BP neural network-based self-tuning PID temperature control for coating ovens

Abstract

To meet the painting process requirements of newly developed vehicle models, SAIC-GM-Wuling Automobile Co., Ltd. Qingdao Branch (hereinafter referred to as SGMW) launched a new construction project for the oven of Production Line C. For conventional PID temperature control systems, improper parameter tuning tends to cause large temperature deviations, which fail to satisfy the requirements of the new process. In this project, a non-standard intelligent temperature control system for the coating gas oven is designed, which adopts BP neural network PID (hereinafter referred to as BPNN-PID) parameter self-tuning to meet the spraying and drying process requirements of the coating workshop. Firstly, a PID control system based on the BP algorithm is designed, and an approximate mathematical model of the oven system is established. With MATLAB software, Simulink simulation models for conventional PID control and BPNN-PID control are built and verified through simulation. Compared with conventional PID control, BPNN-PID control has the advantages of fast response speed, small overshoot, and small steady-state error. Finally, the tuned PID parameters are directly input into the host computer. The field operation results show that the tuned system has a small steady-state error and meets the high control accuracy requirements of the new process, which verifies that the scheme has good practical application value.