Searching and Diagnosing Failures of Electric Machines as Part of Electric Drives by Using Neural Networks

The article presents a review of the demand ащк electric machines failure diagnosis. The most common modern methods of the diagnostics and search for the failures of electric drives are considered in this article; the disadvantages of these methods are presented as well. As an alternative for existing methods of technical diagnosis the authors proposed the structure that is built on the basis of a neural network. The general principles of the neural network operation and methods of its application in the industry are also described. The description of the structure that is oriented towards the diagnosing of the biggest quantity of the electric machines types is given. The article also presents the estimated response time of the considered algorithm on the basis of the theory of categories and its experimental research.