Rolling bearing fault diagnosis transfer learning method based on domain invariant sequence transformation

Abstract

The invention discloses a rolling bearing fault diagnosis transfer learning method based on domain invariant sequence transformation, and the method comprises the steps: firstly converting original simulation data into a state observation sequence through employing an HMM model, eliminating the difference between the simulation data and actual fault data in amplitude, retaining key fault feature frequency information, and carrying out the fault diagnosis of a rolling bearing; the distribution difference between the simulation data and the actual data is eliminated, and domain-invariant fault information is obtained; and then, the state observation sequence is input into a GRU fault diagnosis model, and the time sequence correlation of data is obtained by using the long-time sequence modeling capability of the GRU, so that discriminative fault features are extracted, and fault diagnosis of the rolling bearing is realized.

Description

technical field [0001] The invention belongs to the technical field of fault diagnosis, and more particularly, relates to a transfer learning method for fault diagnosis of rolling bearings based on domain-invariant sequence transformation under only simulation data sets. Background technique [0002] With the development of modern industrial equipment, any minor failure of rolling bearings may cause the equipment to fail to operate normally or even be damaged. Therefore, timely and accurate fault diagnosis of rolling bearings is an important means to ensure the healthy operation of industrial equipment. [0003] In recent years, data-driven fault diagnosis methods for rolling bearings have made great progress without requiring precise physical models and prior knowledge, due to the extensive collection of rolling bearing condition monitoring data by widely used precision sensing devices. Among them, the fault diagnosis method of rolling bearing based on deep learning, as an...

AI Technical Summary

Problems solved by technology

[0005] However, the existing simulation data-driven deep learning fault diagnosis method does not consider the distribution difference between the simulation data and the actual fault data set due to the fact that the data comes from different fields, which leads to the direct application of the deep learning model obtained only through simulation data training to the actual The bearing fault diagnosis will cause serious performance degradation, and the accuracy of fault diagnosis is low

Images