Jingwei Yang,Zhengkun Hou,Yao Dai,Kang Ma,Peizhe Cui,Yinglong Wang,*,Zhaoyou Zhu,Jun Gao

1 College of Chemical Engineering,Qingdao University of Science and Technology,Qingdao 266042,China

2 College of Chemical and Environmental Engineering,Shandong University of Science and Technology,Qingdao 266590,China

Keywords:Azeotrope Pressure-swing distillation Dynamic controllability Control structure

ABSTRACT Dynamic controls of pressure-swing distillation with an intermediate connection(PSDIC)process of ethyl acetate and ethanol separation were investigated.The double temperature/composition cascade control structure can perfectly implement effective control when±20%feed disturbances were introduced.This control structure did not require the control of the flowrate of the side stream.The dynamic controllability of PSDIC with partial heat integration(PHIPSDIC)was also explored.The improved control structure can effectively control±20%feed disturbances.However,in industrial production,simple controller,sensitive and easy to operate,is the optimal target.To avoid the use of component controllers or complex control structure,the original product purities could be maintained using the basic control structure for the PSDIC process if the product purities in steady state were properly increased,albeit by incurring a slight rise in the total annual cost(TAC).This alternative method without a composition controller combined with the energy-saving PSDIC process provides a simple and effective control scheme in industrial production.

1.Introduction

It is well-known that distillation is widely applied as an efficient method for separating mixtures in the chemical process industry[1–3].In addition to conventional distillation,several special distillation technologies were researched for separating azeotropes,such as pressure-swing distillation(PSD)[4–10],extractive and azeotropic distillation[11–19].Because PSD does not introduce the third component,it is an efficient method for separating azeotropic mixtures [20,21].While the PSD technique has many well-known benefits,its significant energy consumption is still a major drawback[22–25].

Steady-state design and dynamic control in PSD processes for separating the azeotropic mixtures have been studied for a long time in many investigations[26–28].Currently,the development of an energy saving method of PSD to improve the energy efficiency is an important research topic[29–34].Several recent papers have studied the energy saving technologies of PSD in more complex distillation systems.Xia et al.[8]proposed a new modification to achieve process intensification of PSD by using self-heat recuperation technology.Vapor recompressions were applied on both columns to achieve“self-heating”with no external heat source required.Aurangzeb and Jana[31]studied a new heat-integration method for separating the mixture of methanol/acetonitrile/benzene by vapor recompressed pressure-swing dividing wall column.This method can achieve the best performance compared to the other schemes.Kiran and Jana[34]studied a hybrid heat-integrated PSD process for separating bioethanol.This scheme showed promising performance in aspect of energy usage,capital cost and TAC savings.Chen et al.[35]designed a fully heat-integrated PSD process with the side withdrawal for ternary mixture separation.These researchers established a new temperature-difference control structure in corresponding fully heat-integrated PSD dynamic process,which achieved good dynamic performance.Huang et al.[36]researched the energy-saving consequent of rectifying/stripping section type heat integration for separating acetonitrile and water in the PSD process.This heat integration method not only improves the thermodynamic efficiency of the process design but also reduces the capital investment.

On the other hand,studying the effective control strategies of the PSD processes is an essential part of the design to ensure a robust control performance when faced with the inevitable disturbances [37–41].Yu et al.[32]presented a novel method for separating methanol/methylal and used PSD with fully heat integration to reduce energy consumption.The control structure of pressure-compensated temperature worked reasonably well and showed great dynamic responses for this PSD with fully heat integration.Luyben has made a great contribution to the study of control dynamics in distillation.Luyben[37]proposed the control structures of heat-integrated PSD process for separating the mixture of trimethoxysilane/methanol.This control design can effectively deal with large disturbances.Wang et al.[39]researched the control of PSD processes with heat integration(no,partial,and full)to separate methanol/tetrahydrofuran.

Ethanol and ethyl acetate which are generally used in chemical and pharmaceutical industry as the excellent industrial solvents are very significant organic chemical raw materials[41].During the production of ethylenediamine,the binary azeotrope of ethyl acetate and ethanol is produced[42].In our previous work,the steady-state design of PSDIC and PHIPSDIC processes was investigated based on process intensification to separate ethyl acetate and ethanol [42].Zhang et al.[43]developed valid control structures to attenuate snowball influence of PHIPSDIC process.The results of this study were of great significance for ensuring the purities of products in facing the feed disturbance in industry.However,the dynamic controllability of the PSDIC for separating azeotropic mixtures has not been explored.The dynamic controllability of PSDIC process is unknown for the introduction of a side stream in the PSD process.Thus,the purpose of our study is to explore control strategies of the PSDIC process,so as to explore the dynamic process of energy-saving PSD process.

