Study And Application On The Predictive PID Control Based On Neural Network

Due to the fact that time delay, inertia, strong nonlinearity, time varying and coupling existed in papermaking industry, routine PID couldn't satisfied with control requests. Neural network is extensively applied on multi-input-output system and it has strong managed ability. It could be approach arbitrary non-linear function with arbitrary approached precision. It has strong self-adaptive and self-learn ability. The paper introduces the method that neural network combined with PID control and learn the research and application on the predictive PID control based on neural networkFirstly, Bring forward a improved algorithm based on RBF neural network. Its primary idea uses total variance of minimization stylebook clustering, educe a kind of K-mean clustering algorithm, adjust the core and width of transfer function, use recursive least square to adjust weights. Simulation shows that it could enhance the modeling and track ability.Bring forward two kinds of neural network algorithm. One is non-linear predictive modeling self-adaptive PID control system based on the RBF neural network. The other is improves cell neural PID algorithm. Simulation shows that non-linear predictive modeling self-adaptive PID control system based on the RBF neural network has a better speed, a stronger robust and well self-learn. Improved cell neural PID algorithm for large time-delay system has strong self-adaptive high control precision and good control effect. Last ,according to the technological control requirements for drug adding process in papermaking industry and the difficulties how to measure and control the consistency of white water, a computer integrated control system based on King-View and PLC was developed and the control strategy of white water consistency was designed. The result of operation indicates that the control system offers quick adjustability, high precision, strong anti-interference capability and better stability and reliability. The technological requirements are fully met by this system and the productive efficiency and economic benefit are improved.