Research Of The Probability Distribution Of Control Parameter Tuning Based On ITAE And Fuzzy Classification

PID controller has simple algorithm, good robustness, and high reliability. It iswidely used in process control and motion control. PID control research has reached avery high level in domestic and abroad, but the theoretical research is often not up todate to the practice. The real industrial setting parameters are often still cannot be usedto achieve a satisfactory effect. In order to meet the demand of training and teachingenvironment, to make simple and quick access to the different performance of thesystem, and to reduce the tuning time of trial and error method, this research was carriedout. It is based on ITAE performance index and fuzzy classification on the performanceof the system. The purpose is to make the parameter combination to realize the differentlevels attributes as probability distribution.At first, this paper expounds the mathematical model of the analog PID controllerand the digital PID control. Then establishing a time domain model of unit feedbackclosed loop control system by Laplace Transform Method to realize the control systemdisperse. Through the Root Locus Method it analysis the transfer function of PIDcontroller, discriminate affects to the result of the system after joining the zero poledistribution, and preliminarily determined the range of the joining Kp, Ti, Tdparameters.After that, make fuzzy classification in the limited step exhaustive. In other words,make exhaustive calculation within the prescribed scope of three PID parameters. Thecalculation error of the integral type index (ITAE) are compared, the fuzzyclassification is roughly divided into four intervals. In the state of the four interval, thecalculate error integral level of indexes are corresponding to the PID parameters spaceand attribute as "good","preferably good","poor" and "bad" these four levels. Findingfour expected groups the parameter of combination, can get the probability distributionof different performance indicators in the four grades.Laboratory configured the low power digital servo controller. It can beprogrammed the12parameters into code, experiment data is able to validate that if thesystem can achieve a uniform probability distribution effect. With the MATLAB GUItools, this paper realized the computer control system visualization, making theobservation of getting the combination of parameters of the probability distribution in sight. These tuning methods of research PID parameters have good effect on teachingand training environment.