Mingyuan Hu ,Hui Tian,2,*

1 College of Chemistry and Chemical Engineering,Yantai University,Yantai 264005,China

2 Collaborative Innovation Center of Comprehensive Utilization of Light Hydrocarbon Resource,Yantai University,Yantai 264005,China

Keywords:Cyclohexene hydration Catalytic distillation Control schemes Dynamic Simulation

ABSTRACT Cyclohexanol is a commonly used organic compound.Currently,the most promising industrial process for synthesizing cyclohexanol,by cyclohexene hydration,suffers from a low conversion rate and difficult separation.In this paper,a three-column process for catalytic distillation applicable in the hydration of cyclohexene to cyclohexanol was established to solve these.Simulation with Aspen Plus shows that the process has good advantages,the conversion of cyclohexene reached 99.3%,and the product purity was ≥99.2%.The stable operation of the distillation system requires a good control scheme.The design of the control scheme is very important.However,at present,the reactive distillation process for cyclohexene hydration is under investigation experimentally and by steady-state simulation.Therefore,three different plant-wide control schemes were established (CS1,CS2,CS3) and the position of temperature sensitive stage was selected by using sensitivity analysis method and singular value decomposition method.The Tyreus-Luyben empirical tuning method was used to tune the controller parameters.Finally,Aspen Dynamics simulation software was used to evaluate the performance of the three control schemes.By introducing ΔF ± 20% and xENE ± 5%,comparison the changes in product purity and yield of the three different control schemes.By comparison,we can see that the control scheme CS3 has the best performance.

1.Introduction

Cyclohexanol is a commonly used organic compound.Its production method has been widely investigated by industries and research groups worldwide.Currently,three main methods are used:phenol hydrogenation [1-6],cyclohexane oxidation [7-10],and cyclohexene hydration.Cyclohexene hydration offers obvious advantages in terms of safety,product selectivity,and cost,becoming the most promising method.Since the hydration reaction of cyclohexene is limited by the equilibrium of the chemical reaction and the mutual solubility between cyclohexene and water is extremely low[11-15],there are many azeotropes between the system components.At present,the cyclohexene hydration process still has problems such as low conversion rate,slow reaction rate,and difficulty in separation.After a long-term study,it was found that adding a co-solvent to the reaction system to improve the mutual solubility of cyclohexene and water,and adopting catalytic distillation can well solve these problems.

In the process of reactive distillation,the reaction process coexists with the gas-liquid mass transfer process.The two processes influence each other and promote each other,so that this unit operation can proceed smoothly [16-21].The reaction distillation technology to hydration of cyclohexene cyclohexanol has several advantages,(1) Increase reaction conversion rate.The hydration reaction of cyclohexene is a reversible reaction.Due to the existence of the distillation process,the reaction products are continuously separated from the system,causing a positive shift in the reaction equilibrium,improving the reaction conversion rate;(2) The reaction temperature is easy to control.The temperature of each stage in the reactive distillation is always the bubble point of the mixture at that stage under system pressure.Therefore,the reaction temperature can be controlled by adjusting the system pressure.In addition,the reaction heat is consumed by the distillation process,the problem of flying temperature of the rectification tower can also be effectively avoided.(3) Reduce system energy consumption.The heat of cyclohexene hydration reaction can be consumed by the distillation process.

The cyclohexene hydration process first developed by Asahi Kasei Corporation of Japan has the disadvantages of slow reaction rate and low equilibrium conversion rate,because the reaction is limited by a chemical equilibrium and the mutual solubility between cyclohexene and water is extremely low.US Patent 0018223A1[22]found that adding isophorone to the reaction system and using high silica zeolite ZSM-5 as a catalyst can improve the yield of cyclohexanol.Under the same reaction conditions,the final yield of the reaction without co-solvent was 12.0%,and the final yield of the reaction added isophorone as co-solvent was 24.3%.Isophorone as a co-solvent can improve the solubility of cyclohexene in water.It hardly participates in the reaction during the reaction,and has a large boiling point difference,which is easy to separate and recover in the post-treatment process.It can be used as a suitable cosolvent for cyclohexene hydration reaction.

