- Email: [email protected]
Design and control of different pressure thermally coupled reactive distillation for amyl acetate synthesis
Accepted Manuscript Title: Design and control of Different Pressure Thermally Coupled Reactive Distillation for Amyl Acetate synthesis Authors: Qingrui Zhang, Tong Guo, Chao Yu, Yonglei Li PII: DOI: Reference:
S0255-2701(17)30350-1 http://dx.doi.org/10.1016/j.cep.2017.09.002 CEP 7066
To appear in:
Chemical Engineering and Processing
Received date: Revised date: Accepted date:
10-4-2017 28-7-2017 4-9-2017
Please cite this article as: Qingrui Zhang, Tong Guo, Chao Yu, Yonglei Li, Design and control of Different Pressure Thermally Coupled Reactive Distillation for Amyl Acetate synthesis, Chemical Engineering and Processinghttp://dx.doi.org/10.1016/j.cep.2017.09.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Design and control of Different Pressure Thermally Coupled Reactive Distillation for Amyl Acetate synthesis Qingrui Zhang*, Tong Guo, Chao Yu, Yonglei Li
College of Chemical Engineering, Qingdao University of Science and Technology, Qingdao 266042, China
*
Corresponding author.
E-mail address: [email protected] (Q. Zhang).
Graphical abstract:
Fig.1. Improved control structure for the different pressure thermally coupled reactive distillation The control strategy of different pressure thermally coupled reactive distillation process was investigated. The improved control structure is proposed to control the AmAc composition by a cascade connection of temperature/composition cascade control and pressure control, and the dynamic performance indicate that the different pressure thermally coupled reactive distillation process robust control can be achieved under the improved control structure.
Highlights：
The DPT-RD process was developed for the AmAc esterification process. The feasibility and effectiveness for the DPT-RD process were investigated. The control strategy of the DPT-RD process was developed and studied.
The results of this study provide theoretical guidance for industrial applications.
Abstract: The different pressure thermally coupled reactive distillation (DPT-RD) process is an attractive heat-integrated design in chemical engineering development processes. In this study, a rigorous simulation of the DPT-RD process for amyl acetate synthesis in a neat operation mode is performed using the Aspen Plus simulator. The economic feasibility and controllability of the DPT-RD process were evaluated based on the research of steady-state optimization and dynamic control. The optimum DPT-RD process in terms of the total annual cost (TAC) is screened. The results show that 33.74% reduction in energy consumption and 7.96% savings in TAC can be achieved by the DPT-RD process compared to the conventional reactive distillation. The control strategy of the DPT-RD process was developed and the basic control structure with proportional control and the improved control structure with temperature/composition cascade control were presented. The dynamic results indicate that the improved control structure can effectively handle disturbances of ±20% in feed flow rate and of 5% in feed composition. Overall results demonstrate that the DPT-RD process is a promising approach for amyl acetate synthesis. Keywords: Amyl acetate; Different pressure thermally coupled reactive distillation; Process optimization; Control structures 1.Introduction With the increasing demand for energy, process intensification has attracted attention of the industry and research community to reduce energy consumption and capital investment. Reactive distillation (RD) combines reaction and separation in a single unit attributed to the potential advantages over the catalyst usage, reactive heat utilization, and cost savings [1,2]. RD has been extensively studied for decades, but it has received renewed attention in recent years due to the novel research trends about process development with higher degrees of intensification. In terms of thermal effects of the reaction, RD systems can be divided into three broad categories: those with high thermal effect, those with considerable thermal effect, and those with negligible or no thermal effect [3]. Research on heat-integrated technology has mainly focused on the RD systems involving reactions with high thermal effect and produced remarkable results [4-7]. Hernandez et al. investigated the design and control strategies of reactive dividing wall column process for the production of ethyl acetate and verified the energy saving effect [8]. The synthesis of tert-amyl methyl ether (TAME) by heat-integrated reactive distillation column was investigated by Vanaki and co-workers and the simulation results showed that 22% of energy savings could be achieved compared to the conventional reactive distillation (CRD) column [9]. Other forms of thermal integration have also been studied such as multi-effect reactive distillation and heat pump-assisted reactive distillation [10-12]. However, previous studies on the RD systems involving reactions with negligible or no thermal effects have only focused on internal mass integration [3,13] and research on energy savings of the system is scarce. For these systems, the effect of internal heat integration seems much weaker than that of internal mass integration; therefore, other thermal integration methods such as external heat integration can be considered. Gao and co-workers proposed an interesting structure of different pressure thermally coupled
reactive distillation (DPT-RD) for TAME synthesis [14]. The DPT-RD process is a highly coupled process of reaction, separation, and energy integration. The connection between the high-pressure (HP) column and the low-pressure (LP) column is realized by a compressor and a throttle valve, which doesn’t have a significant impact on the original material coupling. The top stream of the HP column is used as the heating medium of the LP column reboiler; thus, the DPT-RD process has great potential for energy savings through this type of external heat integration. The DPT-RD process simultaneously achieved the material coupling and heat integration of the RD systems involving reactions with negligible or no thermal effect; thus, it will be a good choice for energy-saving improvements on the RD systems. The RD process operated in the “neat” mode is economical because the one-column process has lower capital investment and energy costs than the two-column system [15] that enhances the advantages of the DPT-RD process in terms of investment. However, the control of the DPT-RD process operated in neat mode is difficult to achieve. The control of the neat RD process must maintain the precise balance of the fresh feed streams because the imbalance will inevitably result in a gradual buildup of one of the reactants and a loss of conversion and product purities [16]. On the other hand, the controllability and effectiveness of this process is a large challenge because of the high degree of coupling and nonlinearity of the DPT-RD process. Therefore, it is essential to study the control strategy of the DPT-RD process. The main purpose of this study is to investigate the feasibility and effectiveness of further heat integration for the RD systems involving reactions with negligible or no thermal effects by the DPT-RD process and develop a control strategy. Taking the amyl acetate (AmAc) esterification process as an example, the steady state optimization and dynamic control of the DPT-RD process are studied. The optimum design of the steady state process is obtained based on the total annual cost (TAC). The economic comparison between the DPT-RD and CRD processes in optimal conditions is also presented. Two control structures are proposed to achieve the robust control of the DPT-RD process and maintain the purity of products for different feed disturbances. Their dynamic responses are also analysed and discussed. In this work, both steady state optimization and dynamic control are carried out with Aspen Plus and Aspen Dynamics (version 7.2). 2. Process description In this study, the esterification of AmAc was carried out in the presence of an acid ion-exchange polymeric resin, Amberlyst 15. The second-order reversible reaction can be described by the following expression (1): C7H14O2 H 2O CH3COOH C5H12O
（1）
The reaction kinetics can be expressed as follows [17]: r mcat (kFCHAcCAmOH kBCAmAcCH2O)
（2）
kF 31.1667 exp(
51740 ) RT
（3）
kB 2.25533exp(
45280 ) RT
（4）
There are many types of azeotropes in the AmOH-HAc-AmAc-H2O quaternary system and the NRTL activity coefficient model is used to predict the phase equilibrium of these four components.
