllable to ensure operational availability. 25
This paper is the first to address the challenges related to process dynamics and control of this 26
highly integrated system. After showing the control difficulties associated with the original 27
design owing to thermal unbalance, an efficient control structure is proposed which introduces 28
a by-pass and an additional external duty stream to the side reboiler. The range of the external 29
duty is rather small, about 5% of the combined duty of the reboilers, but sufficient to stabilize 30
the system by controlling the temperature on the pre-concentration side of the column. Two 31
quality control loops ensures product purity when the system is affected by feed flowrate and 32
composition disturbances. 33
34
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
2
1. Introduction 1
Bioethanol is a renewable fuel produced in largest amounts, by various routes: corn-to-2
ethanol, sugarcane-to-ethanol, basic and integrated lignocellulosic biomass-to-ethanol. In all 3
cases, the raw materials are pre-treated and then fermented to yield diluted bioethanol of 5-4
12% concentration (Vane, 2008; Huang et al., 2008; Frolkova and Raeva, 2010). This needs 5
to be further concentrated to 99-99.8 %wt (depending on the standard), in an energy intensive 6
process that involves distillation combined usually with extractive distillation (Kiss and Ignat, 7
2013). The concentration of diluted streams from fermentation is not an issue related only to 8
ethanol, but a more generic problem encountered in biorefineries (Kiss et al., 2016). 9
Various energy improvements of the distillation process have been proposed for the ethanol 10
separation and purification, such as the use of internally heat-integrated distillation columns 11
(HIDiC) that are based on the vapor recompression principles (Kiss and Olujic, 2014; Ponce 12
et al., 2015). However, the HIDiC technology is hindered by large equipment costs (leading to 13
long payback times) and it is also limited to producing only hydrated ethanol. 14
Dividing-wall column (DWC) is a proven process intensification distillation technology that 15
can deliver 25-35% savings in energy and investment costs (Dejanovic et al., 2010; Yildirim 16
et al., 2011; Kiss, 2013). In spite of the integration of two distillation columns into a single 17
shell, DWC proved to be controllable providing that a suitable control scheme is selected – 18
based on simple PID or more advanced controllers (Serra et al., 1999; Diggelen et al., 2010; 19
Kiss and Bildea, 2011; Rewagad and Kiss, 2012). The addition of heat pumps to dividing-20
wall column technology was also proposed for various processes but the process dynamics 21
was not investigated (Chew et al., 2014; Luo et al., 2015; Liu et al., 2015; Shi et al., 2015). 22
Using DWC for various processes, including bioethanol dehydration by azeotropic and 23
extractive distillation was proposed by several authors (Bravo-Bravo et al., 2010; Sun et al., 24
2011; Kiss and Suszwalak, 2012). Later, the control analysis of extractive dividing-wall 25
columns revealed satisfactory controllability properties of such integrated systems (Tututi-26
Avila et al., 2014; Zhang et al., 2014). 27
In a follow-up study, Kiss and Ignat (2012) proposed a novel extractive DWC configuration 28
that integrates the three distillation columns of a classic process – pre-concentration 29
distillation column (PDC), extractive distillation column (EDC), and solvent recovery column 30
(SRC) – into a single unit. On top of this integrated extractive DWC system, Luo et al. (2015) 31
added also a vapor recompression (VRC) heat pump to further increase the energy savings up 32
to 40% (see Figure 1). This elegant solution was also featured at the joint PSE-2015 & 33
ESCAPE-25 conference (www.pse2015escape25.dk) among the best contributions in the 34
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
3
process design section. But the authors were also challenged to prove the controllability of 1
such a highly-integrated system. This paper addresses for the first time the challenges related 2
to the process dynamics and control of the proposed integrated heat pump assisted E-DWC. 3
4
2. Problem statement 5
The smart combination of heat pumps (e.g. vapor recompression) with extractive dividing-6
wall column (E-DWC) technology offers significant advantages in terms of low total annual 7
costs and major energy savings of about 40%. But the controllability of the process is just as 8