In the industrial process,the stable operation of the reactive distillation column requires a reliable control scheme.Since the reactive distillation technology involves a unit that couples a chemical reaction process with the distillation process.The design of the control scheme for the reactive distillation process is very complex,and there is no universal control method.It is necessary to carry out detailed control system design according to the physicochemical properties and reaction mechanism of different systems,so that the reactive distillation system can be effectively controlled.At present,the reactive distillation process for cyclohexene hydration is under investigation experimentally and by steadystate simulation.No literature has reported dynamic control of the process toward the production of cyclohexanol.In the future,the study of control schemes will definitely be the focus of cyclohexene catalytic distillation hydration process.

Our team has undertaken many years of research on this method.Based on previous research,a three-column catalytic distillation process for cyclohexene hydration for cyclohexanol production is established here as a result.In view of the fact that there are many azeotropes in the system,the product is difficult to separate.The reactive distillation column used in this article only has a rectification section and a reaction section,and an additional azeotropic distillation column is used to complete the unfinished separation of the reactive distillation column.Then,the dynamic control scheme of this process is studied.Based on the analysis of system characteristics.We propose a feed-forward control structure to control the content of cyclohexene in the RD column,and this structure was similar to the R/F structure.After analyzing the control performance,it can be seen that this control scheme has good control performance.

2.Process Studied

2.1.Thermodynamic model

According to our previous research [23-25],the NRTL activity coefficient model can describe the non-ideality of the liquid phase in the vapor-liquid equilibrium and in the vapor-liquid-liquid phase equilibrium of the system.We compare the boiling point of the pure material and the boiling temperature of the binary azeotrope with the calculated values of NRTL,obtained at 101.3 kPa in Table 1.The absolute error for the temperature was 0.74 °C,and the maximum error was 0.003 mol,which indicated that the parameters calculated by Aspen Plus were with adequate accuracy.The binary interaction parameters at 101.325 kPa as shown in Table 2.The ternary diagrams and the binary diagrams has been given in the Supplementary Materials.

2.2.Kinetics model

The direct hydration of cyclohexene to cyclohexanol is a reversible reaction,and the reaction equation is as follows:

According to literature,the ion-exchange resin A-36wat catalyst and the kinetic modelrNOLfor the direct hydration of cyclohexene to cyclohexanol can be used to establish the following relationship:

where the reaction rate constant of the forward reaction [26]:;and the reaction rate constant of the backward reaction) Where the units ofkfandkbare L·(mol·min·g)-1,and the unit of activation energy is J·mol-1.

2.3.Description of the process

It is well known that the mutual solubility of cyclohexene and water is extremely small,which is disadvantageous for the progress of the hydration reaction.We increased the mutual solubility of cyclohexene and water by adding a cosolvent isophorone,thereby increasing the conversion of the reaction.In a previous study,our team established a process for the production of cyclohexanol by adding the catalytic distillation process of cyclohexene hydration reaction using the cosolvent isophorone [27].In this work,the three-column process was established based on the previous research.

Fig.1 shows the workflow of the process.The process is include a catalytic distillation column (RD),two purified distillation column (PD1,PD2),and two phase separators (Decanter1,Decanter2).The RD column includes two parts,a rectification section and a reaction section,and the feed position is the stage 9 at the top of the reaction section.The PD1 is an azeotropic distillation column.Its function is to separate the light component (cyclohexene and water)from the heavy component(cyclohexanol and isophorone).The PD2 is used to separate the cyclohexanol product and isophorone.We feed water into the Decanter1 to improve the liquidliquid phase separation effect.The liquid-liquid phase separation effect of Decanter1 was better than the Decanter2,so the origanic phase of Decanter2 was feed to Decant1 for further separation.

Table 1 Comparison of the calculated values and literature values [23-25] at 101.3 kPa

Table 2 Binary interaction parameters for the binary systems at 101.325 kPa

Fig.1. The workflow of the process.