Furthermore, we adopt the Hayden-O’Connell model to account for the vapor non-ideality because HAc is associated in the vapor phase. The parameters and coefficients in phase equilibrium calculations are listed in Table 1 [18]. In the reactive distillation process for the AmAc synthesis, the two reactants are fed in correct amounts to assure the stoichiometry of the reaction. Figure 1 shows a flow sheet of the CRD process. The catalyst is fixed in the middle of the reactive distillation column; separation and reaction occur in the RD column at the same time; and then the light and heavy products H2O and AmAc are obtained from the top and bottom of the RD column, respectively. It is noticed that H2O is obtained by a decanter at the top of the column and the organic phase is recycled back to the column. The RD column is divided into two parts in the DPT-RD process, the rectifying and reactive section (HP column) and the stripping section (LP column). The corresponding flow sheet of the DPT-RD process is shown in Figure 2. The vapor stream of the LP column is compressed and transferred to the bottom of the HP column. The bottom liquid of the HP column is depressurized by a throttling valve and fed into the top of the LP column. The vapor stream from the HP column is used as the heat source of the LP column and then cooled by an auxiliary condenser. The effective heat integration between the two columns through the heat exchanger may result in energy savings. The feasibility and effectiveness of the DPT-RD process will be investigated and discussed in Section 3. 3. Process optimization TAC is used as the objective function in process optimization (Eq. 5) and the economic feasibility of the DPT-RD process is evaluated [19]. TAC ($/year) = OC + CI/T
（5）
where OC is the operation cost, CI is the capital investment, and T is the payback period with 3 years. Operating costs contain the utility cost (steam and cooling water), the catalyst cost, and the electricity cost for the compressor. Major capital costs include the cost for the reactive distillation column vessel (including column internals), heat exchangers, and the compressor. The cost of small items such as reflux drums, pumps, valves, and pipes is negligible compared to the major capital investment costs. The operating time is assumed to be 8,000 h/yr. We used the price of the steam and equipment size and costs recommended by Luyben [20]. 3.1 Process parameter analysis In the DPT-RD process, a number of design variables should be determined: the number of stages in the HP column (NHP), the number of stages in the LP column (NLP), the number of reactive trays (NRX), the HAc feed location (NFA), and AmOH feed location (NFB) of the HP column. As the main design parameters of the process, these will have a huge impact on the process and equipment investment. By recycling the organic phase back to the column by a decanter, the reflux ratio of the HP column is fixed under certain conditions because the reflux flow rate is determined by the phase equilibrium of the mixed components. The pressure in the HP column and LP column is also an important design variable and has a great influence on the product purity and energy requirements. 3.2 Sensitivity analysis Since pressure has a significant effect on the DPT-RD process, it is essential to maintain the
appropriate pressure in each column. Excessive temperature and pressure will reduce the service life of the catalyst; thus, the pressure of the HP column (PHP) remains at 1 bar on the basis of the CRD process. Figure 3 shows the temperature profiles of the DPT-RD process at different pressure values. The effect of the pressure of the LP column (PLP) on compressor duty and heat exchanger duty is illustrated in Figure 4. △T1 represents the minimum temperature difference between the two towers and △T2 represents the temperature difference between the top of the LP column and the bottom of the HP column. In Figures 3 and 4, both △T1 and △T2 increase as the pressure of the LP column decreases, meanwhile the heat exchanger cost reduction and the compressor cost increases; in contrast, the cost of the heat exchanger is increased and the cost of the compressor is reduced. Therefore, the minimum TAC can be obtained as a trade-off between the cost of the heat exchanger and the compressor, which will be discussed in Section 3.3. To ensure a normal heat transfer, the minimum heat transfer temperature difference is set at 10℃ in this system. 3.3 Economic optimization According to the analysis of the process parameters, NHP, NLP, NRX, NFA, NFB, and PLP should be optimized. Figure 5 reveals how the TAC is affected by NFA, NFB, NRX, NHP, NLP, and PLP. Figure 5 (a) shows that TAC is decreased with the decrease in NFA or the increase in NFB and TAC reaches a minimum value when NFA = 6 and NFB = 3. The feed position is closer to the top of the column or the spacing between the two feed locations is smaller resulting in incomplete reaction, which needs more vapor so that the operating cost of the process increases. Figure 5 (b) shows the variations in TAC with the increase in NRX; TAC has a minimum value at NRX = 22. The conversion of reaction is greatly affected by NRX. At NRX ˂ 22, the conversion rate reduced and more reboiler heat input is needed for separation, while at NRX > 22, the conversion rate increased slowly and the cost of the catalyst and tray is also increased. Thus, TAC has the optimum value at NRX = 22. Figure 5 (c) presents the effect of NHP on TAC. TAC reaches a minimum value on stage 24 with the increase in NHP and then it begins to increase. For this system, due to the presence of azeotrope, the separation of the light product H2O cannot be achieved by increasing the number of rectifying trays. The mixtures should also be separated further in the decanter. However, with the increase in NHP to over 24, the increase in equipment costs is higher than the reduction in operating costs; thus, TAC begins to increase. Figure 5 (d) shows the influence of PLP and NLP on TAC. The optimum values for NLP and PLP are 6 and 0.14 bar, respectively. Figure 4 displays that the minimum temperature difference is about 13.8℃ when the PLP is equal to 0.14 bar, which makes the heat exchange feasible. Based on the above optimization conditions, the detailed information of the DPT-RD process is presented in Figure 2. The DPT-RD process can achieve significant energy savings compared to the CRD process because the energy of the condenser in the HP column of the DPT-RD system is recovered through heat exchange with the reboiler of the LP column. Since the qualities of the heat and power energy are different, a unified measure of energy calculation should be adopted. The conversion coefficient of standard coal [21] is used as the benchmark for energy consumption and the energy efficiency is calculated by converting the energy consumption into standard coal consumption. The coal equivalent conversion coefficient and the calculation results of coal equivalent energy consumption are listed in Table 2. Approximately 33.74% of energy savings can be achieved in the DPT-RD process. In order to test the economic advantages of the DPT-RD process, the process parameters and TAC of the CRD and the DPT-RD processes were compared (Table 3). The feed conditions and
product specifications in the DPT-RD process are consistent with those in the CRD process to ensure a fair comparison. The comparison results show that DPT-RD can reduce TAC by 7.96%, despite a 23.86% increase in capital investment (high expenditure on the compressor), because of a 37.92% decrease in operating cost. 4. Control structure design Although the DPT-RD process can be implemented smoothly at the steady state, the “neat” flowsheet is difficult to control because the feed flow must be consistent with the stoichiometric ratio, which places a heavy demand on the control structure. The nonlinearity is increased and the control freedom is reduced because of the high degree of integration of the DPT-RD process, which increases the control difficulty of the RD system. As a result, it is of great significance to study the controllability of the heat-integrated system. In order to achieve the smooth operation of the DPT-RD dynamic process and ensure the desired product quality, the dynamic process was simulated by the Aspen Dynamics simulator. Two control schemes were developed for the DPT-RD process and their dynamic performances were investigated for the disturbances in feed flow rate and feed composition. 4.1 Basic control structure In the basic control structure, there are eight inventory control loops including the control of four liquid levels, two pressures, and two feeds. Besides, the two feed flow rates are proportional and the multiplier receives its signal from the temperature controller, which supervises the stage 1 temperature in the LP column. The controlled tray is chosen based on the slope criterion (Figure 6) as well as the selection of the temperature controlled by the tray of the HP column [22]. The basic control system configuration for the DPT-RD design flowsheet and the control loops are summarized in Table 4. Conventional proportional and integral (PI) controllers are used for all controllers except four liquid level controllers and the liquid level is controlled by proportional-only controllers with Kc = 2. For the pressure control loops and the feed flow loops, the gain and integral time are set at Kc = 20, τI = 12 min and Kc = 0.5, τI = 0.3 min, respectively. The dead time is set at 1 min and 3 min for the temperature and composition controllers. The ultimate gains and periods are obtained through the relay-feedback test and the Tyreus-Luyben tuning method is used for controller tuning. Figure 7 reveals the relevant dynamic responses when the DPT-RD process suffers the feed flow rate and composition disturbances in 2 h. The fresh feed composition changes from 100% to 95%, meaning that the feed composition is changed from pure material to 95 mol% AmOH/HAc and 5 mol% impurity H2O. The results show that two product purities can arrive at a new stable value and the value is close to their desired specifications when the DPT-RD process suffers the composition changes and the maximum offset is 0.15%. However, the AmAc purity shows a large offset (about 5%) for the desired specification that cannot be ignored in the industrial production process when the +20% feed flow rate disturbance was introduced. The deviation of the AmAc purity as a result of the proportion coefficient of the proportional control changes with the feed flow rate. Feed ratio has a large deviation from the stoichiometric ratio. Excess reactant (AmOH) is extracted from the bottom of the LP column with the AmAc product, which leads to the low purity of the AmAc product. The temperature controller only provides an estimate of the