important as the cost savings, considering that the plant must be available and fully operable 9
in order to deliver the expected design savings. The problem is that a more integrated system 10
has less flexibility and fewer degrees of freedom hence being more difficult to control. To 11
solve this problem, we investigate the dynamics of the process and propose a rather simple 12
but effective process control structure that is able to reject common disturbances, within short 13
settling times and low overshooting. 14
15
3. Results and discussion 16
Plantwide control deals with the strategy of solving the dynamic problems of a full process, 17
under the production (rate and quality) and safety constrains, while minimizing the costs. In 18
this work, an efficient control structure is developed by addressing these tasks of plantwide 19
control. To achieve this goal, dynamic simulations (carried out in Aspen Dynamics) were 20
used to understand the dynamics of the process and to prove the effectiveness of the proposed 21
control structures. The starting point is the VRC assisted E-DWC design proposed by Luo et 22
al. (2015), which achieves 40% energy savings compared to the conventional separation 23
sequence. Note that further energy integration could be attempted by using the product 24
streams for feed pre-heating. However, as there is no liquid flow from the main column to the 25
pre-fractionator, the feed stream acts as reflux. Therefore, feeding below the boiling point 26
improves separation efficiency. Figure 1 presents for details of the VRC assisted E-DWC, 27
including the mass and energy balance. Steady state temperature and composition profiles are 28
shown in Figure 2. The reader is referred to the original paper (Luo et al., 2015) for more 29
details concerning the principle of operation and the efficiency of the system. 30
31
3.1 Process dynamics 32
The original Aspen Plus steady state simulation was exported to Aspen Dynamics. The 33
simulation consists of two separate RADFRAC models which correspond to the PDC and 34
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
4
EDC-SRC sides of the dividing-wall column and are inter-connected by vapor and liquid 1
streams. The dynamic simulation model is presented in Figure 3, which includes the setpoint 2
of the main controllers. 3
The safety of a process is typically achieved by a combination of inherent safe design, safety 4
relief valves and instrument protective functions. The process control system supports safety 5
by controlling the liquid levels and gas pressures, thus ensuring that all the material is 6
contained within the process boundaries. The default inventory controllers provided by Aspen 7
Plus / Aspen Dynamics are: 8
• Level of the PDC sump, by manipulating the bottoms flow rate. According to the 9
steady state boilup ratio V/B, a fraction of this stream is withdrawn as product, while 10
the rest is sent to the side reboiler. 11
• Level of the EDC-SRC sump, by manipulating the bottoms flow rate. 12
• Pressure of the PDC column (top), by manipulating the vapor distillate rate. Note that 13
this controller, required by the flow-driven dynamic simulation, is not part of the 14
practical implementation where the pressures above the top tray of the PDC and on the 15
EDC-SRC side are equal. 16
• Pressure of the EDC-SRC column (top), by manipulating the vapor distillate rate. 17
Some basic changes were made to the dynamic simulation exported by Aspen Plus: 18
• The flow rate of solvent entering the EDC-SRC column was set on flow control. The 19
solvent make-up becomes the manipulated variable in a level control loop. 20
• A heat exchanger was added to allow controlling the temperature of the solvent 21
entering the EDC-SRC column. The duty of this heat exchanger is nearly zero at 22
design condition, but changes when the system is affected by disturbances. 23
• A vapor-split controller was added. This controller is not part of the practical 24
implementation, but is necessary to ensure that the vapor split remains at the design 25
value in the flow-driven dynamic simulation. 26
• A compressor-outlet pressure controller was added. By manipulating the compressor 27
brake-power, the pressure of the compressed vapor is kept at the design value. The 28
practical implementation of controlling the outlet pressure and flow is briefly 29
described in next section. 30
After implementing the basic inventory control, the simulation can be run. Starting from the 31