The conversion rate of the cyclohexene hydration reaction is very low,and it is difficult to ensure the purity of the product and the conversion rate of the reaction using only one catalytic rectification column.There will be two situations in the simulation.First,when the product purity reaches the expected value,the reaction conversion rate is very low;second,if you want to ensure that the reaction has a higher conversion rate,the product purity is difficult to reach the expected value.Both of these are things are undesirable,so this article uses the RD column and PD1.RD column does not have a stripping section,PD1 is an azeotropic distillation column,used to complete the separation work that RD column did not complete.In this way,the concentration of cyclohexene in the RD column can be ensured by adjusting the reflux ratio of the organic phase in the RD column to ensure the conversion rate of the reaction.The PD1 is an azeotropic rectification column,and the azeotrope is cyclohexene.The ratio of cyclohexene and water in the PD1 can be controlled by adjusting the reflux ratio of the organic phase,and the water and cyclohexene can be well separated from the product and the co-solvent.The distillate product of column PD2 was high-purity(≥99.2%(mass))cyclohexanol.The bottom stream was nearly 100.0% (mass) isophorone cosolvent.We fed it to the RD column and recycled it.The specifications of the columns are shown in Table 3.

Before the dynamic simulation.It is important to remember that pumps and control valve install in the steady-state simulation,and the additional information that must be provided is the physical sizes of the various pieces of equipment.The procedure for sizing the distillation column shell(diameter and height)is by Aspen Plus[31].The sizes of the reflux drum and the column base,a commonly used heuristic is to set these holdups such that there are5 min of liquid holdup when the vessel is 50% full,based on the total liquid entering or leaving the vessel.For the reflux drum,this is the sum of the liquid distillate and the reflux.For the column base,it is the liquid entering the reboiler from the bottom tray.Assuming a length to diameter ratio of two,the diameters and lengths can be calculated:

Table 3 Specifications of the columns (RD column,PD1,PD2)

The equipment size as shown in Table 4.

3.Control Studied

For decades,catalytic distillation has been used in industrial applications,and many studies on the dynamic control of catalytic distillation have mostly studied the control of product purity whendisturbance occurred [20,28].In this work,we studied the yield and purity control of the product when disturbance occurred according to the characteristics of the system.Three control schemes were proposed.We intended to design an effective plant-wide control scheme for this system.It is expected that basic information can consequently be provided for the industrial production of cyclohexanol.

Table 4 Equipment size

3.1.Control schemes

Distillation control systems characteristically include two levels of focus[29].The first-level focus is concerned with stabilization of the basic operation of the column (we called this the inventory control scheme) and the second-level focus addresses the separation taking place in the column(we called this the separation control scheme).

The first level is the more mundane aspect of any distillation control system.It typically includes flow controls on the material and energy streams,inventory controls on the reboiler and accumulator,and pressure controls at the accumulator and/or base of the column.Therefore,we call this the inventory control scheme.

In the catalytic distillation process of cyclohexene hydration to cyclohexanol,there are 11 inventory control loops—8 liquid level control loops (organic phase,aqueous phase liquid level control in Decanter1;organic phase,aqueous phase liquid level control in Decanter2;bottom level control in the RD column;bottom level control in column PD1;bottom level control in column PD2;reflux-drum liquid level control) and 3 pressure control loops(RD column pressure,PD1 pressure,and PD2 pressure).The arrangement of the inventory control loops was as follows.The liquid level control was controlled by the valves of each outlet.The RD column pressure was controlled by manipulating the flow rate of the vapor distillate.The pressures of columns PD1 and PD2 were controlled by condenser heat removal.

In addition,there were four fresh feed flow rate control loops in the process to control the feed composition of the RD column.One temperature control loop was used for maintaining the temperature of Decanter 1 at 35℃by manipulating the feed temperature of Decanter 1.