composition, especially for the binary system because the temperature fixes the composition at constant pressure. However, this is no longer the case for a multivariate system such as the DPT-RD process that produces azeotropes (HAc/AmOH/AmAc, HAc/AmOH) when the +20% feed flow rate disturbance is introduced. The formation of azeotropes causes the composition of the temperature control tray to change so that the proportional control is no longer effective. Besides, the proportional control can manage the composition disturbance effectively (Figure 7 (b)) and the proportional control strategy allows the HAc feed flow to change accordingly to compensate the composition difference so that the stoichiometric ratio between AmOH and HAc is maintained. Hence, how to achieve the effective proportional control of the feed flow rate disturbances should be considered. 4.2 Improved control structure Because the basic control structure cannot handle the feed flow rate disturbances effectively, an improved control structure is proposed to control the AmAc composition by a cascade connection of temperature/composition cascade control and pressure control. It has an indirect impact on the feed proportion control due to the restriction of the two product components (the composition of the aqueous phase is determined by the decanter; the AmAc product composition is controlled by the temperature/composition cascade control); thus, the control effect can be improved effectively. The improved control structure for the DPT-RD process is presented in Figure 8. The dynamic performances of the improved control structure are also evaluated by feed flow rate and composition disturbances (Figure 9). The first fluctuation of the FHAc dynamic response curve (lower left of Fig. 9(a)) shows that for a +20% feed flow rate disturbance, FHAc does not reach the required amount of feed at the peak, which is similar to the case of the basic control structure. However, as a result of the addition of the temperature/composition cascade control, timely feedback is made after the purity of AmAc drops and the effect is passed on to the HP column by means of pressure regulation. This has an indirect effect on the feed proportion control. The feed flow rate of HAc then begins to increase and is stabilized after some oscillation, which is consistent with the feed flow rate of AmOH. Therefore, the entire dynamic process becomes smooth and product purities are maintained. The results show that the AmAc product composition can always come back to its original value and two products can be maintained very close to their desired purities when the system undergoes different feed disturbances. This shows that the temperature/composition cascade control is effective in this improved control structure. In addition, the control effect of the feed proportion is improved and the feed flow rate changes accordingly to compensate the feed flow and the composition changes so that the progress of the reaction and the purity of the product are ensured. On the other hand, the settling time is shorter than that of the basic control structure, meaning that the DPT-RD process can stabilize quickly in the face of large feed disturbances. Given the above, the results indicate that the DPT-RD process robust control can be achieved under the improved control structure and it can maintain the products at their desired purities with different feed disturbances. 5. Conclusions In this study, design and control of the DPT-RD process for AmAc synthesis in a neat operation
mode were investigated and an optimal design in terms of TAC was presented. A number of parameters (NHP, NLP, NRX, NFA, NFB, and PLP) were optimized and the optimal results of DPT-RD were obtained by minimizing the TAC of the process. The simulation results show that savings of 33.74% in energy consumption can be achieved compared to the CRD process. TAC decreased only by 7.96% due to the high cost of the compressor. Two control structures (basic and improved control structure) of the DPT-RD process were established using the Aspen Dynamics simulator. For the basic control structure, the DPT-RD process can handle composition disturbances well but the AmAc purity greatly deviated for the desired specification when the feed flow rate disturbances were introduced. Furthermore, an improved control structure with temperature/composition cascade control was explored to get a more robust control. The dynamic simulation results show that the improved control structure can provide good dynamic controllability for the feed flow rate and feed composition disturbances, which demonstrates that the DPT-RD process in a neat operation mode can be controlled effectively with reasonable control. In summary, this study solved the energy-saving optimization and control problem of the DPT-RD process for AmAc synthesis. The DPT-RD process can also be used in other RD systems involving reactions with negligible or no thermal effects and can provide theoretical guidance for these types of systems in industrial applications. Acknowledgment The authors are thankful for support by the Natural Science Foundation of Shandong Province (No. ZR2013BM001).
Reference [1] A Orjuela, M A Santaella, P A Molano, Process Intensification by Reactive Distillation, Springer International Publishing. (2016) 131-181. [2] D.B. Kaymak, W.L. Luyben, Quantitative comparison of reactive distillation with conventional multiunit reactor/column/recycle systems for different chemical equilibrium constants, Ind. Eng. Chem. Res. 43 (10) (2004) 2493-2507. [3] J. Sun, K. Huang, S. Wang. Deepening Internal Mass Integration in Design of Reactive Distillation Columns, 1: Principle and Procedure, Ind. Eng. Chem. Res. 48(4) (2009) 2034-2048. [4] X. Suo, Q. Ye, R. Li, X. Dai, H. Yu, The partial heat-integrated pressure-swing reactive distillation process for transesterification of methyl acetate with isopropanol, Chem. Eng. Process. 107 (2016) 42-57. [5] T.L. Hsiao, K.C. Weng, H.Y. Lee, Design and control of hybrid heat-integrated configuration for an ideal indirect reactive distillation process, J. Taiwan. Inst. Chem. Eng. 000 (2016) 1-13. [6] J.R. Alcántara-Avila, M. Terasaki, H.Y. Lee, J.L. Chen, K.I. Sotowa, T. Horikawa, Design and control of reactive distillation sequences with heat-integrated stages to produce diphenyl carbonate, Ind. Eng. Chem. Res. 56 (2017) 250-260. [7] S.V. Mali, A.K. Jana, A partially heat integrated reactive distillation: feasibility and analysis, Sep. Purif. Technol. 70 (1) (2009) 136-139. [8] S. Hernández, R. Sandoval-Vergara, F.O. Barroso-Muñz, R. Murrieta-Dueñs, H. Hernández-Escoto, J.G. Segovia-Hern á ndez, V. Rico-Ramirez, Reactive dividing wall distillation columns: simulation and implementation in a pilot plant, Chem. Eng. Process. 48 (2009) 250-258. [9] A. Vanaki, R. Eslamloueyan, Steady-state simulation of a reactive internally heat integrated distillation column (R-HIDiC) for synthesis of tertiary-amyl methyl ether (TAME), Chem. Eng. Process. 52 (2012) 21-27. [10] H.Y. Lee, Y.C. Lee, I.L. Chien, H.P. Huang, Design and control of a heat-integrated reactive distillation system for the hydrolysis of methyl acetate. Ind. Eng. Chem. Res. 49 (2010) 7398-7411. [11] N.V.D. Long, Q.M. Le, C.N. Le, M. Lee, A novel self-heat recuperative dividing wall column to maximize energy efficiency and column throughput in retrofitting and debottlenecking of a side stream column[J]. Appl. Energy. 159 (2015) 28-38. [12] Y. Liu, J. Zhai, L. Li, L. Sun, C. Zhai, Heat Pump Assisted Reactive and Azeotropic distillations in Dividing Wall Columns, Chem. Eng. Process. 95 (2015) 289-301. [13] K Huang, Q Lin, H Shao, C Wang, S Wang, A fundamental principle and systematic procedures for process intensification in reactive distillation columns, Chem. Eng. Process. 49 (3) (2010) 294-311. [14] X. Gao, F. Wang, H. Li, X. Li, Heat-integrated reactive distillation process for TAME synthesis, Sep. Purif. Technol. 132 (2014) 468-478. [15] W.L. Luyben, Economic and dynamic impact of the use of excess reactant in reactive distillation systems, Ind. Eng. Chem. Res. 39 (8) (2000) 2935-2946. [16] W.L. Luyben, C.C. Yu . Reactive distillation design and control, John Wiley & Sons, New Jersey, 2008. [17] M.J. Lee, H.T. Wu, C.H. Kang, H.M. Lin, Kinetic behavior of amyl acetate synthesis catalyzed by acidic cation exchange resin, J. Chinese. Inst. Chem. Eng. 30 (2) (1999) 117-122. [18] S.F. Chiang, C.L. Kuo, C.C. Yu, D.S.H. Wong, Design alternatives for the amyl acetate process: coupled reactor/column and reactive distillation, Ind. Eng. Chem. Res. 41 (13) (2002) 3233-3246. [19] J.M. Douglas, Conceptual Design of Chemical Process, McGraw-Hill, New York, 1988.
[20] W.L. Luyben, Principles and case studies of simultaneous design, John Wiley & Sons, 2012. [21] General principles for calculation of total production energy consumption, GB/T 2589-2008. [22] W.L. Luyben, Distillation Design and Control Using Aspen Simulation, John Wiley & Sons, 2013.