steady state, as long as the system is not disturbed, the process variables (flow rates, 32
temperatures, pressures) stay at their nominal values. However, the simulation breaks for any 33
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
5
small disturbances. Quality control loops were added, as follows: 1
• A temperature controller was added on the condensate stream. 2
• A temperature controller was added in the EDC-SRC. Temperature on stage 33 (the 3
most sensitive tray) is controlled by the EDC-SRC reboiler duty. 4
After these changes, the simulation is successful as long as it starts from the steady state and 5
no disturbance is introduced. However, the simulation still breaks for any small increase of 6
the feed rate, which shows that the operating point is unstable. What happens is the following. 7
Increasing the flow rate of PDC feed (cold liquid) leads to a lower vapor flow rate along PDC. 8
As a result, the pressure drops and less vapor is fed to the EDC-SRC section (the PDC 9
pressure controller closes the vapor distillate valve). The pressure on the EDC-SRC section 10
drops and the EDC-SRC pressure controller decreases the EDC-SRC vapor distillate rate. 11
Then, less heat is transferred in the (heat-integrated) side reboiler, followed by further 12
decrease of the vapor flow rate along the PDC. Simultaneously, more EG is found in the 13
bottom section of the EDC-SRC which leads to further temperature increase (despite the 14
lower pressure). The SRC-EDC temperature controller reacts and decreases the reboiler duty, 15
leading to an additional reduction of the vapor flow along the EDC-SRC section (and 16
therefore to further decrease of the SRC-EDC vapors distillate flow). Clearly, an appropriate 17
control structure is required. 18
The key to stabilizing the system is adding a small duty to the last stage of the PDC column. 19
This duty can then be used as new manipulated variable in a control loop which ensures that 20
the vapor flow rate from the PDC to the SRC-EDC side is constant when the system is 21
affected by disturbances. When the valve-position controller VPC (Figure 3) is added, feed 22
flow rate and composition disturbances can be introduced without affecting the stability of the 23
process. However, the control structure shown in Figure 3 is not able to maintain the quality 24
of the products, namely the purity of the water and ethanol streams. Moreover, measuring the 25
flow rate of the vapor going from the PDC to the SRC-EDC side is difficult in practice. 26
Dynamic simulations also reveal that the solvent feed and PDC bottoms heat exchangers 27
require heating or cooling, depending on the disturbance. Therefore, the control structure 28
presented in Figure 3 is further refined, as explained in the next section. 29
30
3.2 Process control 31
As the original design of the VRC assisted E-DWC has control difficulties associated with the 32
thermal unbalance, we propose here a novel efficient control structure that introduces an 33
additional duty stream to the side reboiler. The range of the additional duty is rather small, 34
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
6
about 5% of the combined duty of the reboilers, but this is sufficient to counteract the possible 1
thermal unbalance. Moreover, the design changes and additional controllers ensures the 2
quality of the water and ethanol product streams. 3
Figure 4 shows the control structure for the VRC assisted E-DWC. For convenience, the 4
setpoint of several controllers are included, while detailed information about controller 5
settings can be found in Table 1. The production rate is set by changing the setpoint of the 6
flow controller manipulating the feed stream flow rate. A ratio controller ensures that the 7
amount of solvent used (make-up and recycle stream) is proportional to the feed flowrate. 8
This ratio is set by a quality controller which achieves constant mass fraction of water 9
impurity in the ethanol product. The temperature of the solvent added to the E-DWC unit is 10
controlled by a split range setup, in which one of the heat exchangers cools-down or heats-up 11
the solvent stream, as required. The compressor is operated at constant outlet pressure – the 12
pressure at which the valve on the outlet line opens – and variable flow rate, which is 13
achieved by means of an internal recycle or other means. The duty of the PDC (external) 14