We intended to design an effective plant-wide control scheme for this system.The next step was the design of the second level focus,which was the separation control scheme.Our goal was to achieve the expected product purity (≥99.2%) when the system faced a disturbance.This is not a simple task because there very complex phase equilibria and very complex distillation flows are involved.

As the first unit of the whole process,the stability of the RD column directly affects the downstream unit and affects the production of cyclohexanol.Catalytic distillation is a complicated process.The reaction and separation processes are carried out simultaneously and interact with each other.The control of the RD column is thus particularly important.Thus,we propose two different control schemes for the RD column,as shown in Fig.2(a) and (b).

The CS1 structure control with a temperature on a tray in the RD column was controlled by manipulating the reboiler heat input.The organic phase reflux ratio was held constant,as shown in Fig.2(a).The CS2 structure control with a temperature on a tray in the RD column was controlled by manipulating the reboiler heat input and organic phase reflux ratio,as shown in Fig.2(b).

The controls of columns PD1 and PD2 were the same in the control scheme 1 (CS1) and control scheme 2 (CS2).The temperature of a tray of column PD1 was controlled by manipulating the reboiler heat input,and the organic phase reflux ratio was held constant.The temperature of a tray of column PD2 was controlled by manipulating the reboiler heat input and the reflux rate was held constant.The selection of the manipulated variables and temperature control tray will be described in the next section.

3.2.Sensitivity analysis

For the control of the separation of the distillation column,first of all,it is necessary to ensure that the purity of the product reaches the expected specifications.In the industry,temperature control with minimal lag is often used to ensure product quality.If tray temperatures are to be used,the criterion is to select the best tray on which the temperature should be held constant.The position of the temperature controller is determined by the traditional open-loop sensitivity analysis [30,31].A very small change(0.1% of the design value) is made in one of the manipulated variables(e.g.,reflux flow rate).The resulting changes in the temperatures of all the trays are examined to see which tray experiences the largest change in the temperature.Dividing the change in the tray temperature by the change in the manipulated variable gives the open-loop steady-state gain between the temperature on that tray and each manipulated variable.The tray with the largest temperature change is the most ‘‘sensitive”and is selected to be controlled.A large gain indicates that the temperature on that tray can be effectively controlled by the corresponding manipulated variable.A small gain indicates that valve saturation can easily occur and the operability region could be limited.

I started to talk to some friends nearby when there was a tug9 on my sleeve, my arm was pulled over by a determined10 young Josh Frager, and he began putting a multicolored, woven yarn11 bracelet12 around my wrist

The open-loop steady-state gain of the three columns is shown in Fig.3.Fig.3(a)shows that the gain of the 17 trays was the largest when the reboiler heat input changed,which indicates that the temperature of the 17th tray can be effectively controlled by the reboiler heat input.Similarly,in Fig.3(b)and(c),it can be seen that column PD1 manipulates the reboiler heat input to control the temperature of the 9th tray,and column PD2 manipulates the reboiler heat input to control the 17th tray.

This completes the selection of the manipulated variable and temperature control tray for the CS1.For the selection of the manipulated variable and temperature control tray of the RD column in the CS2,we need to use another method for further analysis.The RD column in the CS2 has a double-ended control,which not only considers the loop sensitivity problem,but also balances the interactions between two controllers.We use singular value decomposition analysis for further analysis [32].

A 20×2 gain matrix K was formed using our calculated RD column reboiler heat input and open-loop steady state gain between the organic phase reflux ratio and the tray temperature.This matrix was decomposed using the ‘‘svd”function in Matlab into three matrices:K=UσVT(U is a 20×2 matrix,σ is a 2×2 matrix,V is a 2 × 2 matrix).The two U vectors were plotted against the tray number,as shown in Fig.3(d).The tray or trays with the largest magnitudes of U indicated locations in the column that could be controlled most effectively.In the figure,U1 is associated with the reboiler heat input.U2 is associated with the organic phase reflux ratio.The SVD results are similar to the sensitivity results.This suggests that stage 7 can be controlled by the organic reflux ratio and stage 17 by the reboiler heat input.