89.8°C, 1bar Qc=-888.6kW
AmOH 50kmol/h 25°C, 1.5bar
F=50kmol/h 50°C, 1bar HAc=0.003 AmOH=0.005 AmAc=314ppm H2O=0.991
Organic Reflux
HAc 50kmol/h 25°C, 1.5bar
NR=2 NRX=22 NS=12
Qr=1009.4kW
F=50kmol/h 149.1°C, 1.245bar HAc=0.005 AmOH=0.004 AmAc=0.991 H2O=trace
Fig.1. Conventional reactive distillation process for AmAc synthesis
125.1°C, 1.156bar
Compressor duty=263.0kW
4 98.9°C, 1bar
PHP=1bar
7 PLP=0.14bar 1
Organic Reflux
AmOH 50kmol/h 25°C, 1.5bar
Qc=-393.823kW
HAc 50kmol/h 25°C, 1.5bar
2
6 F=50kmol/h 50°C, 1bar XD,HAc=0.003 Qr=425.935kW XD,AmOH=0.005 XD,H2O=0.991 5 XD,AmAc=325ppm 85.1°C 0.165bar 3
H2O
F=50kmol/h 85.3°C, 1bar XB,HAc=0.005 XB,AmOH=0.004 XB,H2O=trace XB,AmAc=0.991 AmAc
78.7°C, 0.14bar
Fig.2. Different pressure thermally coupled reactive distillation process for AmAc synthesis HP column LP column 0.14bar LP column 0.15bar LP column 0.10bar
140
Temperature (℃)
130 120
△T2
110 100 △T1
90 80 70
0
5
10
15
20
25
30
Number of stage
Fig.3. Temperature profile of DPT-RD process at different pressures
Compressor duty (kw)
Heat exchanger duty (kw)
450 Compressor Heat exchanger
340
420
320 390
300 280
360
260
330
240
0.08
0.10
0.12 PLP(bar)
0.14
300 0.16
Fig.4. Effect of the pressure of the LP column on compressor duty and heat exchanger duty (a)
(b) 629
760 NFB=2
TAC (10 $/year)
NFB=4
3
680
3
TAC (10 $/year)
628
NFB=3
720
640
6
7 8 9 NFA: feed location of HAc
18
20
22
24
26
NRX: number of reactive trays (d) 660 NLP=12
611
NLP=9
640
610
TAC (10 $/year)
609
3
3
625
10
(c) 612
TAC (10 $/year)
626
624
600
608 607 606 605
627
NLP=6
620 600 580 560 0.08
23 24 25 26 27 NHP: number of stages in HP column
0.10
0.12
0.14
0.16
PLP (bar)
Fig.5. Effect of (a) NFA and NFB (b) NRX (c) NHP (d) NLP and PLP on TAC 5
LP column
HP column 4
1.3
ΔT (℃)
ΔT (℃)
1.4
1.2
3 2 1
1.1 1
2
3
4
Number of stage
5
6
0
0
5
10
15
Number of stage
Fig.6. Temperature slope profiles in the LP column and HP column (a)
20
25
70
1.00
+20% F -20% F
XAmAc (mol%)
FAmAc (kmol/h)
60 50 40
+20% F -20% F
0.98
0.96
0.94 30
0
5
10
15
0
20
5
10
+20% F -20% F
50
40
30
0
5
10
15
20
0.990
0.988
0.986
0
5
10
+20% F -20% F
TLP, stage 1 (℃)
FHAc feed (kmol/h)
40
0
5
10
20
15
+20% F -20% F
79.6
50
30
15
Time (h)
Time (h)
60
20
+20% F -20% F
0.992
XH2O (mol%)
FH2O (kmol/h)
60
15
Time (h)
Time (h)
79.2 78.8 78.4 78.0
20
0
5
10
15
20
Time (h)
Time (h)
Fig.7. Dynamic responses of the basic control structure in (a) ±20% feed flow rate (b) 5% composition changes (b) 0.9930 0.9925 95mol% AmOH 95mol% HAc
49
48
XAmAc (mol%)
FAmAc (kmol/h)
50
95mol% AmOH 95mol% HAc
0.9920 0.9915 0.9910 0.9905
47
0
5
10 Time (h)
15
20
0
5
10 Time (h)
15
20
54 95mol% AmOH 95mol% HAc
XH2O (mol%)
FH2O (kmol/h)
53
95mol% AmOH 95mol% HAc
0.9930
52 51
0.9925 0.9920 0.9915
50 0
5
10
15
20
0.9910
0
5
54
79.6
50 48
0
5
10
20
15
95mol% AmOH 95mol% HAc
79.2
95mol% AmOH 95mol% HAc
TLP, stage 1 (℃)
FHAc feed (kmol/h)
52
15
Time (h)
Time (h)
46
10
20
Time (h)
78.8 78.4 78.0
0
5
10
15
20
Time (h)
Fig.7. Dynamic responses of the basic control structure in (a) ±20% feed flow rate (b) 5% composition changes (continued)
Fig.8. Improved control structure for the DPT-RD
(a) 65
0.996 +20% F -20% F
+20% F -20% F
55
0.992
XAmAc (mol%)
FAmAc (kmol/h)
60
50 45
0.988
0.984
40 0
5
10
15
20
0
5
Time (h)
Time (h)
70
0.996
+20% F -20% F
+20% F -20% F 0.992
XH2O (mol%)
60
FH2O (kmol/h)
20
15
10
50
0.988 0.984
40 0.980
0
5
10
15
20
0
5
Time (h)
10
15
20
Time (h)
Fig.9. Dynamic responses of the improved control structure in (a) ±20% feed flow rate (b) 5% composition changes 80.0 +20% F -20% F
65
+20% F -20% F 79.5
TLP, stage 1 (℃)
FHAc feed (kmol/h)
60 55 50 45
79.0 78.5
40 35
0
5
10
15
78.0
20
0
5
10
15
20
Time (h)
Time (h)
(b) 0.9920
95mol% AmOH 95mol% HAc
95mol% AmOH 95mol% HAc
XAmAc (mol%)
FAmAc (kmol/h)
50
49
48
0.9916
0.9912
0.9908 47
0
5
10 Time (h)
15
20
0
5
10 Time (h)
15
20
0.9930
95mol% AmOH 95mol% HAc
53
XH2O (mol%)
FH2O (kmol/h)
0.9925 52 51 50
95mol% AmOH 95mol% HAc
0.9920 0.9915 0.9910
0
5
10
15
0
20
5
10
15
20
Time (h)
Time (h)
79.6
54
95mol% AmOH 95mol% HAc 79.2
95mol% AmOH 95mol% HAc
TLP, stage 1 (℃)
FHAc feed (kmol/h)
52 50 48 46
0
5
10 Time (h)
15
20
78.8 78.4 78.0
0
5
10
15
20
Time (h)
Fig.9. Dynamic responses of the improved control structure in (a) ±20% feed flow rate (b) 5% composition changes (continued)
Tables and Figures
Table 1. Activity coefficient models parameters for AmAc system Comp.i
HAc
HAc
HAc
AmOH
AmOH
AmAc
Comp.j
AmOH
AmAc
H2O
AmAc
H2O
H2O
bij(K)
-316.8
-37.943
-110.57
-144.8
100.1
254.47
bji(K)
178.3
214.55
424.018
320.6521
1447.5
2221.5
cij
0.1695
0.2000
0.2987
0.3009
0.2980
0.2000
Table 2. Coal equivalent (CE) conversion coefficient and the CE energy consumption Energy consumption on CE (kg·h-1) DPT-RD
CRD
Vapor consumption
0.1286kg·kg-1
124.19
(reboiler) Standard coal
Cooling
water
equivalent
consumption
coefficient
(condenser)
0.4857kg·t-1
48.39
0.404kg·(kw·h)-1
106.25
109.19
Electricity consumption (compressor) Total (% difference)
154.64 (-33.74)
233.38 (0)
Table 3. Comparison of CRD and DPT-RD processes DPT-RD
parameter
CRD
HP column
LP column
RD column
Number of stages
24
6
37
Reactive stages
3-24
3-24
AmOH feed stage
3
3
HAc feed stage
6
7
Feed flow rate (kmol/h)
50
50
Operating pressure (bar)
1
0.14
1
Column diameter (m)
0.997
1.702
1.028
Operating cost Capital cost
(103$/year)
(103$/year)
Total annual cost
(103$/year)
186.95
301.12
351.18
283.54
538.13
584.66
Table 4. Control loops of multi-variable control Control circuits
Controlled variables
Control objectives
Manipulated variables
LC1
base leaves in HP column
50%
flow-out of HP column bottom
LC2
base leaves in LP column
50%
flow-out of LP column bottom
LC3
organic phase leaves in decanter
50%
flow-out of organic phase
LC4
aqueous phase leaves in decanter
50%
flow-out of aqueous phase
PC1
overhead pressure in HP column
1 bar
PC2
overhead pressure in LP column
0.14 bar
FC1
feed rate of AmOH
50 kmol·h
FC2
feed rate of HAc
50 kmol·h
TC1
temperature of stage 1 in LP column
78.87℃
flow of vapor from the top of the HP column brake power of compressor −1 −1
feed rate of AmOH feed rate of HAc ratio of the feed streams
S0255-2701(17)30350-1 http://dx.doi.org/10.1016/j.cep.2017.09.002 CEP 7066
To appear in:
Chemical Engineering and Processing
Received date: Revised date: Accepted date:
10-4-2017 28-7-2017 4-9-2017
Please cite this article as: Qingrui Zhang, Tong Guo, Chao Yu, Yonglei Li, Design and control of Different Pressure Thermally Coupled Reactive Distillation for Amyl Acetate synthesis, Chemical Engineering and Processinghttp://dx.doi.org/10.1016/j.cep.2017.09.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Design and control of Different Pressure Thermally Coupled Reactive Distillation for Amyl Acetate synthesis Qingrui Zhang*, Tong Guo, Chao Yu, Yonglei Li
College of Chemical Engineering, Qingdao University of Science and Technology, Qingdao 266042, China
*
Corresponding author.