reboiler is provided by a fraction of the compressed vapors (about 94% at design conditions) 15
and by hot utility. The external duty provided by the hot utility is used to control the 16
temperature on the PDC side of the dividing-wall column. When the tuning of this loop is 17
sufficiently aggressive (large gain, small integral time), the system is stabilized. The ‘valve-18
position control’ (VPC) loop ensures both rapid response to disturbances and energy 19
efficiency. Thus, when higher duty is required by the PDC temperature controller, the flow 20
rate of hot utility is increased. This has a fast effect on the controlled variable. Afterwards, the 21
VPC controller gradually increases the flow rate of compressed vapors passing through the 22
reboiler (by-pass fraction is reduced), until the utility flow rate returns to its design / set point 23
value. The set point of the PDC temperature controller is given, in cascade fashion, by a 24
quality controller keeping constant the mass fraction of ethanol impurity in the water product 25
stream. Note that the dynamics of the concentration transducers (measurements) on the 26
ethanol and water streams was properly modelled including 1 minute sampling interval and 1 27
minute dead-time. 28
In practice, the feed contains small amounts (about 0.1%wt.) of dissolved CO2, which have to 29
be removed from the system. To this end, a small purge stream from the reflux drum could be 30
used as manipulated variable, in a pressure control loop (Loy et al., 2015; Batista et al., 2012). 31
Table 1 provides the controller tuning parameters. The control loops were tuned by a simple 32
version of the direct synthesis method (Luyben and Luyben, 1997), according to which, the 33
desired closed-loop response for a given input is specified. With the model of the process 34
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
7
known, the required form and the tuning of the feedback controller are then back-calculated. 1
For all controllers, the acceptable control error (∆εmax) and the maximum available control 2
action (∆umax) were specified. Afterwards, the controller gain (expressed in engineering units) 3
was calculated as Kc = ∆umax/ ∆εmax and translated into percentage units. First order open-loop 4
models were assumed, in order to calculate the integral time of the pressure, temperature and 5
concentration control loops (Bildea and Kiss, 2011). As rough evaluations of the process time 6
constants τ, 12 min, 20 min and 40 min were used, respectively. It can be shown (Luyben and 7
Luyben, 1997) that the direct synthesis method requires that the reset time of a PI controller is 8
equal to the time constant of the process (τi = τ). For the level controllers, a rather large reset 9
time τi = 60 min was chosen, as no tight control is required. When the PDC temperature 10
control loop is sufficiently fast, the system is stabilized and disturbances can be successfully 11
rejected. To tune the concentration controllers, the ultimate gain and the period of oscillations 12
at stability limit were found using the ATV (Auto Tuning Variation) method. These values 13
were further used to calculate the controller parameters according to Tyreus-Luyben settings 14
(Luyben and Luyben, 1997). 15
Figure 5 and Figure 6 show the dynamic response of the system for feed flow rate 16
disturbances. Starting from the steady state, the feed rate is ramped up by 10% (from t = 2 h 17
to t = 4 h), brought back to the initial value (at t = 12 h), decreased by 10% (at t = 22 h), and 18
returned again to the initial value (at t = 32 h). Both the ethanol and water flow rates change 19
accordingly to their feed rate and reach their new steady state values in a rather short time 20
with almost no overshooting (Figure 5, left). It is remarkable that the composition of the 21
ethanol and water stream products remains practically unchanged (Figure 5 right, where the 22
mass fraction of solvent in water is shown, as this variable is not measured / controlled). 23
Figure 6 shows the key internal variables, for the same scenario. The various internal flow 24
rates follow the change of the feed rate (plot a). The ethanol purity is maintained by changing 25
the solvent/feed ratio (plot b). When the feed rate increases (from t = 2 h to t = 12 h), higher 26
side-reboiler duty is required. This is achieved by increasing the external duty (plot c) and by 27
reducing the by-pass around the side-reboiler (plot d). Note that when lower duty is required 28