Fig.2. (a)Control scheme1(CS1);(b)Control scheme 2(CS2).Reflux ratio controller;Level controller;pressure controller;Temperature Controller;Flow controller;Multiplier.

The σ matrix is a 2 × 2 diagonal matrix whose elements are‘‘singular values.”The ratio of the larger to the smaller is the‘‘condition number,”which can be used to assess the feasibility of dual-temperature control.The singular values of the steady-state gain matrix were σ1=22.4786 and σ2=17.9476,which gives a condition number CN=σ1/σ2=1.25.This indicates that the two temperatures were fairly independent,so a dual-temperature control scheme should be feasible,at least from the steady-state point of view.

3.3.Controller tuning

Temperature control has significant inherent dynamic lags and dead times relative to the level and flow controllers.These should be incorporated in the control loop.This is necessary so that we can use realistic controller tuning constants and do not predict dynamic performance that is better than the performance achievable in a real plant installation.Tuning the controller during dynamic control is very important and time consuming.It is necessary to adjust the PID parameters in turn according to the degree of response of each control loop [33].

The tray temperature controllers were tuned by inserting a 1 min dead time in the loop and using the relay-feedback test to determine the ultimate gain (KU) and ultimate frequency (PU).Then,the Tyreus-Luyben empirical tuning method was applied to determine the gain (KC) and integration time (τ1).Table 5 gives the tuning constants.The tuning constants of the flow,pressure,and liquid level controllers are empirical values [31],i.e..The flow control gainKc=0.5,the integration time τ1=0.3 min.The pressure control gainKc=20,the integration time τ1=12 min.The liquid level controller only needs proportional control,so the gain of the controller wasKc=2,and the integration time was set to be very large (τ1=9999 min) to cancel the integral action.The valve is half open in steady state.

Fig.3. (a) RD column open loop sensitivity analysis;(b) PD1 open loop sensitivity analysis;(c) PD2 open loop sensitivity analysis;(d) RD column SVD analysis.

Table 5 Temperature controller tuning constants

4.Performance Evaluation

We wanted to determine how the control scheme proposed above responds when facing disturbances.We observed whether the purity of the product was close to the expected purity by introducing a disturbance of±20%(ΔF±20%)to the feed flow and±5%(xF,ENE±5%) to the composition of cyclohexene.The feed flow rate and the composition of cyclohexene were all for a mixed feed F.

4.1.Performance of CS1

The CS1 control scheme is shown in Fig.2(a):the organic phase reflux ratio was held at a steady-state value.The controller tuning parameters are shown in Table 5.After 5 h of system operation,the disturbances of ±20% to the feed flow rate (ΔF± 20%) and ±5% to the composition of cyclohexene (xF,ENE±5%) were added.The dynamic response of the purity of cyclohexanol product was observed.Fig.4(a) and (b) shows that the CS1 structure could quickly recover the purity of cyclohexanol to the expected value when facing the disturbance in the feed flow rate.However,as shown Fig.5(a),at the change ofxF,ENE-5%,although the temperature of the three columns could quickly return to the set point,the purity of the final product cyclohexanol was not satisfactory.The results show that this simple control scheme can better handle the ΔF± 20% [as shown Fig.4(a) and (b)] andxF,ENE+5% [as shown Fig.5(b)]disturbances.Product purity does not reach the expected value atxF,ENE-5% disturbance.

4.2.Performance of CS2

The CS2 control scheme is shown in Fig.2(b).The temperature of the 7th and 17th trays of the RD column was controlled by manipulating the organic phase reflux ratio and the reboiler heat input.The controller tuning parameters are shown in Table 5.After 5 h of system operation,the disturbances of ΔF±20%andxF,ENE±5%were added.The dynamic response of the purity of the cyclohexanol product was observed.This control scheme could better handle the disturbances ΔF+20%andxF,ENE+5%,as shown in Figs.4(b)and 5(b).However,the scheme did not handle the changes ofxF,ENE-5%and ΔF-20%well,as shown in Fig.4(a)and 5(a),since the product purity did not reach the expected value.