E-mail address: [email protected] (Q. Zhang).
Graphical abstract:
Fig.1. Improved control structure for the different pressure thermally coupled reactive distillation The control strategy of different pressure thermally coupled reactive distillation process was investigated. The improved control structure is proposed to control the AmAc composition by a cascade connection of temperature/composition cascade control and pressure control, and the dynamic performance indicate that the different pressure thermally coupled reactive distillation process robust control can be achieved under the improved control structure.
Highlights：
The DPT-RD process was developed for the AmAc esterification process. The feasibility and effectiveness for the DPT-RD process were investigated. The control strategy of the DPT-RD process was developed and studied.
The results of this study provide theoretical guidance for industrial applications.
Abstract: The different pressure thermally coupled reactive distillation (DPT-RD) process is an attractive heat-integrated design in chemical engineering development processes. In this study, a rigorous simulation of the DPT-RD process for amyl acetate synthesis in a neat operation mode is performed using the Aspen Plus simulator. The economic feasibility and controllability of the DPT-RD process were evaluated based on the research of steady-state optimization and dynamic control. The optimum DPT-RD process in terms of the total annual cost (TAC) is screened. The results show that 33.74% reduction in energy consumption and 7.96% savings in TAC can be achieved by the DPT-RD process compared to the conventional reactive distillation. The control strategy of the DPT-RD process was developed and the basic control structure with proportional control and the improved control structure with temperature/composition cascade control were presented. The dynamic results indicate that the improved control structure can effectively handle disturbances of ±20% in feed flow rate and of 5% in feed composition. Overall results demonstrate that the DPT-RD process is a promising approach for amyl acetate synthesis. Keywords: Amyl acetate; Different pressure thermally coupled reactive distillation; Process optimization; Control structures 1.Introduction With the increasing demand for energy, process intensification has attracted attention of the industry and research community to reduce energy consumption and capital investment. Reactive distillation (RD) combines reaction and separation in a single unit attributed to the potential advantages over the catalyst usage, reactive heat utilization, and cost savings [1,2]. RD has been extensively studied for decades, but it has received renewed attention in recent years due to the novel research trends about process development with higher degrees of intensification. In terms of thermal effects of the reaction, RD systems can be divided into three broad categories: those with high thermal effect, those with considerable thermal effect, and those with negligible or no thermal effect [3]. Research on heat-integrated technology has mainly focused on the RD systems involving reactions with high thermal effect and produced remarkable results [4-7]. Hernandez et al. investigated the design and control strategies of reactive dividing wall column process for the production of ethyl acetate and verified the energy saving effect [8]. The synthesis of tert-amyl methyl ether (TAME) by heat-integrated reactive distillation column was investigated by Vanaki and co-workers and the simulation results showed that 22% of energy savings could be achieved compared to the conventional reactive distillation (CRD) column [9]. Other forms of thermal integration have also been studied such as multi-effect reactive distillation and heat pump-assisted reactive distillation [10-12]. However, previous studies on the RD systems involving reactions with negligible or no thermal effects have only focused on internal mass integration [3,13] and research on energy savings of the system is scarce. For these systems, the effect of internal heat integration seems much weaker than that of internal mass integration; therefore, other thermal integration methods such as external heat integration can be considered. Gao and co-workers proposed an interesting structure of different pressure thermally coupled
reactive distillation (DPT-RD) for TAME synthesis [14]. The DPT-RD process is a highly coupled process of reaction, separation, and energy integration. The connection between the high-pressure (HP) column and the low-pressure (LP) column is realized by a compressor and a throttle valve, which doesn’t have a significant impact on the original material coupling. The top stream of the HP column is used as the heating medium of the LP column reboiler; thus, the DPT-RD process has great potential for energy savings through this type of external heat integration. The DPT-RD process simultaneously achieved the material coupling and heat integration of the RD systems involving reactions with negligible or no thermal effect; thus, it will be a good choice for energy-saving improvements on the RD systems. The RD process operated in the “neat” mode is economical because the one-column process has lower capital investment and energy costs than the two-column system [15] that enhances the advantages of the DPT-RD process in terms of investment. However, the control of the DPT-RD process operated in neat mode is difficult to achieve. The control of the neat RD process must maintain the precise balance of the fresh feed streams because the imbalance will inevitably result in a gradual buildup of one of the reactants and a loss of conversion and product purities [16]. On the other hand, the controllability and effectiveness of this process is a large challenge because of the high degree of coupling and nonlinearity of the DPT-RD process. Therefore, it is essential to study the control strategy of the DPT-RD process. The main purpose of this study is to investigate the feasibility and effectiveness of further heat integration for the RD systems involving reactions with negligible or no thermal effects by the DPT-RD process and develop a control strategy. Taking the amyl acetate (AmAc) esterification process as an example, the steady state optimization and dynamic control of the DPT-RD process are studied. The optimum design of the steady state process is obtained based on the total annual cost (TAC). The economic comparison between the DPT-RD and CRD processes in optimal conditions is also presented. Two control structures are proposed to achieve the robust control of the DPT-RD process and maintain the purity of products for different feed disturbances. Their dynamic responses are also analysed and discussed. In this work, both steady state optimization and dynamic control are carried out with Aspen Plus and Aspen Dynamics (version 7.2). 2. Process description In this study, the esterification of AmAc was carried out in the presence of an acid ion-exchange polymeric resin, Amberlyst 15. The second-order reversible reaction can be described by the following expression (1): C7H14O2 H 2O CH3COOH C5H12O
（1）
The reaction kinetics can be expressed as follows [17]: r mcat (kFCHAcCAmOH kBCAmAcCH2O)
（2）
kF 31.1667 exp(
51740 ) RT
（3）
kB 2.25533exp(
45280 ) RT
（4）
There are many types of azeotropes in the AmOH-HAc-AmAc-H2O quaternary system and the NRTL activity coefficient model is used to predict the phase equilibrium of these four components.
