The Research On Intelligent PID Algorithm And Its Application In Temperature Control

The proportional,integral,and derivative(PID) control algorithm is one of the earliest developed control strategy,and one of the most widely used control strategy in industrial process.With the process becoming more and more complex,such as non-linear,parameter variable etc,it is very hard to get satisfied control performance for conventional PID.It is necessary to improve PID with advanced intelligent algorithms.Intelligent PID algorithm is a recently developed method by combining the intelligent control and conventional PID.It has the ability of self-learning, self-adaptive and self-organize.Intelligent PID controller has the characters of simple configuration,strong anti-disturbance ability and good robustness.Some common kinds of intelligent PID controller were researched,including Neural Network PID, Fuzzy adaptive PID and optimal PID based on genetic algorithm(GA),their control performance were analyzed through simulation and application in temperature control. The main content of the paper are as follows:The control plant is temperature process with characters of non-linear,parameter variable.Its approximate mathematical model was acquired through identifying its dynamic curve.PCC of B&R was chosen as the controller,the intelligent PID algorithms were realized through Automation Basic(AB) language based on Automation Studio(AS) software and applied in the temperature process.The theory of neural network PID was researched,single neural PID with gain self-tuning based on proportional,summation,derivative(PSD) algorithm and improved BP neural network PID were adopted to control the temperature process. The results of simulation and application showed that,single neural element adaptive PID can achieve high dynamic performance.It is simple,occupies little CPU and stack,and has a better anti-disturbance ability and robustness than conventional PID; BP neural network PID has strong self-learning ability,and farther improvement in anti-disturbing ability and robustness than the single neural element PID,but it occupies more CPU and stack.The principle of fuzzy control was analyzed.The theory of variable universe was used in fuzzy PID control to improve the self-adaptive.The variable universe fuzzy PID control can realize self-tuning of fuzzy factors and PID parameters.The results of simulation and application showed that,fuzzy adaptive PID can improve the performance of control system,with shorter settling time,better anti-disturbance ability and robustness than the conventional PID.Genetic Algorithm(GA) was adopted to optimize PID parameters.According to the two different kinds of target functions which is dynamic performance(such as overshoot,settling time) and ITAE,optimal PID parameters were obtained through GA.The result showed that,GA is a kind of optimization method to explore the overall optimum and independent of the initial conditions.It is an effective and practical optimization method of PID parameters.