Fig.4. (a) Cyclohexanol product purity change when ΔF -20%;(b) cyclohexanol product purity change when ΔF+20%.

Fig.5. (a) Cyclohexanol product purity change when xF,ENE -5%;(b) cyclohexanol product purity change when xF,ENE+5%.

5.Alternative CS3

5.1.Established the alternative CS3

It can be seen from the above results that there are deficiencies in the CS1 and CS2.Therefore,we conducted a more in-depth analysis of the system based on the thermodynamic model and the chemical kinetics model.We hope to provide a better control scheme for this process.

It is well-known that the mutual solubility between cyclohexene and water is extremely small.Using the Aspen Plus process simulation software,we analyzed the composition of the tray of RD column and found that even if the cosolvent isophorone was added,it would still forms a vapor-liquid-liquid-solid reaction system in the RD column.The catalyst we used was a strongly acidic cation exchange resin A-36wat.On the external surface,where a hydrophilic sulfonic acid group was present,the catalysts were always surrounded by the aqueous phase during the reaction.Thus,it could be considered that the cyclohexene hydration reaction occurs in the aqueous phase.Thus,the content of cyclohexene in the aqueous phase can effect the conversion of the reaction,

By changing the reflux ratio of the organic phase,we observed the effect of different compositions of cyclohexene of the aqueous phase of the 8,12,16,and 20 trays of the RD column on the yield of cyclohexanol.As shown in Fig.6,the yield of cyclohexanol increases first and then decreases as the content of cyclohexene increases.From the chemical kinetic point of view,increase the content of cyclohexene in the aqueous phase is bound to increase the yield of cyclohexanol.However,why does cyclohexanol production decrease when the composition of cyclohexene is too high?This is because the catalytic distillation process is a complex process,and the reaction process is carried out simultaneously with the separation process.In the steady state simulation,we know that the top of the RD column is distillate the azeotrope of cyclohexene and water.When the composition of the cyclohexene in the RD column is too high,a large amount of azeotrope of cyclohexene and water is distillate from the top of the RD column.This results in a decrease in the content of cyclohexene and water in the reaction section.Therefore,the production of cyclohexanol will decrease.

Fig.6. Effect of different mass fractions of cyclohexene on cyclohexanol yield.

Fig.7. Alternative control scheme 3 (CS3).

Fig.8. (a)Cyclohexanol yield change when ΔF-20%;(b)cyclohexanol yield change when ΔF+20%;(c)cyclohexanol yield change when xF,ENE-5%;(d)cyclohexanol yield change when xF,ENE+5%.

Through the above analysis,we can know the content of cyclohexene in the system is the important factor.The RD column is the first unit of the process.Fluctuations in the composition of cyclohexene in the RD column not only affect downstream devices,but also affect the conversion of the system.Control of the cyclohexene content in the RD column is particularly important.In the steady-state simulation,we ensured that the content of cyclohexene in the RD column was optimal by adjusting the reflux ratio of the organic phase.Therefore,when the content of the cyclohexene feed changed,if there was no change in the reflux ratio of the organic phase,it must affect the composition of cyclohexene in the RD column,thereby affecting the purity and yield of the product.

Therefore,we propose a feed-forward control scheme to control the content of cyclohexene in the RD column.This structure was similar to the R/F structure,but with one difference:this structure was controlled by the ratio of the organic phase reflux ratio(RR)to the feed molar ratio of water to cyclohexeneIfFW/FENEincreased (decreased),the PID controller would result in the organic phase reflux ratio increasing (decreasing),and consequently the cyclohexene content in the RD column will increase(decrease).The control scheme 3 (CS3) was established as shown in Fig.7.The temperature of stage 17 was controlled by manipulating the reboiler heat input.The tuning parameters are shown in Table 5.This structure was essentially the same as CS1 whenFW/FENEwas fixed.