Distillation operation process is complex and changeable in actual production.Different control structures need to be developed to cope with different operation conditions.Although Zhang et al.[43]developed some effective control structures of PHIPSDIC process,it is also meaningful to propose different effective control structures.In this study,the“over-purity”design for basic control structure of PSDIC was studied to achieve effective control.For the system that can further improve the purity of the product in the steady state,the new alternative scheme that we proposed may be feasible.Although the energy cost and TAC under steady state are slightly increased,the use of complex controllers can be avoided.And the TAC values for different product purities in the side-stream pressure-swing distillation process were also compared,and the balance among the product purity,cost and controllability was analyzed.This alternative scheme to achieve the required purity without using composition controllers or complex control structure is feasible and can be further studied and applied in other PSD processes.Buckley[44]discussed the overpurity design in the 1960s.Based on this approach,the TAC values for different product purities in the PSDIC were also compared,and the balance among the product purity,cost and controllability was analyzed.Zhu et al.[45]researched the separation of pressure-sensitive azeotropes by extractive distillation.And the results showed that the process control is improved to a great extent when TAC was slightly increased.Wang et al.[46]separated isobutanol and n-heptane by extractive distillation.The process is evaluated from the perspective of economy and dynamic control performance.The results showed that the dynamic control performance was improved when the flow rate of solvent and TAC were increased.There is a trade-off among purity,TAC and control performance.Based on the above research,dynamic control was also used to explore the process of the PSDIC for different product purities.In term of control,dynamic control performances with different product purities were evaluated to ensure a proper PSDIC process.

2.Process Design and Temperature Control Stage Select

The optimized PSDIC process in steady state for ethyl acetate and ethanol separation was developed in our previous work[42].The difference between the PSDIC and PSD was that an intermediate stream was present in the PSDIC.Compared with PSD,this process required less energy.The detailed process was shown in Fig.1(a).The UNIQUAC was selected to describe physical property of steady-state simulation.The initial feed flow was set to 100 kmol·h?1.And the feed composition was chosen as ethyl acetate 50 mol%and ethanol 50 mol%.The highpressure column operated at 607.95 kPa had 48 stages and was fed on the 26th stage,and the location of recycle feed was stage 38.The lowpressure column operated at 50.663 kPa had 27 stages and was fed on stage 18.The flow rate of side-stream was 56 kmol·h?1,the outlet was located at stage 23 and the feed was fed on stage 10.The ethanol and ethyl acetate product purities were both required to be 99.6 mol%at the bottom of the two columns.The pressure drop for each tray of the PSDIC process was set to 0.689 kPa.The holdups of the two columns were set as the liquid entering or leaving the vessel to allow 5 min of liquid holdup when the vessel was half full.Aspen Plus Dynamics was used to explore dynamic performance of the PSDIC.

The temperature-sensitive control stages in the PSDIC process should also be chosen carefully.The slope criterion method is the most common method,and we use it to select and analyze the temperature control stage in the PSDIC process[44].The temperature and corresponding slope curves of these two columns are shown in Fig.S1.There were obvious fluctuations in stage 43 of C1 and stage 22 of C2.Therefore,select stages 43 and 22 as the control stages in two columns,respectively.

3.Control of the PSDIC Process

3.1.Basic temperature control structure for PSDIC

In this work,dynamic simulation process is realized by using Aspen plus Dynamics V8.4 software.For the design of control strategies and selection of the pairing of the closed loops,we use the methods developed by Ma et al.[47].According to the results of degrees of freedom analysis,to implement effective control of two columns,a minimum of five basic controllers must be added in a basic control system including the pressure controller,base level controller,reflux drum level controller,tray temperature controller and feed flow controller[48].The control structure of basic temperature with fixed reflux ratios is shown in Fig.2.The corresponding control structures and settings are as follows:

(1) Fresh feed is controlled by feed inlet valve(reverse acting).

(2) The pressures of columns C1 and C2 are manipulated by controlling condenser heat load(reverse acting).

(3) The base levels are maintained by controlling the bottoms flow rates of two columns(direct acting).

(4) The reflux drum levels are maintained by controlling distillate flow rate in the two columns(direct acting).

(5) The temperature-sensitive stages in two columns are manipulate by the corresponding reboiler duty(reverse acting).

(6) The reflux ratios in two columns are fixed.

For the basic controller tuning parameters in PSDIC,the pressure controller and fresh feed flow controller are set to KC=20,τI=12 min,KC=0.5,τI=0.3 min,respectively.The level controller is set to KC=2,τI=9999 min.Two deadtime elements with the 1 min deadtime are used for two temperature control loops.Temperature controllers are tuned by using Tyreus-Luyben tuning rules and relay-feedback tests.Table S1 presents the tuning parameters of two temperature controllers.