(from t = 22 h to t = 32 h), this is achieved by increasing the by-pass, while the external duty 29
remains at the minimum value of 2 GJ/h. 30
The system can also withstand larger feed flow rate disturbances, for example +/- 20%, if they 31
occur as ramp with the maximum slope of about 5% / hour. Moreover, a larger range of the 32
external duty is required (0-16 GJ/h). Although this seems to be a limitation of the 33
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
8
controllability of this highly integrated system, fast changes of the feed rate (for example in 1
the form of step signals) are not likely to be encountered in practice. 2
Figure 7 and Figure 8 show the dynamic response to various changes of feed ethanol 3
concentration, from the initial value of 10 %wt. to 12 %wt., and 8 %wt., similarly to the 4
previously described scenario. The flowrates of ethanol and water products follow the 5
amounts existing in the feed (Figure 7 - left). The changes in the concentration of impurities 6
(solvent in water, and water in ethanol) are quite minor and the system stabilizes shortly 7
(Figure 7 - right). The key internal variables shown in Figure 8 have limited variations, and 8
return to their initial values when the disturbance is removed. 9
Figure 9 and Figure 10 show dynamic simulation results for changing the setpoint of the 10
ethanol concentration controller from the initial value of 99.8 %wt to 99.5 %wt (at t = 2 h), 11
and to 99 %wt (at t = 12 h), after which the setpoint is brought back to the initial value in two 12
steps (at t = 22 h and t = 32 h). The setpoint changes are implemented as ramp signals, over 8-13
hour period. The product flow rates change according to the mass balance (Figure 9 – left), 14
while the ethanol concentration follows it setpoint (Figure 9 – right). The internal variables 15
have limited changes (Figure 10) and return to initial values once the disturbance is removed. 16
Tututi-Avila et al. (2014) analyzed the control properties of a conventional extractive 17
dividing-wall column used for ethanol dehydration, where the feed stream is a concentrated 18
ethanol solution (93 %wt). The control structure involved conventional temperature, pressure 19
and level control loops. Use of the vapor split as manipulated variable in a temperature 20
control loop was necessary to properly reject step feed flow rate and composition 21
disturbances, product purities returning close to their specifications in about 1.5 h. For most 22
disturbances, the VRC assisted E-DWC is equally fast. However, the dynamics is slower 23
when the ethanol feed concentration decreases, the new steady state being achieved in about 5 24
hours. By using concentration controllers, the control structure of the VRC assisted E-DWC 25
achieves better composition control, without the need of vapor split manipulation. In contrast 26
to the conventional E-DWC system, the VRC assisted E-DWC cannot withstand step 27
disturbances, due to higher degree of integration. 28
29
4. Conclusions 30
Integrated processes combining vapor recompression with dividing-wall column technology 31
have more interactions of variables and fewer degrees of freedom as compared to classic 32
DWC systems. This makes it challenging to control and questions the expected benefits. 33
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
9
In particular, the system is unstable even when appropriate basic inventory (level and 1
pressure) controllers are provided. The reason is the positive feedback due to vapor-2
recompression: decrease of the pre-fractionator temperature leads to lower flow rate of the 3
recompressed vapor, which results in less heat being transferred in the side reboiler, and 4
further temperature decrease. Some design changes are necessary to provide the manipulated 5
variables necessary for stabilizing the system. Thus, the side reboiler has an additional 6
external duty, used for pre-fractionator temperature control. Moreover, only a fraction of the 7
recompressed vapor is used for heating the side reboiler. Manipulating the by-pass provides 8
additional heat when needed, but also ensures that the external duty is not shut down and is 9
kept to a minimum value. Concentration measurements are necessary to achieve good quality 10
control of the ethanol and water streams. 11
An effective control structure that achieves the control objective is the main result of this 12
study, which is the first to prove the controllability of vapor recompression assisted extractive 13
