Multi-target-domain equipment fault diagnosis method based on AdaDCLF

Abstract

The invention discloses an AdaDCLF-based multi-target domain equipment fault diagnosis method, which combines an MK-MMD distance to measure the distribution deviation of a source domain and a target domain, then adaptively adjusts a domain confusion loss factor, and focuses on the importance of target domain data corresponding to the maximum domain confusion loss. According to the AdaDCLF adaptive adjustment domain confusion loss factor method, the generalization ability of the model on the multi-target domain is enhanced, the diagnosis accuracy of the equipment fault diagnosis model on the multi-target domain is remarkably improved, and the problem that the generalization ability of a single-target domain adaptive algorithm on the multi-target domain is not strong is solved, so that intelligent diagnosis of equipment faults is realized.

Description

technical field [0001] The invention relates to a device fault diagnosis method, in particular to an AdaDCLF-based multi-target domain device fault diagnosis method. Background technique [0002] With the rapid development of industrial big data and the Internet of Things in the context of Industry 4.0, intelligent health management of key industrial equipment in factories has become more and more important, which is also an urgent problem that smart factories need to solve. In actual industrial applications, most equipment fault location and maintenance rely on past experience or expert knowledge. This diagnosis method requires a lot of manpower and time costs, and the efficiency is not high. In response to this problem, intelligent equipment fault diagnosis combines artificial intelligence technology and fault diagnosis technology to become an important branch of intelligent health management. Traditional traditional machine learning methods, including support vector mach...

AI Technical Summary

Problems solved by technology

However, these domain adaptation methods are often based on a single target domain. In practical applications, there may be more than one target domain. If only one of the target domains is adapted, the performance of the training model in other target domains may be reduced.

Images