The dynamic performances for±20%disturbances are presented in Fig.3(a)and(b).The disturbances are introduced as 0.5 h and the dynamic process is paused as 20 h.Although each variable in Fig.3(a)and(b)can be quickly restored to a stable state,the purities cannot return to initial values which introduced the±20%feed disturbances.The initial values of the two product purities are set to 99.60 mol%.However,when introducing the+20%feed flowrate disturbances,purity of ethanol can only reach 99.38 mol%.When ?20%feed composition disturbances are introduced,the purity of ethanol can only reach 99.50 mol%rather than 99.60 mol%.

To observe the intermediate stream under feed disturbances,the side-stream flow rate is analyzed.As shown in Fig.3,transient deviation of the intermediate stream has varying degrees of changes under feed disturbances.The control of the intermediate stream is of interest.Therefore,the control structure with intermediate stream flow rate controller is proposed,and this structure,the responses to the feed disturbances are shown in Fig.S2.However,the control effect of intermediate stream flow rate controller is not obvious.Therefore,it is necessary to explore new control structures of PSDIC process.

Fig.1.Flowsheet for the(a)PSDIC process(b)PHIPSDIC process(1atm=101325 Pa).

3.2.Reflux-to-feed ratios control structure for PSDIC

According to the basic control structure,another commonly used control structure using the reflux-to-feed ratios instead of the reflux ratios is explored for the PSDIC.The reflux-to-feed ratios control structure for PSDIC and the disturbance results are shown in Fig.S3.However,as can been seen,the disturbance results of reflux-to-feed ratios control structure show the same performance to the basic control structure,that is when introducing the+20%feed flowrate disturbances,purity of ethanol can only reach 99.38 mol%,when introducing the ?20%feed composition disturbances are introduced,the purity of ethanol can only reach 99.50 mol% rather than 99.60 mol%.And it has not achieved the effect control for the PSDIC.

3.3.Reboiler duty/Feed flowrate control structure for PSDIC

Under the disturbance of feed flow,the composition of column C2 has large transient disturbance.Therefore,based on the basic control structure,the control effect of reboiler duty/feed flowrate control structure on column C2 was explored.According to the basic control structure,reboiler duty of feed flowrate control structure is shown in Fig.S4(a).It is expected when the disturbance of feed flow rate is introduced,the reboiler duty must be changed,which may significantly improve the dynamic characteristics for the PSDIC process.The responses to the disturbances of feed flow rate and composition are shown in Fig.S4(b)and(c).The purities of products are not held close to their required specifications,particularly when introducing±20%composition disturbances,which shows that this control structure is also not ideal.When introducing the+20%feed composition disturbances,purity of ethyl acetate can only reach 99.55 mol%,when introducing the ?20%feed composition disturbances are introduced,the purity of ethanol can only reach 99.50 mol%rather than 99.60 mol%.

Fig.2.Basic control structure with fixed ratio of the PSDIC.

3.4.Pressure-compensated temperature control structure for PSDIC

Fig.3.Dynamic responses of the basic control structure for the PSDIC:(a)±20%feed flow rate disturbance;(b)±20%feed composition disturbance.

Fig.4.Double temperature/composition cascade control structure of the PSDIC.

Fig.5.Dynamic responses of double temperature/composition cascade control structure for the PSDIC:(a)±20%feed flow rate disturbance;(b)±20%feed composition disturbance.

Fig.6.(a)Improved control structure for the PHIPSDIC,(b)flowsheet equations for PHIPSDIC.

Based on the problem of ±20% composition disturbances for reboiler duty/feed flowrate control structure,the pressure-compensated temperature control structure is studied.Temperature control structures are usually used to pursue product composition in distillation column,however,the changes of column pressure will have a negative impact on the estimation of components.Therefore,the control structure of pressure-compensated temperature is proposed to test whether the control structure can perform the effective control of PSDIC process.In this control structure,measurements of both pressure and temperature on the temperature-sensitive control stage are used to estimate the corresponding composition.In order to better control the high-pressure column C1,and in consideration of the possible application of pressure-compensated temperature control structure in part heat integration process,so as to obtain a larger temperature difference to ensure the heat input of the reboiler of the low-pressure column C2,the pressure-compensated temperature control structure is first applied to the high-pressure column C1.Fig.S5 shows the control structure and the flowsheet equations.When running the relay-feedback tests on pressure-compensated temperature controller CC43,KCand τIobtained by Tyreus-Luyben tuning are 0.41 and 10.56 min.The KCand τIon the temperature controller TC2 are 3.31 and 7.92 min,respectively.The dynamic performance characteristics of this control structure to±20%composition and feed flow rate disturbances are shown in Fig.S6.It is observed that these disturbances are not handled effectively.When introducing the+20%feed flow rate disturbances,purity of ethanol can only reach 99.41 mol%,when introducing the+20%feed composition disturbances are introduced,the purity of ethyl acetate can only reach 99.54 mol%and when introducing the ?20%feed composition disturbances are introduced,the purity of ethanol can only reach 99.50 mol%rather than 99.60 mol%.

