Intelligent mechanical fault diagnosis method based on deep transfer learning

Abstract

The invention relates to a mechanical fault intelligent diagnosis method based on deep transfer learning, and belongs to the technical field of mechanical equipment fault diagnosis. The method comprises the steps that S1, vibration signals of various faults of the mechanical equipment under different working conditions are collected and preprocessed; s2, Gaussian noise with different signal-to-noise ratios is added to the divided sample set, and various noise environments are simulated; s3, constructing a CAE-DTLN network by using an association alignment and domain adversarial transfer learning method; s4, inputting the training set into the CAE-DTLN, and carrying out iterative updating training on the CAE-DTLN by utilizing a source domain labeled sample classification error, a CORAL index and a domain classifier discrimination error; and S6, inputting the test set into the trained CAE-DTLN, and carrying out migration fault diagnosis on the mechanical equipment under different noises and working conditions. According to the method, anti-noise migration diagnosis can be realized, and the method has a good migration diagnosis effect and strong robustness and generalization ability.

Description

technical field [0001] The invention belongs to the technical field of mechanical equipment fault diagnosis, and relates to an intelligent diagnosis method for mechanical faults based on deep transfer learning. Background technique [0002] Existing industrial systems usually operate continuously and stably for a long time, and the frequency of failures is low. However, once a failure occurs, the failure will deteriorate rapidly, and the failure will cause a great threat. If it is not controlled in time, it will lead to serious accidents such as machine destruction and human death. Therefore, it is of great social significance to ensure the long-term stable operation of industrial systems and avoid serious accidents, and can bring huge economic and social benefits. [0003] In order to ensure the safety of industrial systems and promote the development of intelligent manufacturing, more and more industrial complex systems use the industrial Internet of Things platform to est...

AI Technical Summary

Problems solved by technology

However, opportunities are often accompanied by challenges. Due to the high complexity of the environment and incomplete information of industrial systems, the data-driven fault diagnosis technology encounters huge challenges in its actual application.

For data-driven fault diagnosis technology, the main problem comes from the data itself. In actual industrial applications, the cost of labeled data is very high; at the same time, the distribution of data under different working conditions is also different.

Images