The Application Research On Neural PID Control In Industrial Process Control

PID control method is one of the traditional control methods and gets good effects under many application situaltions. But with the increase in complexity of manufacture technics and demands of industrial process performance, the conventional PID control can not meet the requirement of closed loop optimized control. Intelligent control independent of model of a plant and based on knowledge offers a new idea for improving the process control quality, of which neural network , as one of modern information process technologies, has some advantages in many applications. Neural network control became a regarded research direction. And neural PID control is one of the main techniques in research and application due to the dominant place of PID types control in industrial process control recently. If the controller which has better performance and is wieldy like PID can be found, there will be significance in both theory and practice.In this article, several kinds of neural PID controllers are introduced. According to the need of industrial process, three typical neural PID controllers (including the Single neuronal PID Controller, the Neural Network PID Controller and the Liked PID-NN Controller) are studied by simulation in order to show their characteristic and applicability. Aiming at their shortcomings some projects are proposed to optimize the controllers in this article.In the end, the proposed LPNC is applied to control the temperature of the crystallizing process in the automatic control system of the sorbitol product line. The neural PID control is proved to be effective and adaptive by simulation and realized using ANSI B language in RSView32.