3.5.Temperature/composition cascade control structure for PSDIC

Although this temperature controller has an advantage of quickly response,it cannot maintain desired purity.By contrast,the composition controller is moderative,but it can ensure that purities of products achieve the desiring values.The temperature/composition cascade control structure not only controls the speed but also maintains the purity of the product.Therefore,this cascade control structure used in PSDIC process is aimed to control the purities of the two columns.Due to the large transient disturbance of the composition of column C2 under the feed flow disturbance,the temperature/composition cascade control structure is firstly selected for column C2.The cascade control structure for PSDIC is shown as Fig.S7.And the corresponding tuning parameters of the temperature/composition controllers are shown in Table S2.The addition of the composition controllers can further the overall control.The disturbance results are shown in Fig.S8.The purities of products maintain values close to the required specifications under the±20%feed flow rate disturbances.However,a value which is closed to its required specification cannot be maintained for ethyl acetate under the+20%feed composition disturbance.It is possible that the column C1 is not equipped with a composition controller.

3.6.Double temperature/composition cascade control structure for PSDIC

Based on the problem of +20%feed composition disturbance for temperature/composition cascade control structure,the double temperature/composition cascade control structure is proposed.Because the intermediate stream flow rate controller has no obvious effect for entire control loop,intermediate stream flow rate controller is ignored.The deadtime of the composition control loop is 3 min.Fig.4 shows the corresponding control structure,and Table S3 presents corresponding tuning parameters of the composition and temperature controllers.

Fig.5 shows the dynamic performances introduced±20%feed disturbances.It is observed that temperature/composition cascade control structure can remain constant and maintain the desired purity values of the dynamic process after 3 h.Although the PSDIC process is more complex than conventional PSD,they can be controlled effectively by using the temperature/composition cascade control structure.And the comparison of six control structures for PSDIC is shown in Table S4.

4.Control of PHIPSDIC

Our previous paper revealed that the TAC of the PHIPSDIC was lower than that of the conventional partial heat integration PSD (Fig.1).Therefore,it is also essential to explore an effective control structure for PHIPSDIC.In the pressure compensation temperature control structure,the input signal of the temperature controller of the high-pressure column is replaced by the component signal.The composition signal is based on the temperature and pressure value on the sensitive plate to calculate the ethyl acetate content on the plate.The pressure-compensated temperature controller CC43 obtaining the component signal further transmits the output signal to the Reboiler duty/Feed flowrate controller QR1/F.Fig.6 shows the calculation formula of the component signal,and the calculation result is the input value of the delay timemodule.Because the gas phase of the high-pressure column provides heat for the low-pressure column,the column top pressure controller is set to the manual state.The pressure-compensated temperature control structure is also explored.However,these results indicate that the system could not reach the desired values and the control effect is not ideal.Then,the improved control structure of the PHIPSDIC was proposed.The control structure and the flowsheet equations are shown in Fig.6.The corresponding tuning parameters of this control structure are shown in Table S5.The dynamic performances are shown in Fig.7 after introducing±20%feed disturbances.The improved structure can retain constant levels after 4 h and rapidly brings the ethanol purity back to its initial value.The only slight deficiency of this approach is that ethyl acetate purity reaches 99.55 mol%instead of 99.6 mol%after introducing+20%feed composition disturbance.

Table 1 Detailed costs of the PSDIC process with different purities

Fig.7.Dynamic responses of the improved control structure for the PSDIC with partial heat integration:(a)±20%feed flow rate disturbance;(b)±20%feed composition disturbance.

5.Discussion and Analysis

Studies of the dynamic controllability of PSD process are always a valuable and important topic.Although the effective control structure of the PSDIC was developed,it is not a perfect control structure because of the addition of two composition controllers.In practical production,the plurality of distillation columns is designed to achieve a designated separation between two components.Composition controllers have high maintenance costs.The composition controllers are also often expensive because of the composition analyzers they contain.The stability of composition controllers is sometimes not suitable for on-line continuous control.Therefore,finding a good solution scheme without using composition controllers to achieve highly effective control is significant for practical production.

