Study On Kernel-Function-Based Nonlinear Predictive PID Control Method And The Key Techniques Of Its PLC Implementation

Microbial fermentation process is the foundation of the modern bioengineering, biotechnology and their industrialization. With the advancements of fermentation technology and the scale extending of its industrial production, the development of fermentation process control technology plays an important role to improve the product yield and quality. The PID controller is convenient, while it has a certain robustness of model error, has become the most widely used industrial controller. But there are problems while applied in the systems with nonlinearity and time-delay characteristics, so it’s difficult for PID to control the complicated process. Fermentation involves the growth and reproduction of living organisms whose mechanism is complicated, is with highly nonlinear and time-varying and large time delay characteristics. The traditional PID controller cannot satisfy the actual needs of the fermentation process control. Nonlinear predictive PID method based on kernel function is able to effectively handle large hysteresis control problem, which has stronger robustness and reliability when the parameters of the controlled system change, can satisfy the special requirement of the fermentation process. So researching the nonlinear predictive PID control method based on kernel function and solving the key technical problems encountered in its realization based on PLC is of important theoretical significance and application value.In consideration of the nonlinear, time-varying and large time-delay features of fermentation process control, the improved nonlinear predictive PID recursive controller is constructed by the research of predictive PID method using SVM error prediction method based on kernel function in this paper; the predictive PID method based on the recursive least squares SVM is researched, the truncation error’s influence to the prediction PID controller robustness is analysed; the realization technology of the proposed kernel predictive PID method based on PLC is studied, the nonlinear predictive PID algorithm is efficiently realized on the low-end PLC with Limited resources, which supplies technology basement for the fermentation measurement control system based on PLC; the purposed method is improved and optimized through the simulation experiment and practical application, the effectiveness of the proposed method in this paper is verified.Experiment studies show that nonlinear predictive PID method based on kernel function is able to deal with the fermentation process control problems effectively, the control system possesses strong robustness under the disturbances of Gauss noise and observation noise and truncation errors while the purposed kernel predictive PID algorithm is realized based on low-end PLC with only integer arithmetic function. The purposed algorithm ensures that the controlled system can run reliability and possess good control effect while parameters change, and bring effective solutions for the performance optimization problems of the complicated fermentation process control system.