5.2.Performance of the CS3

After 5 h of system operation,the disturbances of ±20%(ΔF±20%)for the feed flow rate and±5%(xF,ENE±5%)for the composition of cyclohexene were added.The dynamic response of the yield of cyclohexanol was observed.As shown in Fig.8(a) and (b),for ΔF+20%,the CS1 control scheme showed a higher stable yield than the CS3.For ΔF-20%,the CS3 involved a faster settling time.In the case of a change in the feed flow rate,the seemingly identical CS1 and CS3 control schemes showed this difference because,when the feed flow rate changes,the change of the cyclohexene recycling stream from Decanter 1 causes fluctuations in the flow rate of the cyclohexene feed,which causes fluctuations in theFW/FENEin feed F,thereby causes the CS1 and CS3 to show different dynamic responses.As shown in Fig.8(a)and(b),the stable purity and stable yield of the CS1 and CS3 control schemes were similar when the feed flow rate changed.Thus,we can consider their performances to be similar.

One of the purposes of our CS3 was to ensure that cyclohexanol production is maintained at a high level when the content of cyclohexene changes in the feed.As can be seen in Fig.8(c)and(d),the CS3 maintained a higher cyclohexanol production at a disturbance ofxF,ENE±5% compared to that offered by the first two control schemes.

In order to compare the first two control schemes with this one,we also observed the dynamic response of cyclohexanol product purity on introducing the same perturbation.It can be seen in Figs.4 and 5 that this control scheme could still better control the purity of the product.

6.Conclusions

We applied the catalytic distillation technology to the cyclohexene hydration process to produce cyclohexanol.A steady-state model for the production of cyclohexanol by catalytic distillation and hydration of cyclohexene was established.Optimization was achieved using the Aspen Plus process simulation software.The conversion reached 99.3%,which is much higher than that given by the Asahi Kasei process,and the purity of the cyclohexanol product was ≥99.2%.

We established three different control schemes through different methods,hoping to better control the purity and yield of the products.The performances of the three control schemes were compared by investigating the dynamic response of system to changes in the feed flow and the feed content of cyclohexene.The detailed comparison of the three structures in terms of the steady-state deviation indicated that the product specifications were not met.We also compared the three structures in terms of stable purity,stable yield,settling time,and fluctuation size.As shown in Table 6.

For ΔF-20%,CS2 could not achieve the expected product purity(≥99.2%),while CS1 and CS3 showed the same stable purity and stable yield,but because CS3 showed a shorter settling time and less fluctuation,we considered that the performance of the CS3 structure was slightly better than CS1 for ΔF-20%.For ΔF+20%,the three structural products could achieve the desired product purity (≥99.2%),but the stable yield of CS1 was slightly larger than that of CS3 and much larger than that of CS2,so we considered that,for ΔF+20%,the performance of the CS1 structure was better.ForxF,ENE-5%,it was obvious that CS1 and CS2 could not give the expected product purity,and the stable output of CS3 was higher than that of CS1 and CS2.ForxF,ENE+5%,CS3 performed better than CS1 and CS2 in terms of product yield control.All three structures could achieve the desired product purity,but CS3 showed less fluctuation.When the content of cyclohexene was changed,CS3 showed better control performance than CS1 and CS2 in terms of product purity and yield.

On comparison,CS3 exhibited better control performance in terms of both yield and product purity.In this paper,the dynamic control scheme of catalyzed hydration of cyclohexene coupled with distillation toward production of cyclohexanol is discussed to provide essential data and a theoretical basis for the industrialization of the cyclohexanol production process.However,specific considerations should be kept in mind for the industrialization of the process,taking into account economic and performance parameters.For example,when the product purity is expected to be ≥99.0%,the CS1 structure shows better control performance and is also simpler and faster.

Table 6 Comparison of performance of three control schemes

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.

Acknowledgements

The authors acknowledge the Natural Science Foundation of Shandong Province,China (ZR2017QB006),the Focus on Research and Development Plan in Yantai city(2018XSCC038),and the Qingchuang Science and Technology Plan Innovation Team of Shandong Province (2019KJC012).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2020.11.029.