DWC. The system is robust to ramp feed rate and feed composition disturbances and to ramp 14
changes of ethanol purity setpoint. As step changes are not tolerated, this study provides a 15
clear warning to process and control engineers that such changes should be implemented in a 16
slow fashion. 17
18
Acknowledgement 19
Financial support of the European Commission through the European Regional Development 20
Fund and of the Romanian state budget, under the grant agreement 155/25.11.2016 (Project 21
POC P-37-449, acronym ASPiRE) is gratefully acknowledged. 22
23
24
References 25
1. Batista F. R. M., Follegatti-Romero L. A., Bessa L. C. B. A., Meirelles A. J. A., 26
Computational simulation applied to the investigation of industrial plants for bioethanol 27
distillation, Computers and Chemical Engineering, 46 (2012), 1-16. 28
2. Bildea C. S., Kiss A. A., Dynamics and control of a biodiesel process by reactive 29
absorption, Chemical Engineering Research and Design, 89 (2011), 187-196. 30
3. Bravo-Bravo C., Segovia-Hernández J.G., Gutiérrez-Antonio C., Duran A.L., Bonilla-31
Petriciolet A., Briones-Ramírez A., Extractive dividing wall column: Design and 32
optimization, Industrial & Engineering Chemistry Research, 49 (2010), 3672-3688. 33
4. Chew J. M., Reddy C. C. S., Rangaiah G. P., Improving energy efficiency of dividing-34
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
10
wall columns using heat pumps, Organic Rankine Cycle and Kalina Cycle, Chemical 1
Engineering and Processing: Process Intensification, 76 (2014), 45-59. 2
5. Dejanović I., Matijašević L., Olujić Ž., Dividing wall column - A breakthrough towards 3
sustainable distilling, Chemical Engineering and Processing, 49 (2010), 559-580. 4
6. Diggelen R. C. van, Kiss A. A., Heemink A. W, Comparison of control strategies for 5
dividing-wall columns, Industrial & Engineering Chemistry Research, 49 (2010, 288-307. 6
7. Frolkova A. K., Raeva V. M., Bioethanol dehydration: State of the art, Theoretical 7
Foundations of Chemical Engineering, 44 (2010), 545-556. 8
8. Huang H. J., Ramaswamy S., Tschirner U. W., Ramarao B. V., A review of separation 9
technologies in current and future biorefineries, Separation and Purification Technology 10
62 (2008), 1-21. 11
9. Kiss A. A., Bildea C. S., A control perspective on process intensification in dividing-wall 12
columns, Chemical Engineering and Processing, 50 (2011), 281-292. 13
10. Kiss A. A., Suszwalak D. J-P. C., Enhanced bioethanol dehydration by extractive and 14
azeotropic distillation in dividing-wall columns, Separation & Purification Technology, 15
86 (2012), 70-78. 16
11. Kiss A. A., Ignat R. M., Innovative single step bioethanol dehydration in an extractive 17
dividing-wall column, Separation & Purification Technology, 98 (2012), 290-297. 18
12. Kiss A. A., Ignat R. M., Optimal economic design of a bioethanol dehydration process by 19
extractive distillation, Energy Technology, 1 (2013), 166-170. 20
13. Kiss A.A., Advanced distillation technologies - Design, control and applications. Wiley, 21
Chichester, UK, 2013. 22
14. Kiss A. A., Olujic Z., A review on process intensification in internally heat-integrated 23
distillation columns, Chemical Engineering and Processing: Process Intensification, 86 24
(2014), 125-144. 25
15. Kiss A. A., Lange J. P., Schuur B., Brilman D. W. F., van der Ham A. G. J., Kersten S. R. 26
A., Separation technology - Making a difference in biorefineries, Biomass and Bioenergy, 27
95 (2016), 296-309. 28
16. Liu Y., Zhai J., Li L., Sun L., Zhai C., Heat pump assisted reactive and azeotropic 29
distillations in dividing wall columns, Chemical Engineering and Processing: Process 30
Intensification, 95 (2015), 289-301. 31
17. Loy Y. Y., Lee X. L., Rangaiah G. P., Bioethanol recovery and purification using 32
extractive dividing-wall column and pressure swing adsorption: An economic comparison 33
after heat integration and optimization, Separation and Purification Technology, 149 34
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
11
(2015), 413-427. 1
18. Luo H., Bildea C. S., Kiss A. A., Novel heat-pump-assisted extractive distillation for 2
bioethanol purification, Industrial & Engineering Chemistry Research, 54 (2015), 2208-3
2213. 4
19. Luyben W. L., Luyben M. L., Essentials of process control, McGraw-Hill, US, 1997. 5
20. Ponce G. H. S. F., Alves M., Miranda J. C. C., Maciel Filho R., Wolf Maciel M. R., Using 6
