Design of Control Structure of Cumene Process

Jittima Thipsukhum1 and Montree Wongsri2

1 2

Master Degree Chulalongkorn University Engineering Chulalongkorn University.

Abstract
The structural design of control structures of a chemical plant is complex because a chemical process consists of a reaction section, a separation section, material and energy recovery in order to take full advantage in the productions. The general problem of plantwide control structure design is trying to achieve good controllability loop. A crucial step is to select matches of controlled variables (CVs) and manipulated variables (MVs) in an optimal way. Therefore, this research studies the design of control structure of the cumene process using Wongsri's design procedure (Wongsri, 2009). The steady state and dynamic simulation was done by HYSYS software. New designed plantwide control structures were designed and we used the design of Luyben (2010) as a reference. Material flow and thermal disturbance were used for this case to evaluate the control performance. Keywords: Plantwide control structure, Controlled variable, Fixture point theorem

1. Introduction
Nowadays, industries are very competitive both in quality and cost of production. Therefore, production process should have high quality and high efficiency. The process should always operate under the design condition, use little energy, low waste production and meet required specification of the products. In the real situation, the process will not operate smoothly. All factors do not meet the designed condition. The process always changes due to disturbance from the external factors and the internal factors. However, no matter what factors cause the change, in case of having deviation or disturbance come in to the process, the effect should be eliminated from the process as soon as possible process will have the least deviation from the designed condition. The appropriate choice of control structure among many possible control structures is important tasks in control system. In the past, Umeda, Kuriyama & Ichidawa (1987) presented the unit-based control system design methodology for control complete plants. The effect of recycles impacts on this method. Later Luyben, Tyreus and Luyben (1997) presented a general heuristic design procedure. Their procedure generated an effective plantwide control structure for an entire complex process flowsheet and not simply individual units. The nine step of the proposed procedure center around the fundamental principles of plantwide control: energy management, production rate, product quality, operational, environmental and safety constraints, liquid level and gas-pressure inventories, makeup of reactants, component balances and economic or process optimization.
BANGKOK UNIVERSITY RESEARCH CONFERENCE

195

Cumene can be found in crude oil and is a part of processed high-octane gasoline. Cumene is an important intermediate for industrial products such as phenolic resins, epoxy, nylon-6 and polycarbonate resins etc. Generally, processes for cumene manufacture consist of a packed bed reactor (liquid or vapor phase) followed by a separation train that removes the light inerts (propane), recycles the unreacted benzene and separates the cumene product from heavies generated by further alkylation of cumene to diisopropylbenzene. The heavies are sent to a transalkylation reactor for reactions with benzene to recover cumene. In this research, it will focus on new plantwide control design procedure of Wongsri (2009) applied to cumene process. In this study, will design plantwide control structures of cumene process and simulate them by using HYSYS simulator to study about dynamic behavior and evaluate the performance of the designed structures. An effective designed structure can achieve the control objective to reduce the cost of production and operate the process within safety and environment constraint.

2. Objective
To validate plantwide control structures of cumene process by using Wongsris procedure.

3. Literature Review
Luyben (1996) presented one of the central problems in developing a steady-state process flowsheet is finding the number of variables that must be specified to completely define the process. This number is called the design degrees of freedom. In principle, the design degrees of freedom are easily calculated by simply subtracting the number of equations from the number of variables. However, for typically complex industrial processes, there are many hundreds of variables and equations, and it is not a trivial job to make sure that the correct variables and equations have been defined. Luyben, Tyreus and Luyben (1997) presented a general heuristic design procedure. The nine step of the proposed procedure center around the fundamental principles of plantwide control: energy management, production rate, product quality, operational, environmental and safety constraints, liquid level and gas-pressure inventories, makeup of reactants, component balances and economic or process optimization. This procedure was illustrated with three industrial examples: the vinyl acetate monomer process, Eastman process and HDA process. The procedure produced a workable plantwide control strategy for a given process design. Kasemchainun (2003) present plantwide control strategies for designing control structures of vinyl acetate monomer plant to achieve the objective of this process can be operate within constraint of safety, environment, and operation. Three alternative plantwide control structures are designed, tested and compared the performance with Luybens structure. All of controls structures can operate within constraint achieve the objectives and a good control structure was quickly response to disturbance.

