Fuzzy PID Temperature Control Algorithms Research Base On The Smith Predictor

The time delay exists in the actual industrial control systems widely, due toits character，there will be time lag when the signal or energy is transferring. Theregulation of the controller will reflect to system output through a period of time,so it will be very difficult for determining the parameters of the controller, if theparameter is seleced inappropriate, it will lead to the control performancedeterioration and even instability. If the system is a large time delay, nonlinear,time-varying, the problem will become more prominent. Temperature controlbelongs to the system which is more difficult to be controlled.This paper mainly discussed the problem of temperature control, Themethods of time delay system that will be introduced and analysed, but theintelligent control methods are not mature on the application of time-delaysystems, and it is developing at now stage, If the intelligent control methods canbe developed completely and applied in time delay system, it will be verysignificant.In this paper, The fuzzy PID controller is used in time delay systems, itcombines fuzzy control and PID control with their respective advantages, so itimproves the robustness and adaptability of the system. The fuzzy PID controlleras the important point is studied, and discussed its design problem. In additionthe Smith predictor is applied in the delay systems, and Smith predictor controlhas been improved, it is applied mainly when the charged object does not matchestimate the model, Based on the situation, an improved method is proposed, thenine points control method is used to adjusted the inertia time constant whichexists in the feedback channel. In order to certify the method which is efficient,the MATLAB as a tool is used for simulation. The method is certified effectively,and it can improve the performance of system assuredly.When the plant model does not match with the predictor model, the systemcontrol performance will become terrible, the problem is analysed in this paper.in allusion to this issue, an effective method that the Smith predictor is taken theplace of neural network is proposed, this method makes use of self-learningability of neural network fully, that solves the problem which the modelparameters don’t match with each other. This method is also studied by matlabsimulation, which achieves good resultsFinally, the algorithm of Chapter II is applied to the packaging machine,which controls the heating equipment. through the experiments verifies thealgorithm.