Asynchronous Motor Fault Detection Based On Multi-case Tagging And Support Vector Data Description

With simple structure, asynchronous motors widely used in industrial production equipment as important driver.Once serious fault occurs to asynchronous motors not only damage the motor itself, but also will lead the entire production system to be great damaged, and even endanger the lives and safety of staff.The fault detection of Asynchronous motor is very important,This article will study the multi-loading fault detection of the machine.as follows:In this paper, an fault detection RBF-PCA-based method of asynchronous motor will be presented. when RBF kernel function parameters has been determined, stator current RBF model establishes.Because RBF does not fall into local optimum, so that the differences of the motor running will express on the weight vector of RBF, and finally PCA detection model will be establish on historical rights of all conditions.This paper presents an improved LDA-SVDD algorithm to establish fault detection model of multiple operation modes of asynchronous motor. stator current RBF model establishes,RBF weights sample of all conditions need to be calibrated by their respective categories, That is tagged, so that FDA can be characterized and classified.After the FDA classification, samples were mixed degree higher area, the FDA to classification in class in this area, to each kind of different regions of the FDA's score vector SVDD model is set up.In order to detect the accuracy of the methods described above,It needs stator current data,so multi-loop induction motor mathematical model need to be established to achieve fault condition simulation and multi-motor.