196

BANGKOK UNIVERSITY RESEARCH CONFERENCE

Suntisrikomol (2008) study of Fixture Point Theorem for hydride alkylation process (HDA) to select appropriate the set of controlled variables from a large number of candidate output. The fixture point control theorem state that the most disturbed points must be satisfactorily controlled by giving them consideration before other controlled variables. The maximum (scaled) gain is used to selecting and paring controlled variables with manipulated variables. In this study, the set of the first rank of controlled variables are same as Luyben (1998). There are select three set of controlled variables (second and third rank from fixture point) and five control structures were design and compared. In order to illustrate the dynamic behaviors of the control structures when economic disturbance load occur, the performance of controlled structures were presented in IAE value and compared with reference structure. The design structures are fast response and the most effective on compared with Araujo et al, 2006 and Luyben, 1998. Detjareansri (2009) proposed the control structures for alkylation process using new design procedure of Wongsri. The designed control structures are evaluated the dynamic performance compared with Luyben, 2002.She designed eight for new plantwide control structures. Material flow and thermal disturbance are used for this case to evaluate the control performance. All the designed control structures have a good performance because it handle disturbances entering the process and can maintain product quality as compared by integral absolute error (IAE). The Wongsris procedure is a useful design procedure for complex industrial process that leads to a good performance plantwide control system. Luyben (2010) presented of design and control of the cumene process. The purpose of this work is to use the cumene process to illustrate some interesting design optimization features and principles. The two dominant design optimization variables are reactor size and benzene recycle. Increasing either one reduces the amount of undesirable byproduct that is produced, but increasing either of these variables raises capital and energy costs. The result can infer that design optimization variables affect capital and energy costs. The economics of the cumene process illustrate that reducing reactant losses and the production of undesirable products can produce savings that are much larger than capital or energy costs.

4. Research Methodology
The methanol process flowsheet is studied in this work following by Luyben (2010). 4.1 Process Studied The cumene process features the desired reaction of benzene with propylene to form cumene and the undesirable reaction of cumene with propylene to form p-diisopropylbenzene. The main reaction for production of cumene involves the reaction of benzene with propylene in a high-temperature, high-pressure gas-phase reactor. (1) C6H6 + C3H6 C9H12 There is also a sequential reaction of cumene and propylene to form pdiisopropylbenzene (PDIB).
BANGKOK UNIVERSITY RESEARCH CONFERENCE

197

C9H12 + C3H6 C12H18 (2) The kinetic data is taken from a case study given by Turton et al. (2003). These exothermic reactions are irreversible and occur in the vapor phase. The kinetic expressions assumed to be valid for the system are: (3) (4) All reaction rates have units of kmol s-1 m-3. Concentration units are molarities. Activation energies have units of kj/kmol. The activation energy of the undesirable reaction is larger than that of the desirable reaction. Therefore low reactor temperatures improve selectivity. In addition selectivity is improved by keeping the concentration of cumene and propylene low in the reactor. This can be achieved by using a large excess of benzene, but the excess must be recovered and recycled. 4.2 Steady-state simulation The steady state model of cumene process is simulated in Aspen HYSYS. The Aspen NRTL physical properties model is used in all units of the process for realistic simulation. Figure 1 show the flowsheet of the cumene process equipment size and condition used by Luyben (2010).

Figure 1: flowsheet of the cumene process: Reference: William L. Luyben. 2010

198

BANGKOK UNIVERSITY RESEARCH CONFERENCE

The fresh feed streams of benzene and mixed C3 (propylene and propane) enter the process as liquids and fed into a vaporizer. The saturated gas leaves the vaporizer at 210 C and 25 bars. It is preheated in two heat exchangers. The first recovers heat from the hot reactor effluent at 358.5 C. The second adds additional heat to bring the reactor inlet temperature up to 358 C. Reaction Section, The reactor is a cooled tubular reactor that generates high-pressure steam from the exothermic reactions. We assume that the reactor inlet temperature is set to the same value as the coolant (steam) temperature in the reactor. Reactor effluent is cooled in the feed-effluent heat exchanger, and sent to a condenser. Then the two-phase stream from the heat exchanger is fed to a flash tank. The gas from the tank is used as fuel. The liquid is fed into the first distillation column C1 to separate Benzene. Cumene is separated in c2 column. 4.3 Dynamic Simulation For dynamic simulation in this work, the plantwide control structure is used to design control structure by using Wongsris procedure (2009). To evaluate the dynamic performance of control structure designed. In this work, two disturbances (feed flow rate and temperature outlet of reactor) have been studied. New design procedure of Wongsri (2009) presented plantwide control structure design procedure based on heuristics and mathematical analysis. In this procedure, the precedence of control variables is established The major disturbances are directed or managed explicitly to achieve the minimal interaction between loops by using the extended (thermal) disturbance propagation method to cover the material disturbances. The purposed plantwide control structure design procedure for selection the best set of control structure is intuitive, simple and straightforward. New design procedures of Wongsri (2009) are: Step1. Establishment of control objective. Step2. Selection of controlled variables to maintain product quality and to satisfy safety operational and environmental constrains and to setting the production rate. The selected CVs are ranked using the Fixture Point theorem. Step3. Selection of manipulated variables and measurements via DOF analysis. Step4. Energy management via heat exchanger networks. Step5. Selection of control configuration using various tools available. Step6. Completing control structure design by checking the component balance. Step7. Selection of controller type. Step8. Validation via rigorous dynamic simulation.

