A Method for Mechanical Product Failure Pattern Recognition Based on Evidence Neural Network Model

Abstract

The present invention is a mechanical product fault pattern recognition method based on evidence neural network model, which defines a new type of evidence neural network model based on evidence theory and radial basis neural network model, and based on the evidence neural network model to The signal characteristics of mechanical products are used for mode judgment and failure mode identification, taking into account the characteristics of limited sample size and high precision requirements. Based on evidence theory and neural network model, the basic trust distribution function of evidence theory is used to measure the neural network training process. Uncertainty, and using the evidence fusion rule to fuse information, the evidence neural network can effectively improve the accuracy of fault mode recognition.

Description

technical field [0001] The present invention is a mechanical product fault pattern recognition method based on the evidence neural network model, which utilizes evidence information fusion in evidence theory to quantify and optimize the parameters in the neural network training process, and to judge and analyze fault state information in mechanical products. Identification belongs to the technical field of reliability. Background technique [0002] With the rapid development of science and technology, the functional requirements of modern mechanical equipment are becoming more and more complex, and the requirements for work efficiency are getting higher and higher. And due to the complexity of the structure and function of the mechanical equipment, the harsh working environment, the increase in maintenance costs and other factors, the requirements for the performance indicators of the mechanical equipment are getting higher and higher. Once mechanical equipment fails, it wi...

AI Technical Summary

Problems solved by technology

However, the neural network model is based on a large number of samples and the limitation of accuracy, which affects the application of fault identification in mechanical products to a certain extent.

Images