Servo motor fault diagnosis method

Abstract

The invention discloses a servo motor fault diagnosis method which comprises the following steps of: collecting data samples of an AC servo motor in a normal operation state and in a fault state; training an ellipsoidal basis function (EBF) neural network by using a training sample; applying a harmony search (HS) algorithm to a fuzzy clustering FCM to form a HS-KCM algorithm so as to obtain the ellipsoid center value of a hidden layer neuron in each dimension of an input space; further obtaining the ellipsoid center value and the input and output values of each neuron; then training the weightvalue of the output unit corresponding to the hidden layer neuron by the BP algorithm; and finally inputting a test sample to determine the operating state of the AC servo motor. The method has the advantages of good universality, high accuracy and good self-learning ability, and solves the problem that an existing AC servo motor fault diagnosis method has a poor learning ability and is prone tomissed diagnosis and misdiagnosis.

Description

technical field [0001] The invention relates to the field of AC servo motor fault diagnosis, more specifically, the invention relates to a servo motor fault diagnosis method. Background technique [0002] With the rapid development of science and technology, the application range of AC servo motors is becoming wider and wider. Therefore, the fault diagnosis of AC servo motors is particularly important, and the complexity of the structure of servo motors is getting higher and higher, which brings greater challenges to the fault diagnosis of AC servo motors. Existing methods for fault diagnosis of AC servo motors can be divided into two categories, namely methods based on the mathematical model of the diagnostic object and methods based on knowledge base. The first type of method is only suitable for those who can obtain accurate mathematical models, and the detection object has relatively large limitations. The second type of method avoids the difficulty of extracting the ma...

AI Technical Summary

Problems solved by technology

However, the fuzzy reasoning method still has poor learning ability and is prone to missed diagnosis and misdiagnosis. However, the EBF neural network algorithm has been widely used in fault diagnosis because of its extensive adaptability and self-learning ability.

Images