Furthermore, we adopt the Hayden-O’Connell model to account for the vapor non-ideality because HAc is associated in the vapor phase. The parameters and coefficients in phase equilibrium calculations are listed in Table 1 [18]. In the reactive distillation process for the AmAc synthesis, the two reactants are fed in correct amounts to assure the stoichiometry of the reaction. Figure 1 shows a flow sheet of the CRD process. The catalyst is fixed in the middle of the reactive distillation column; separation and reaction occur in the RD column at the same time; and then the light and heavy products H2O and AmAc are obtained from the top and bottom of the RD column, respectively. It is noticed that H2O is obtained by a decanter at the top of the column and the organic phase is recycled back to the column. The RD column is divided into two parts in the DPT-RD process, the rectifying and reactive section (HP column) and the stripping section (LP column). The corresponding flow sheet of the DPT-RD process is shown in Figure 2. The vapor stream of the LP column is compressed and transferred to the bottom of the HP column. The bottom liquid of the HP column is depressurized by a throttling valve and fed into the top of the LP column. The vapor stream from the HP column is used as the heat source of the LP column and then cooled by an auxiliary condenser. The effective heat integration between the two columns through the heat exchanger may result in energy savings. The feasibility and effectiveness of the DPT-RD process will be investigated and discussed in Section 3. 3. Process optimization TAC is used as the objective function in process optimization (Eq. 5) and the economic feasibility of the DPT-RD process is evaluated [19]. TAC ($/year) = OC + CI/T
（5）
where OC is the operation cost, CI is the capital investment, and T is the payback period with 3 years. Operating costs contain the utility cost (steam and cooling water), the catalyst cost, and the electricity cost for the compressor. Major capital costs include the cost for the reactive distillation column vessel (including column internals), heat exchangers, and the compressor. The cost of small items such as reflux drums, pumps, valves, and pipes is negligible compared to the major capital investment costs. The operating time is assumed to be 8,000 h/yr. We used the price of the steam and equipment size and costs recommended by Luyben [20]. 3.1 Process parameter analysis In the DPT-RD process, a number of design variables should be determined: the number of stages in the HP column (NHP), the number of stages in the LP column (NLP), the number of reactive trays (NRX), the HAc feed location (NFA), and AmOH feed location (NFB) of the HP column. As the main design parameters of the process, these will have a huge impact on the process and equipment investment. By recycling the organic phase back to the column by a decanter, the reflux ratio of the HP column is fixed under certain conditions because the reflux flow rate is determined by the phase equilibrium of the mixed components. The pressure in the HP column and LP column is also an important design variable and has a great influence on the product purity and energy requirements. 3.2 Sensitivity analysis Since pressure has a significant effect on the DPT-RD process, it is essential to maintain the
appropriate pressure in each column. Excessive temperature and pressure will reduce the service life of the catalyst; thus, the pressure of the HP column (PHP) remains at 1 bar on the basis of the CRD process. Figure 3 shows the temperature profiles of the DPT-RD process at different pressure values. The effect of the pressure of the LP column (PLP) on compressor duty and heat exchanger duty is illustrated in Figure 4. △T1 represents the minimum temperature difference between the two towers and △T2 represents the temperature difference between the top of the LP column and the bottom of the HP column. In Figures 3 and 4, both △T1 and △T2 increase as the pressure of the LP column decreases, meanwhile the heat exchanger cost reduction and the compressor cost increases; in contrast, the cost of the heat exchanger is increased and the cost of the compressor is reduced. Therefore, the minimum TAC can be obtained as a trade-off between the cost of the heat exchanger and the compressor, which will be discussed in Section 3.3. To ensure a normal heat transfer, the minimum heat transfer temperature difference is set at 10℃ in this system. 3.3 Economic optimization According to the analysis of the process parameters, NHP, NLP, NRX, NFA, NFB, and PLP should be optimized. Figure 5 reveals how the TAC is affected by NFA, NFB, NRX, NHP, NLP, and PLP. Figure 5 (a) shows that TAC is decreased with the decrease in NFA or the increase in NFB and TAC reaches a minimum value when NFA = 6 and NFB = 3. The feed position is closer to the top of the column or the spacing between the two feed locations is smaller resulting in incomplete reaction, which needs more vapor so that the operating cost of the process increases. Figure 5 (b) shows the variations in TAC with the increase in NRX; TAC has a minimum value at NRX = 22. The conversion of reaction is greatly affected by NRX. At NRX ˂ 22, the conversion rate reduced and more reboiler heat input is needed for separation, while at NRX > 22, the conversion rate increased slowly and the cost of the catalyst and tray is also increased. Thus, TAC has the optimum value at NRX = 22. Figure 5 (c) presents the effect of NHP on TAC. TAC reaches a minimum value on stage 24 with the increase in NHP and then it begins to increase. For this system, due to the presence of azeotrope, the separation of the light product H2O cannot be achieved by increasing the number of rectifying trays. The mixtures should also be separated further in the decanter. However, with the increase in NHP to over 24, the increase in equipment costs is higher than the reduction in operating costs; thus, TAC begins to increase. Figure 5 (d) shows the influence of PLP and NLP on TAC. The optimum values for NLP and PLP are 6 and 0.14 bar, respectively. Figure 4 displays that the minimum temperature difference is about 13.8℃ when the PLP is equal to 0.14 bar, which makes the heat exchange feasible. Based on the above optimization conditions, the detailed information of the DPT-RD process is presented in Figure 2. The DPT-RD process can achieve significant energy savings compared to the CRD process because the energy of the condenser in the HP column of the DPT-RD system is recovered through heat exchange with the reboiler of the LP column. Since the qualities of the heat and power energy are different, a unified measure of energy calculation should be adopted. The conversion coefficient of standard coal [21] is used as the benchmark for energy consumption and the energy efficiency is calculated by converting the energy consumption into standard coal consumption. The coal equivalent conversion coefficient and the calculation results of coal equivalent energy consumption are listed in Table 2. Approximately 33.74% of energy savings can be achieved in the DPT-RD process. In order to test the economic advantages of the DPT-RD process, the process parameters and TAC of the CRD and the DPT-RD processes were compared (Table 3). The feed conditions and
product specifications in the DPT-RD process are consistent with those in the CRD process to ensure a fair comparison. The comparison results show that DPT-RD can reduce TAC by 7.96%, despite a 23.86% increase in capital investment (high expenditure on the compressor), because of a 37.92% decrease in operating cost. 4. Control structure design Although the DPT-RD process can be implemented smoothly at the steady state, the “neat” flowsheet is difficult to control because the feed flow must be consistent with the stoichiometric ratio, which places a heavy demand on the control structure. The nonlinearity is increased and the control freedom is reduced because of the high degree of integration of the DPT-RD process, which increases the control difficulty of the RD system. As a result, it is of great significance to study the controllability of the heat-integrated system. In order to achieve the smooth operation of the DPT-RD dynamic process and ensure the desired product quality, the dynamic process was simulated by the Aspen Dynamics simulator. Two control schemes were developed for the DPT-RD process and their dynamic performances were investigated for the disturbances in feed flow rate and feed composition. 4.1 Basic control structure In the basic control structure, there are eight inventory control loops including the control of four liquid levels, two pressures, and two feeds. Besides, the two feed flow rates are proportional and the multiplier receives its signal from the temperature controller, which supervises the stage 1 temperature in the LP column. The controlled tray is chosen based on the slope criterion (Figure 6) as well as the selection of the temperature controlled by the tray of the HP column [22]. The basic control system configuration for the DPT-RD design flowsheet and the control loops are summarized in Table 4. Conventional proportional and integral (PI) controllers are used for all controllers except four liquid level controllers and the liquid level is controlled by proportional-only controllers with Kc = 2. For the pressure control loops and the feed flow loops, the gain and integral time are set at Kc = 20, τI = 12 min and Kc = 0.5, τI = 0.3 min, respectively. The dead time is set at 1 min and 3 min for the temperature and composition controllers. The ultimate gains and periods are obtained through the relay-feedback test and the Tyreus-Luyben tuning method is used for controller tuning. Figure 7 reveals the relevant dynamic responses when the DPT-RD process suffers the feed flow rate and composition disturbances in 2 h. The fresh feed composition changes from 100% to 95%, meaning that the feed composition is changed from pure material to 95 mol% AmOH/HAc and 5 mol% impurity H2O. The results show that two product purities can arrive at a new stable value and the value is close to their desired specifications when the DPT-RD process suffers the composition changes and the maximum offset is 0.15%. However, the AmAc purity shows a large offset (about 5%) for the desired specification that cannot be ignored in the industrial production process when the +20% feed flow rate disturbance was introduced. The deviation of the AmAc purity as a result of the proportion coefficient of the proportional control changes with the feed flow rate. Feed ratio has a large deviation from the stoichiometric ratio. Excess reactant (AmOH) is extracted from the bottom of the LP column with the AmAc product, which leads to the low purity of the AmAc product. The temperature controller only provides an estimate of the