It is observed that the basic control structure could achieve fast and stable control in the PSDIC process,but this structure cannot meet the purity requirements.It is important to study the use of a basic controller to achieve effective control.For the system that can further improve the purity of the product in the steady state,the new alternative scheme that we proposed may be feasible.First,the product purity in steady-state design is improved and the new Aspen Plus file is exported into the Aspen Dynamics simulation file.Then,the dynamic response of PSDIC process using the basic control structure under the new purity of product can be tested.As a general rule,with the increase in the product purity,the TAC also increases correspondingly.Therefore,the analysis of the balance among product purity,economy and controllability is essential.This method is feasible only when the cost is increased slightly.

In this section,the economics of the process with the increased product purities of 99.7 mol%,99.8 mol%,and 99.9 mol% were investigated.The total annual costs for the purities at 99.7 mol%,99.8 mol%,99.9 mol%are obtained and detailed economic calculation results are shown in Table 1.It is observed that the TAC increases slightly when the purities are 99.7 mol%and 99.8 mol%.When the purity is improved to 99.9 mol%,the TAC is increased by 2.25%.

It is possible that the new process will not reach new purity levels after introducing the feed disturbances.It is also possible that the purity can be returned to its original purity level at the cost of incurring a small TAC increase.Therefore,the control effect using basic control structure under the new purity levels is investigated.

5.1.Basic temperature control structure for the PSDIC with purities of 99.7 mol%

First,the product purity is increased from the initial 99.6 mol%to 99.7 mol%in the Aspen Plus file,and the Aspen Plus file is exported to the dynamics file as a pressure-driven simulation.The parameter settings are the same as before.For the new purity PSDIC process,temperature sensitive trays must be selected again.Stage 43 in column C1 and stage 21 in column C2 are selected as the temperature control stages.The basic control structure is same to the control structure in Fig.2.The temperature controller parameters of new steady state with the purities of 99.7 mol%are shown in Table S6.It can be seen from Fig.8 that the purity cannot return to the desired value of 99.6 mol%under ?20%composition disturbance.Therefore,the dynamic controllability for the PSDIC with the purities of 99.8 mol%requires further study.

5.2.Basic temperature control structure for the PSDIC with the purities of 99.8 mol%

Fig.8.Dynamic responses of the basic control structure for the PSDIC with purities of 99.7 mol%:(a)±20%feed flow rate disturbance;(b)±20%feed composition disturbance.

The dynamic controllability of the PSDIC with the purities of 99.8 mol%is proposed.The temperature sensitive trays are the same as in the PSDIC with the purities of 99.7 mol%.The corresponding temperature controller parameters are shown in Table S7.Fig.9 shows the dynamic performances of the PSDIC with purities of 99.8 mol%using the basic control structure.The results show that the product purity of both columns can reach the original specified value of 99.6 mol%under the±20%feed disturbances.

The study of the relationships among the three variables of controllability,purity and TAC are highly significant for the development of simple,efficient and robust dynamic control strategies.This alternative scheme to achieve the required purity without using composition controllers can be further studied and applied in other PSD processes.

6.Conclusions

In this work,six control strategies of dynamic controllability of the PSDIC process for separating azeotrope of 50 mol%ethyl acetate and 50 mol%ethanol are explored.The basic temperature control structure and other common control structures cannot achieve effective control in the PSDIC process.The double temperature/composition cascade control structure can achieve stable and effective control of PSDIC process when±20%feed disturbances are introduced.For the PHIPSDIC,the improved control structure is developed,and this control structure can achieve fast and stable control in the PHIPSDIC process.It is noteworthy that the PSDIC process with and without heat integration can be effectively controlled by adding a composition controller.On the other hand,improving the steady-state purity to achieve effective dynamic control is an alternative control scheme that does not involve the use of complex controller at the price of incurring a small TAC increase.After adding disturbances,the purities can reach the original values using the basic temperature control structure.The results show that we can sacrifice TAC properly to improve the purity of products,and then achieve good control effect with using basic control structure,which can provide guidance for industrial production.

Nomenclature

PSD pressure-swing distillation

PSDIC pressure-swing distillation with intermediate connection

PHIPSDIC PSDIC with partial heat integration

TAC total annual cost

KCgain

τIintegral time

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig.9.Dynamic responses of the basic control structure for the PSDIC with purities of 99.8 mol%:(a)±20%feed flow rate disturbance;(b)±20%feed composition disturbance.

Acknowledgements

This work is supported by the National Natural Science Foundation of China(No.21776145 and 21676152).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2020.07.059.