an internally heat-integrated distillation column for ethanol-water separation for fuel 7
applications, Chemical Engineering Research and Design, 95 (2015), 55-63. 8
21. Rewagad R. R., Kiss A. A., Dynamic optimization of a dividing-wall column using model 9
predictive control, Chemical Engineering Science, 68 (2012), 132-142. 10
22. Serra M., Espuña A., Puigjaner L., Control and optimization of the divided wall column, 11
Chemical Engineering and Processing, 38 (1999), 549-562. 12
23. Shi L., Huang K., Wang S.-J., Yu J., Yuan Y., Chen H., Wong D. S. H., Application of 13
vapor recompression to heterogeneous azeotropic dividing-wall distillation columns, 14
Industrial & Engineering Chemistry Research, 54 (2015), 11592-11609. 15
24. Sun L. Y., Chang X. W., Qi C. X., Li Q. S., Implementation of ethanol dehydration using 16
dividing-wall heterogeneous azeotropic distillation column, Separation Science and 17
Technology, 46 (2011), 1365-1375. 18
25. Tututi-Avila S., Jiménez-Gutiérrez A., Hahn J., Control analysis of an extractive dividing-19
wall column used for ethanol dehydration, Chemical Engineering and Processing: Process 20
Intensification, 82 (2014), 88-100. 21
26. Vane L. M., Separation technologies for the recovery and dehydration of alcohols from 22
fermentation broths, Biofuels, Bioproducts and Biorefining, 2 (2008), 553-588. 23
27. Yildirim O., Kiss A. A., Kenig E. Y., Dividing wall columns in chemical process industry: 24
A review on current activities, Separation & Purification Technology, 80 (2011), 403-417. 25
28. Zhang H., Ye Q., Qin J., Xu H., Li N., Design and control of extractive dividing-wall 26
column for separating ethyl acetate-isopropyl alcohol mixture, Industrial & Engineering 27
Chemistry Research, 53 (2014), 1189-1205. 28
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
12
Tables 1
2
Table 1. Controller tuning parameters 3
Controller PV, value & range OP, value & range Kc, %/%
Ti, min
Solvent feed to PDC
FC Flow rate = 16.736 t/h Not modelled in flow-driven simulation
QC Water in ethanol product = 800 ppm solvent : feed = 0.1339 0.581 17.1
400 … 2000 ppm 0…0.3
LC Level Modeled as instantaneous unit
TC Temperature = 52.5 °C Cooling duty = 0 GJ/h 10 60
45 …62 °C -4…0 GJ/h
TC Temperature = 52.5 °C Heating duty = 0 GJ/h 10 50
40…70 °C 0…4 GJ/h
Fresh feed Not modelled in flow-driven simulation
FC Flow rate = 125 t/h
Column
PC Column pressure = 1 bar Vapor flow rate = 1142 kmol/h
20 12
0 … 2 bar 0…2284 kmol/h
PC Compressed vapor = 3.7 bar Compressor power = 1.8 MW
1 10
3.5…4 bar 0…3.6 MW
RC Reflux ratio L/D = 3.20
LC Sump level = 2.05 m EG recycle = 16.44 t/h 10 60
0…4.1 m 0…32.88 t/h
TC Tray 33 temperature = 184.18 °C Reboiler duty = 37.7 GJ/h
1 20
170…200 °C 0…75 GJ/h
TC Reflux temperature = 105.3 °C Cooling duty = -4.233 GJ/h
10 30
95…115 °C -24…0 GJ/h
Side reboiler
LC Level Modelled as instantaneous unit
PC Pressure = 1.25 bar Modelled as instantaneous unit
TC PF Tray 6 temperature = 100.1 °C Trim duty = 2 GJ/h 49.3 9.2
90…110 °C 0…8 GJ/h
QC Ethanol in water product = 89 ppm SP of tray 6 TC = 101 °C 0.14 6.6
0…200 ppm 90...110 °C
VPC Trim duty = 2 GJ/h By-pass = 0.061 1 10
0…8 GJ/h 0…0.5
Dynamics and control of a heat pump assisted E-DWC for bioethanol dehydration Patraşcu, Bîldea, Kiss
13
Figure captions (auto-updated) 1
2
Figure 1. Flowsheet of the VRC assisted extractive-DWC for bioethanol dehydration 3
4
Figure 2. Steady state temperature and liquid phase composition profiles of pre-fractionator 5
(dashed lines) and SRC-EDC (continuous lines) 6
7
Figure 3. Dynamic simulation model of the VRC assisted extractive-DWC for bioethanol 8
dehydration including the control loops necessary for stabilizing the process 9
10
Figure 4. Control structure of the VRC assisted E-DWC for bioethanol dehydration 11
12
Figure 5. Dynamic response of products flow rate and purity, for ±10% disturbance of feed 13
flow rate 14
15
Figure 6. Key internal variables for ±10% flow disturbance. a) Flow rates; b) Ethanol product 16
composition controller; c) Prefractionator temperature controller; d) Side-reboiler external 17
duty controller (VPC) 18
19
Figure 7. Dynamic response of products flow rate and purity, for +/- 2% disturbance of feed 20
concentration 21
22
Figure 8. Key internals variables for +/- 2% disturbance of feed concentration a) Flow rates; 23
b) Ethanol product composition controller; c) Prefractionator temperature controller; d) Side-24
reboiler external duty controller (VPC) 25
26
Figure 9. Dynamic response of products flow rate and purity, for 99.8 to 99.0%wt range 27