composition, especially for the binary system because the temperature fixes the composition at constant pressure. However, this is no longer the case for a multivariate system such as the DPT-RD process that produces azeotropes (HAc/AmOH/AmAc, HAc/AmOH) when the +20% feed flow rate disturbance is introduced. The formation of azeotropes causes the composition of the temperature control tray to change so that the proportional control is no longer effective. Besides, the proportional control can manage the composition disturbance effectively (Figure 7 (b)) and the proportional control strategy allows the HAc feed flow to change accordingly to compensate the composition difference so that the stoichiometric ratio between AmOH and HAc is maintained. Hence, how to achieve the effective proportional control of the feed flow rate disturbances should be considered. 4.2 Improved control structure Because the basic control structure cannot handle the feed flow rate disturbances effectively, an improved control structure is proposed to control the AmAc composition by a cascade connection of temperature/composition cascade control and pressure control. It has an indirect impact on the feed proportion control due to the restriction of the two product components (the composition of the aqueous phase is determined by the decanter; the AmAc product composition is controlled by the temperature/composition cascade control); thus, the control effect can be improved effectively. The improved control structure for the DPT-RD process is presented in Figure 8. The dynamic performances of the improved control structure are also evaluated by feed flow rate and composition disturbances (Figure 9). The first fluctuation of the FHAc dynamic response curve (lower left of Fig. 9(a)) shows that for a +20% feed flow rate disturbance, FHAc does not reach the required amount of feed at the peak, which is similar to the case of the basic control structure. However, as a result of the addition of the temperature/composition cascade control, timely feedback is made after the purity of AmAc drops and the effect is passed on to the HP column by means of pressure regulation. This has an indirect effect on the feed proportion control. The feed flow rate of HAc then begins to increase and is stabilized after some oscillation, which is consistent with the feed flow rate of AmOH. Therefore, the entire dynamic process becomes smooth and product purities are maintained. The results show that the AmAc product composition can always come back to its original value and two products can be maintained very close to their desired purities when the system undergoes different feed disturbances. This shows that the temperature/composition cascade control is effective in this improved control structure. In addition, the control effect of the feed proportion is improved and the feed flow rate changes accordingly to compensate the feed flow and the composition changes so that the progress of the reaction and the purity of the product are ensured. On the other hand, the settling time is shorter than that of the basic control structure, meaning that the DPT-RD process can stabilize quickly in the face of large feed disturbances. Given the above, the results indicate that the DPT-RD process robust control can be achieved under the improved control structure and it can maintain the products at their desired purities with different feed disturbances. 5. Conclusions In this study, design and control of the DPT-RD process for AmAc synthesis in a neat operation
mode were investigated and an optimal design in terms of TAC was presented. A number of parameters (NHP, NLP, NRX, NFA, NFB, and PLP) were optimized and the optimal results of DPT-RD were obtained by minimizing the TAC of the process. The simulation results show that savings of 33.74% in energy consumption can be achieved compared to the CRD process. TAC decreased only by 7.96% due to the high cost of the compressor. Two control structures (basic and improved control structure) of the DPT-RD process were established using the Aspen Dynamics simulator. For the basic control structure, the DPT-RD process can handle composition disturbances well but the AmAc purity greatly deviated for the desired specification when the feed flow rate disturbances were introduced. Furthermore, an improved control structure with temperature/composition cascade control was explored to get a more robust control. The dynamic simulation results show that the improved control structure can provide good dynamic controllability for the feed flow rate and feed composition disturbances, which demonstrates that the DPT-RD process in a neat operation mode can be controlled effectively with reasonable control. In summary, this study solved the energy-saving optimization and control problem of the DPT-RD process for AmAc synthesis. The DPT-RD process can also be used in other RD systems involving reactions with negligible or no thermal effects and can provide theoretical guidance for these types of systems in industrial applications. Acknowledgment The authors are thankful for support by the Natural Science Foundation of Shandong Province (No. ZR2013BM001).
Reference [1] A Orjuela, M A Santaella, P A Molano, Process Intensification by Reactive Distillation, Springer International Publishing. (2016) 131-181. [2] D.B. Kaymak, W.L. Luyben, Quantitative comparison of reactive distillation with conventional multiunit reactor/column/recycle systems for different chemical equilibrium constants, Ind. Eng. Chem. Res. 43 (10) (2004) 2493-2507. [3] J. Sun, K. Huang, S. Wang. Deepening Internal Mass Integration in Design of Reactive Distillation Columns, 1: Principle and Procedure, Ind. Eng. Chem. Res. 48(4) (2009) 2034-2048. [4] X. Suo, Q. Ye, R. Li, X. Dai, H. Yu, The partial heat-integrated pressure-swing reactive distillation process for transesterification of methyl acetate with isopropanol, Chem. Eng. Process. 107 (2016) 42-57. [5] T.L. Hsiao, K.C. Weng, H.Y. Lee, Design and control of hybrid heat-integrated configuration for an ideal indirect reactive distillation process, J. Taiwan. Inst. Chem. Eng. 000 (2016) 1-13. [6] J.R. Alcántara-Avila, M. Terasaki, H.Y. Lee, J.L. Chen, K.I. Sotowa, T. Horikawa, Design and control of reactive distillation sequences with heat-integrated stages to produce diphenyl carbonate, Ind. Eng. Chem. Res. 56 (2017) 250-260. [7] S.V. Mali, A.K. Jana, A partially heat integrated reactive distillation: feasibility and analysis, Sep. Purif. Technol. 70 (1) (2009) 136-139. [8] S. Hernández, R. Sandoval-Vergara, F.O. Barroso-Muñz, R. Murrieta-Dueñs, H. Hernández-Escoto, J.G. Segovia-Hern á ndez, V. Rico-Ramirez, Reactive dividing wall distillation columns: simulation and implementation in a pilot plant, Chem. Eng. Process. 48 (2009) 250-258. [9] A. Vanaki, R. Eslamloueyan, Steady-state simulation of a reactive internally heat integrated distillation column (R-HIDiC) for synthesis of tertiary-amyl methyl ether (TAME), Chem. Eng. Process. 52 (2012) 21-27. [10] H.Y. Lee, Y.C. Lee, I.L. Chien, H.P. Huang, Design and control of a heat-integrated reactive distillation system for the hydrolysis of methyl acetate. Ind. Eng. Chem. Res. 49 (2010) 7398-7411. [11] N.V.D. Long, Q.M. Le, C.N. Le, M. Lee, A novel self-heat recuperative dividing wall column to maximize energy efficiency and column throughput in retrofitting and debottlenecking of a side stream column[J]. Appl. Energy. 159 (2015) 28-38. [12] Y. Liu, J. Zhai, L. Li, L. Sun, C. Zhai, Heat Pump Assisted Reactive and Azeotropic distillations in Dividing Wall Columns, Chem. Eng. Process. 95 (2015) 289-301. [13] K Huang, Q Lin, H Shao, C Wang, S Wang, A fundamental principle and systematic procedures for process intensification in reactive distillation columns, Chem. Eng. Process. 49 (3) (2010) 294-311. [14] X. Gao, F. Wang, H. Li, X. Li, Heat-integrated reactive distillation process for TAME synthesis, Sep. Purif. Technol. 132 (2014) 468-478. [15] W.L. Luyben, Economic and dynamic impact of the use of excess reactant in reactive distillation systems, Ind. Eng. Chem. Res. 39 (8) (2000) 2935-2946. [16] W.L. Luyben, C.C. Yu . Reactive distillation design and control, John Wiley & Sons, New Jersey, 2008. [17] M.J. Lee, H.T. Wu, C.H. Kang, H.M. Lin, Kinetic behavior of amyl acetate synthesis catalyzed by acidic cation exchange resin, J. Chinese. Inst. Chem. Eng. 30 (2) (1999) 117-122. [18] S.F. Chiang, C.L. Kuo, C.C. Yu, D.S.H. Wong, Design alternatives for the amyl acetate process: coupled reactor/column and reactive distillation, Ind. Eng. Chem. Res. 41 (13) (2002) 3233-3246. [19] J.M. Douglas, Conceptual Design of Chemical Process, McGraw-Hill, New York, 1988.
[20] W.L. Luyben, Principles and case studies of simultaneous design, John Wiley & Sons, 2012. [21] General principles for calculation of total production energy consumption, GB/T 2589-2008. [22] W.L. Luyben, Distillation Design and Control Using Aspen Simulation, John Wiley & Sons, 2013.
