Iterative Learning Control Of Chemical Batch Processes

Batch process production is a system that performs the same control task repeatedly in a finite time interval. Its mode of production is flexible, which can meet the demands of production that of small quantity, that of multiple kinds and that of high quality, and it plays an increasingly important part in industries. However, due to its time-varying parameters, nonlinear and difficulty for modeling, even many system parameters are difficult to measure, also with a variety of operating constraints and interference factors, the control of batch processes is a challenging subject for control community. So new advanced control methods should be researched.For batch process’s repetitive feature, iterative learning control is very suitable for batch process system. And the application in batch process control is an important respect of iterative learning control. Iterative learning control can overcome the unfavorable factors well in batch process control, and obtain better effects than that of traditional methods. The application of iterative learning control method in batch process has become a hot research for scholars in recent years.According to the temperature control of batch chemical processes, pH value control, this paper proposes a series of intelligent learning control methods. The main work and key innovations are summarized as follows:1) This paper presents a deep analysis of the characteristics and difficulties of batch process, also, the research status and tendency of batch process is introduced. Combined with the characteristics of iterative learning control, this paper proposes the application of iterative learning control (ILC) in batch process control, and compared with traditional control methods, iterative learning control can achieve higher precision tracking and more rapid convergence property for the characteristics of batch process’s repetitive process; and iterative learning control system does not rely on the precise mathematical model, it is more significant for solving difficulty in modeling and time varying parameters.2) Aiming at the temperature control of batch chemical processes, utilizing the repetitive nature of batch processes, PD-ILC is designed and its convergence property is analyzed. Also, this kind of circumstance about the initial state of each iterative is different to its idea value,so interference factors are considered. At last, the validity of proposed PD-ILC has been confirmed with simulation results. 3) This paper further according to the batch process’s difficulty for modeling, time-varying parameters and nonlinear, wastewater pH value control is researched as an example and AILC is proposed, which incorporating iterative learning control and adaptive control, applying iterative learning control based on the projection algorithm, which improves the proposed algorithm’s adaptability. The algorithm performs well in a simulated pH value control and shows good track capability, rapid convergence and high control precision.