disturbances of ethanol product purity 28
29
Figure 10. Key internals variables for 99.8 to 99.0%wt range disturbance of ethanol product 30
purity. a) Flow rates; b) Ethanol product composition controller; c) Prefractionator 31
temperature controller; d) Side-reboiler external duty controller (VPC) 32
33
Figure 1
Ethanol
Water
Feed
rL
rV
E-DWC
Solvent (EG)
1
1
13
18
30
37
4
31
17
EDC
PD
C
SR
C
FEHE
S-REB
COMP
Solvent (recycle)
Q = 1947 kW
25°C, 1.2 bar125000 kg/h10%w EtOH90%w H2O
39.7°C
204.3°C, 1.25 bar16487.3 kg/h 100%w EG
25°C, 250 kg/h 100%w EG
52.5°C
Q = 12024 kW
104.7C, 1.19 bar112750 kg/h99.8%w H2O00.2%w EG
Reflux ratio = 0.646Boilup ratio = 2.587rL=0.0, rV=0.548Qreb = 10043 kW
115.3°C, 3.7 bar12500 kg/h99.9%w EtOH00.1%w H2O
78°C, 1 bar
115.3°C3.7 bar52542 kg/h99.9%w EtOH00.1%w H2O
149.7°C, 3.7 bar
W = 1808 kWQcnd = –267 kW
Figure 2
50
90
130
170
210
250
0 5 10 15 20 25 30 35 40
Tem
pe
ratu
re /
[°C
]
Stage
0.00
0.20
0.40
0.60
0.80
1.00
0 5 10 15 20 25 30 35 40
Ma
ss f
ract
ion
Stage
WATER
ETHANOL
ETHYLENE GLYCOL
Figure 3
1
4
17
31
37
VPC
LCFC
PC
PC
PC
TC33
TC
TC
FC
1
13
Solvent make-up
Feed
Ethanol
Water
Solvent recycle
LC
L/D = 3.203
V/B = 0.173
3.7 bar
1 bar
1.128 bar
125 t/h
153.3 °C
701.15
kmol/h
115.3 °C
52.5 °CLC
X
32
FC
LC
PDC EDC-SRC
Figure 4
TC
LC
FC
PC
PC
TC
TC
TC
FC
Solvent
make-up
Feed
Ethanol
Water
Solvent recycle
LC
L/D = 3.203
3.7 bar
125 t/h
184.18 °C
115.3 °C
52.5 °CLC
LC
VPC
QC
89 ppm
Ethanol
2 GJ/h
QT
QTQC
800 ppm
Water
TC
Splitrange
1
4
17
31
33
37
6
1 18
13 30
1 bar
By-pass
PC
19
1.25 bar
x
Figure 5
10000
11000
12000
13000
14000
15000
100000
105000
110000
115000
120000
125000
0 5 10 15 20 25 30 35 40 45
Flo
w r
ate
/ [
kg
/h]
Time / [h]
Water
Ethanol
0.00
0.10
0.20
0.30
0.40
0.50
0 5 10 15 20 25 30 35 40 45
Ma
ss f
ract
ion
/ [
%w
t]
Time / [h]
Solvent in Water
Water in Ethanol
10000
11000
12000
13000
14000
15000
100000
105000
110000
115000
120000
125000
0 5 10 15 20 25 30 35 40 45
Flo
w r
ate
/ [
kg
/h]
Time / [h]
Water
Ethanol
0.00
0.10
0.20
0.30
0.40
0.50
0 5 10 15 20 25 30 35 40 45
Ma
ss f
ract
ion
/ [
%w
t]
Time / [h]
Solvent in Water
Water in Ethanol
Figure 6
0
15000
30000
45000
60000
75000
0 5 10 15 20 25 30 35 40 45
Flo
w r
ate
/ [
kg
/h]
Time / [h]
Reflux
Ethanol
Distilate (HP)
to Reboiler
Bypass
0
1
2
3
4
5
98.0
99.0
100.0
101.0
102.0
103.0
0 5 10 15 20 25 30 35 40 45
Du
ty /
[G
J/h
r]
Tem
pe
ratu
re /
[°C
]
Time / [h]
PV
SP OP
(a) (b)(a) (b)
(c) (d)
(a) (b)
(c)
(a) (b)
(d)(c)
(a) (b)
0.130
0.131
0.132
0.133
0.134
0.135
790
800
810
820
830
840
0 5 10 15 20 25 30 35 40 45
So
lve
nt
/ Fe
ed
ra
tio
Wa
ter
/ [p
pm
]
Time / [h]
PV
SP
OP
0.90
0.92
0.94
0.96
0.98
1.00
1
2
3
4
5
6
0 5 10 15 20 25 30 35 40 45
Sp
lit
/ [-
]
Du
ty /
[G
J/h
r]
Time / [h]
PV
SP OP
Figure 7
10000
11000
12000
13000
14000
15000
108000
110000
112000
114000
116000
118000
0 5 10 15 20 25 30 35 40 45
Flo
w r
ate
/ [
kg
/h]
Time / [h]
Water
Ethanol
0.00
0.10
0.20
0.30
0.40
0.50
0 5 10 15 20 25 30 35 40 45
Ma
ss f
ract
ion
/ [
%w
t]
Time / [h]
Solvent in Water
Water in Ethanol
Figure 8
0
15000
30000
45000
60000
75000
0 5 10 15 20 25 30 35 40 45
Flo
w r
ate
/ [
kg
/h]
Time / [h]
Reflux
Ethanol
Distilate (HP) to Reboiler
Bypass
0
2
4
6
8
10
94.0
96.0
98.0
100.0
102.0
104.0
0 5 10 15 20 25 30 35 40 45
Du
ty /
[G
J/h
r]
Tem
pe
ratu
re /
[°C
]
Time / [h]
PV SP
OP
(d)(c)
(a) (b)
0.00
0.05
0.10
0.15
0.20
0.25
600
700
800
900
1000
1100
0 5 10 15 20 25 30 35 40 45
So
lve
nt
/ Fe
ed
ra
tio
Wa
ter
/ [p
pm
]
Time / [h]
PV
SP
OP
0.80
0.84
0.88
0.92
0.96
1.00
1
2
3
4
5
6
0 5 10 15 20 25 30 35 40 45
Sp
lit
/ [-
]
Du
ty /
[G
J/h
r]
Time / [h]
PV
SP
OP
Figure 9
0.00
0.40
0.80
1.20
1.60
2.00
0 5 10 15 20 25 30 35 40 45
Ma
ss f
ract
ion
/ [
%w
t]
Time / [h]
Solvent in Water
Water in Ethanol
12300
12400
12500
12600
12700
12800
112500
112600
112700
112800
112900
113000
0 5 10 15 20 25 30 35 40 45
Flo
w r
ate
/ [
kg
/h]
Time / [h]
Water Ethanol
Figure 10
0.00
0.05
0.10
0.15
0.20
0.25
0
4000
8000
12000
16000
20000
0 5 10 15 20 25 30 35 40 45
So
lve
nt
/ Fe
ed
ra
tio
Wa
ter
/ [p
pm
]
Time / [h]
PV
SP
OP
0
15000
30000
45000
60000
75000
0 5 10 15 20 25 30 35 40 45
Flo
w r
ate
/ [
kg
/h]
Time / [h]
Reflux
Ethanol
Distilate (HP) to Reboiler
Bypass
0
1
2
3
4
5
98.0
99.0
100.0
101.0
102.0
103.0
0 5 10 15 20 25 30 35 40 45
Du
ty /
[G
J/h
r]
Tem
pe
ratu
re /
[°C
]
Time / [h]
PV SP
OP
(d)(c)
(a) (b)
0.90
0.92
0.94
0.96
0.98
1.00
1
2
3
4
5
6
0 5 10 15 20 25 30 35 40 45
Sp
lit
/ [-
]
Du
ty /
[G
J/h
r]
Time / [h]
PV
SP
OP