5. Result
5.1 New designed control structure Selection tray temperature Location, Figure 2 and Figure 3 shows the IAE summation result of tray temperature for Benzene column (C1) and Cumene column (C2) respectively from
BANGKOK UNIVERSITY RESEARCH CONFERENCE

199

Fixture Point method to select the appropriate controlled variables of tray temperature for both columns from a candidate output deviation. The temperature on tray 10st for Benzene column is appropriated controlled variables because it is the most sensitive. Identical, the temperature on tray 15st for Cumene column is appropriated controlled variables because it is the most sensitive. The new plantwide control structure for cumene process following Wongsris procedure as shown in Figure 4. 5.2 Dynamic Performance Results Figure 5 show dynamic responses of designed control structure for cumene process by 20% changes in the set point of the C3 fresh feed flow controller. The solid lines are 20% increase and the dash lines are 20% decreases and the disturbances are made at 0.2 h. Stable regulatory control is achieved the total benzene flow rates increase (Figure 5b) to maintain the quality of product. Figure 5h shows that the composition of benzene derived from the first distillation column is maintained a constant at 94 mol %.As the same composition of cumene derived from the second distillation column is maintained a constant at 99.9mol % (Figure 5l).

Figure 2: IAE summation result of Tray temperature deviation for Benzene column

Figure 3: IAE summation result of Tray temperature deviation for Cumene column.

200

BANGKOK UNIVERSITY RESEARCH CONFERENCE

Figure 4: Designed Control Structure for Cumene Process. The inlet temperature and outlet temperature of reactor are well controlled (Figure 5e and Figure 5f). As the same, both column temperatures manipulating reboiler heat input are well control (Figure 5o and Figure 5p). Stable regulatory control is achieved the total benzene flow rates increase (Figure 5b) to maintain the quality of product. Figure 5h shows that the composition of benzene derived from the first distillation column is maintained a constant at 94 mol %.As the same composition of cumene derived from the second distillation column is maintained a constant at 99.9mol % (Figure 5l). The inlet temperature and outlet temperature of reactor are well controlled (Figure 5e and Figure 5f). As the same, both column temperatures manipulating reboiler heat input are well control (Figure 5o and Figure 5p). The second disturbance is change in the set point of the reactor exit temperature controlled. Figure 6 presents for a change from 358.5 to 365 C (solid lines) and for change from 358.5 to 350 C (dash lines). As expected, both increasing and decreasing reactor exit temperature produces more undesirable product B2 (Figure 6m). Cumene production D2 is well controlled (Figure 6k). Composition of cumene derived from the first distillation column is maintained a constant at 0.096 mol % (Figure 6h). And composition of cumene derived from the second distillation column is maintained a constant at 99.89mol % (Figure 6l).

BANGKOK UNIVERSITY RESEARCH CONFERENCE

201

Figure 5: 20% change in C3 fresh feed flow rate.

202

BANGKOK UNIVERSITY RESEARCH CONFERENCE

Figure 6
BANGKOK UNIVERSITY RESEARCH CONFERENCE

203

6. Conclusion
In this research has designed control structure design for cumene process, using new design procedure of Wongsri (2009). This procedure based on heuristics and mathematical analysis. New design procedure of Wongsri (2009) can find the appropriate set of controlled variables. The best control configurations depend on the direction of controlled variable with manipulated variable. Therefore the Wongsris procedure useful design procedure that leads to a good-performance plantwide control system.

7. Recommendations
Study and design the plantwide control structure of the other process via new design Wongsris procedure.

8. References
Detjareansri, S. (2009). Plantwide control structures design for alkylation process. Masters Thesis. Department of Chemical Engineering, Engineering, Chulalongkorn University. Kasemchainun, S. (2003) Plantwide Control Design of Vinyl Acetate Monomer Plant. Masters Thesis. Department of Chemical Engineering, Chulalongkorn University. Luyben, M. L., Tyreus, B. D., & Luyben, W. L. (1997). Plantwide process control. New York: McGraw-Hill. Luyben, W. L., (1996). Design and Control Degrees of Freedom, Ind. Eng. Chem. Res., 35, 2204 2214. . (2010) Design and Control of the Cumene Process, Ind. Eng. Chem. Res., 49, 719734. Suntisrikomol, S. (2008). Plantwide control structure design procedure applied to the hydrodealkylation process using fixture point theorem. Masters Thesis. Department of Chemical Engineering, Engineering, Chulalongkorn University. Turton, R., Bailie, R. C., Whiting, W. B., Shaelwitz, J. A. (2003). Analysis, Synthesis and Design of Chemical Processes, (2nd ed). Prentice Hall: UpperSaddle River, NJ. Umeda, T., Kuriyama, T., Ichidawa, A. (1987). A logical structure for process control system synthesis. Proceedings of IFAC Congress, Helsinki.

204

BANGKOK UNIVERSITY RESEARCH CONFERENCE