Pneumatic Actuator Fault Diagnosis Based On Relevance Vector Machine

As an important executive terminals of industrial production, actuator is directly relatedto the safety and reliability of the production process. Once the actuator is faulted, it willbring huge losses to the production process. Pneumatic actuator is superior to the hydraulicactuators and electric actuators, with its simple structure, convenient maintenance and nopollution, and is widely used in industrial automation. Self-validating (SEVA) pneumaticactuator not only can realize the function of self-test and self-diagnosis, but also can providethe parameters of their own state. In order to solve the fault diagnosis problem ofself-validating pneumatic actuator, an pneumatic actuator fault diagnosis approach based onrelevance vector machine (RVM) regression modeling and relevance vector machine (RVM)multi-classifier is proposed. The method establishes an actuator model of pneumatic actuatorsusing the relevance vector machine (RVM) multi-regression, and the residuals are generatedby comparing the output of the models and the actual self-validating actuator is used as thenonlinear features. Then, the structure of the RVM for multi-classifier is designed usingk-meaning clustering methods, which is used as fault classifier to identify the condition andfault pattern of the self-validating actuator. The proposed approach is verified using fault datagenerated by DABLib（Matlab-Simlink）model and actuator data from Lublin Sugar Factory.Compare the three methods which are one-and-one (OAO), one-and-another (OAA) andbinary tree classification based on relevance vector. And then, the proposed approach iscompared with support vector machine (SVM) fault diagnosis approach. The results indicatethat the proposed approach overcomes the drawbacks of SVM that choosing kernel function isrestricted by the Mercer theory and the support vector will increase with the sample increase,leading to poor sparse and resolves the small sample and nonlinear problem in SEVApneumatic actuator fault diagnosis.