89.8°C, 1bar Qc=-888.6kW
AmOH 50kmol/h 25°C, 1.5bar
F=50kmol/h 50°C, 1bar HAc=0.003 AmOH=0.005 AmAc=314ppm H2O=0.991
Organic Reflux
HAc 50kmol/h 25°C, 1.5bar
NR=2 NRX=22 NS=12
Qr=1009.4kW
F=50kmol/h 149.1°C, 1.245bar HAc=0.005 AmOH=0.004 AmAc=0.991 H2O=trace
Fig.1. Conventional reactive distillation process for AmAc synthesis
125.1°C, 1.156bar
Compressor duty=263.0kW
4 98.9°C, 1bar
PHP=1bar
7 PLP=0.14bar 1
Organic Reflux
AmOH 50kmol/h 25°C, 1.5bar
Qc=-393.823kW
HAc 50kmol/h 25°C, 1.5bar
2
6 F=50kmol/h 50°C, 1bar XD,HAc=0.003 Qr=425.935kW XD,AmOH=0.005 XD,H2O=0.991 5 XD,AmAc=325ppm 85.1°C 0.165bar 3
H2O
F=50kmol/h 85.3°C, 1bar XB,HAc=0.005 XB,AmOH=0.004 XB,H2O=trace XB,AmAc=0.991 AmAc
78.7°C, 0.14bar
Fig.2. Different pressure thermally coupled reactive distillation process for AmAc synthesis HP column LP column 0.14bar LP column 0.15bar LP column 0.10bar
140
Temperature (℃)
130 120
△T2
110 100 △T1
90 80 70
0
5
10
15
20
25
30
Number of stage
Fig.3. Temperature profile of DPT-RD process at different pressures
Compressor duty (kw)
Heat exchanger duty (kw)
450 Compressor Heat exchanger
340
420
320 390
300 280
360
260
330
240
0.08
0.10
0.12 PLP(bar)
0.14
300 0.16
Fig.4. Effect of the pressure of the LP column on compressor duty and heat exchanger duty (a)
(b) 629
760 NFB=2
TAC (10 $/year)
NFB=4
3
680
3
TAC (10 $/year)
628
NFB=3
720
640
6
7 8 9 NFA: feed location of HAc
18
20
22
24
26
NRX: number of reactive trays (d) 660 NLP=12
611
NLP=9
640
610
TAC (10 $/year)
609
3
3
625
10
(c) 612
TAC (10 $/year)
626
624
600
608 607 606 605
627
NLP=6
620 600 580 560 0.08
23 24 25 26 27 NHP: number of stages in HP column
0.10
0.12
0.14
0.16
PLP (bar)
Fig.5. Effect of (a) NFA and NFB (b) NRX (c) NHP (d) NLP and PLP on TAC 5
LP column
HP column 4
1.3
ΔT (℃)
ΔT (℃)
1.4
1.2
3 2 1
1.1 1
2
3
4
Number of stage
5
6
0
0
5
10
15
Number of stage
Fig.6. Temperature slope profiles in the LP column and HP column (a)
20
25
70
1.00
+20% F -20% F
XAmAc (mol%)
FAmAc (kmol/h)
60 50 40
+20% F -20% F
0.98
0.96
0.94 30
0
5
10
15
0
20
5
10
+20% F -20% F
50
40
30
0
5
10
15
20
0.990
0.988
0.986
0
5
10
+20% F -20% F
TLP, stage 1 (℃)
FHAc feed (kmol/h)
40
0
5
10
20
15
+20% F -20% F
79.6
50
30
15
Time (h)
Time (h)
60
20
+20% F -20% F
0.992
XH2O (mol%)
FH2O (kmol/h)
60
15
Time (h)
Time (h)
79.2 78.8 78.4 78.0
20
0
5
10
15
20
Time (h)
Time (h)
Fig.7. Dynamic responses of the basic control structure in (a) ±20% feed flow rate (b) 5% composition changes (b) 0.9930 0.9925 95mol% AmOH 95mol% HAc
49
48
XAmAc (mol%)
FAmAc (kmol/h)
50
95mol% AmOH 95mol% HAc
0.9920 0.9915 0.9910 0.9905
47
0
5
10 Time (h)
15
20
0
5
10 Time (h)
15
20
54 95mol% AmOH 95mol% HAc
XH2O (mol%)
FH2O (kmol/h)
53
95mol% AmOH 95mol% HAc
0.9930
52 51
0.9925 0.9920 0.9915
50 0
5
10
15
20
0.9910
0
5
54
79.6
50 48
0
5
10
20
15
95mol% AmOH 95mol% HAc
79.2
95mol% AmOH 95mol% HAc
TLP, stage 1 (℃)
FHAc feed (kmol/h)
52
15
Time (h)
Time (h)
46
10
20
Time (h)
78.8 78.4 78.0
0
5
10
15
20
Time (h)
Fig.7. Dynamic responses of the basic control structure in (a) ±20% feed flow rate (b) 5% composition changes (continued)
Fig.8. Improved control structure for the DPT-RD
(a) 65
0.996 +20% F -20% F
+20% F -20% F
55
0.992
XAmAc (mol%)
FAmAc (kmol/h)
60
50 45
0.988
0.984
40 0
5
10
15
20
0
5
Time (h)
Time (h)
70
0.996
+20% F -20% F
+20% F -20% F 0.992
XH2O (mol%)
60
FH2O (kmol/h)
20
15
10
50
0.988 0.984
40 0.980
0
5
10
15
20
0
5
Time (h)
10
15
20
Time (h)
Fig.9. Dynamic responses of the improved control structure in (a) ±20% feed flow rate (b) 5% composition changes 80.0 +20% F -20% F
65
+20% F -20% F 79.5
TLP, stage 1 (℃)
FHAc feed (kmol/h)
60 55 50 45
79.0 78.5
40 35
0
5
10
15
78.0
20
0
5
10
15
20
Time (h)
Time (h)
(b) 0.9920
95mol% AmOH 95mol% HAc
95mol% AmOH 95mol% HAc
XAmAc (mol%)
FAmAc (kmol/h)
50
49
48
0.9916
0.9912
0.9908 47
0
5
10 Time (h)
15
20
0
5
10 Time (h)
15
20
0.9930
95mol% AmOH 95mol% HAc
53
XH2O (mol%)
FH2O (kmol/h)
0.9925 52 51 50
95mol% AmOH 95mol% HAc
0.9920 0.9915 0.9910
0
5
10
15
0
20
5
10
15
20
Time (h)
Time (h)
79.6
54
95mol% AmOH 95mol% HAc 79.2
95mol% AmOH 95mol% HAc
TLP, stage 1 (℃)
FHAc feed (kmol/h)
52 50 48 46
0
5
10 Time (h)
15
20
78.8 78.4 78.0
0
5
10
15
20
Time (h)
Fig.9. Dynamic responses of the improved control structure in (a) ±20% feed flow rate (b) 5% composition changes (continued)
Tables and Figures
Table 1. Activity coefficient models parameters for AmAc system Comp.i
HAc
HAc
HAc
AmOH
AmOH
AmAc
Comp.j
AmOH
AmAc
H2O
AmAc
H2O
H2O
bij(K)
-316.8
-37.943
-110.57
-144.8
100.1
254.47
bji(K)
178.3
214.55
424.018
320.6521
1447.5
2221.5
cij
0.1695
0.2000
0.2987
0.3009
0.2980
0.2000
Table 2. Coal equivalent (CE) conversion coefficient and the CE energy consumption Energy consumption on CE (kg·h-1) DPT-RD
CRD
Vapor consumption
0.1286kg·kg-1
124.19
(reboiler) Standard coal
Cooling
water
equivalent
consumption
coefficient
(condenser)
0.4857kg·t-1
48.39
0.404kg·(kw·h)-1
106.25
109.19
Electricity consumption (compressor) Total (% difference)
154.64 (-33.74)
233.38 (0)
Table 3. Comparison of CRD and DPT-RD processes DPT-RD
parameter
CRD
HP column
LP column
RD column
Number of stages
24
6
37
Reactive stages
3-24
3-24
AmOH feed stage
3
3
HAc feed stage
6
7
Feed flow rate (kmol/h)
50
50
Operating pressure (bar)
1
0.14
1
Column diameter (m)
0.997
1.702
1.028
Operating cost Capital cost
(103$/year)
(103$/year)
Total annual cost
(103$/year)
186.95
301.12
351.18
283.54
538.13
584.66
Table 4. Control loops of multi-variable control Control circuits
Controlled variables
Control objectives
Manipulated variables
LC1
base leaves in HP column
50%
flow-out of HP column bottom
LC2
base leaves in LP column
50%
flow-out of LP column bottom
LC3
organic phase leaves in decanter
50%
flow-out of organic phase
LC4
aqueous phase leaves in decanter
50%
flow-out of aqueous phase
PC1
overhead pressure in HP column
1 bar
PC2
overhead pressure in LP column
0.14 bar
FC1
feed rate of AmOH
50 kmol·h
FC2
feed rate of HAc
50 kmol·h
TC1
temperature of stage 1 in LP column
78.87℃
flow of vapor from the top of the HP column brake power of compressor −1 −1
feed rate of AmOH feed rate of HAc ratio of the feed streams