Font Size: a A A

Investigation Of Advanced PID Controlling For Giant Magnetostrictive Actuator

Posted on:2014-04-07Degree:MasterType:Thesis
Country:ChinaCandidate:A J DongFull Text:PDF
GTID:2268330431454362Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Giant magnetostrictive material characterizes high force, large displacement, highpower density, rapid response and wide range of temperatures, a good material for theproduction of mico-displacement actuator. Giant magnetostrictive material (GMA) madeby GMM presents a good prospect in the field of precision positioning. Response speed,precision positioning and anti-jamming capability of control system are importantperformance indicators for GMA. Traditional PID control method needs large numbers ofexperiments and parameters of PID are difficult to adjust. Also better effect of PID controlis difficult to achieve for GMA. In order to control GMA effectively, fuzzy adaptiveself-tuning PID control method and RBF neural network self-tuning PID control strategyare proposed. Analysis of comparison of aforementioned two advanced PID controlmethod is performed. The research of intelligent control for GMA has important theoryand practical meaning.Thesis analyzed the development and the superior performance of GMM and thecontrol methods of the GMA in detail. By the comparison with the commonly usedhystersis models of GMA, the Jiles-Atherton model is selected as the description of thehystersis. A physical model is established by the abstraction of the dynamic process. Incase of low drive magnetic field, linear model is established by simplifying hystersisnonlinear model. Linear model is mathematical model of GMA.Fuzzy adaptive PID controller is designed. The difference between expected input andoutput of GMA is error. The inputs of fuzzy control are error and error rate of change.Parameters of PID are tuned by fuzzy reasoning. RBF neural network self-tuning PIDcontroller is designed. Parameters of PID are tuned by adjusting parameters of RBF neuralnetwork. The appropriate parameters are selected by trial and error. A simulation of controlmethod is performed from aspects of precision positioning, response speed andanti-jamming capability. A comparison simulation of them is conducted. The designedfuzzy rules exhibit inferior performance than trial and error method for controlling actuator system. The experiments show that RBF network self-tuning PID control strategy obtainsbetter results.
Keywords/Search Tags:giant magnetostrictive actuator, fuzzy control, RBF neural network, PIDcontrol
PDF Full Text Request
Related items