Integrated Process Design And Control Based On Bi-level Optimization Strategies

In the real industrial process, researchers usually use sequential approaches to achieve the optimal design of chemical processes. Sequential approaches implement the process design and control system design in two separate stages respectively. First, based on the steady-state model, the best process design variables are obtained by using steady-state optimization methods; and then based on this designed process, the corresponding control system is designed to meet the requirements of dynamic control performance. However, sequential approaches ignore the natural relationship between process design and process control, the process obtained can not ensure the overall optimal both in economic performance and dynamic performance. Extensive research has shown that it’s necessary to consider process dynamic control performance in the initial stages of process design. In recent years, integrated design and control of chemical processes have been achieved increasing attention by researchers in process design and process control field, who have realized that integration of process design and control can obtain better economic performance and dynamic performance for chemical processes.In this paper, some approaches are presented for the integration of process design and control for CSTR processes. The main work of this paper is as follows:(1) State-of-the-art of the integration of process design and control is surveyed from three aspects:(I) the solution algorithm;(II) the control strategies;(III) the up-to-date review of the integration of process design and control for CSTR.(2) By using bi-level optimization strategies, a framework of integrated process design and control is constructed. In one level, process design parameters are obtained by using dynamic optimization approach; the resulted candidate process design is evaluated for its dynamic performance in another level. And finally, the optimal process design parameters and controller parameters are obtained. To validate the effectiveness of integrated design approach, the traditional PI controller and LQR are integrated in bi-level optimization framework, respectively. The simulation results show that the processes obtained from integrated design approach have better economic performance and dynamic control performance as compared with traditional sequential approach.(3) It is a fact that the hierarchical control structure has been used widely in industrial processes. Therefore, an approach based on hierarchical control structure is proposed for the integration of process design and control using bi-level optimization strategy. At the same time, the external disturbance and parameters uncertainly are also considered in our study. In the level of control decision optimization, the designed objects and conventional controller are tackled as a generalized object, and then the generalized object is controlled by advanced process controller(i.e., LQR). Similarly, in the level of design decision optimization, by considering the external disturbance and system uncertainly, the design decision optimization problem is formulated as a multi-period optimization problem. And the reasonable process design parameters and controller parameters are obtained in the end. Consequently, the economic performance and operability of the designed process can be ensured in the presence of external disturbance and parameters uncertainly.