Imbalanced fault diagnosis method based on attention perception and decision coupling

By using a deep reinforcement learning framework that generates three-channel images and asymmetric reward functions, the problems of majority class bias and single feature representation in imbalanced fault diagnosis are solved, and efficient identification and autonomous diagnosis of weak faults are achieved.

CN122113004AActive Publication Date: 2026-05-29INST OF ENERGY HEFEI COMPREHENSIVE NAT SCI CENT (ANHUI ENERGY LAB) +1
3 Cites 0 Cited by

Patent Information

Application Number
CN202610561294.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-27
Publication Date
2026-05-29
Estimated Expiration
2046-04-27

AI Technical Summary

Technical Problem

When dealing with imbalanced fault diagnosis, existing technologies often suffer from majority class bias due to traditional methods, resulting in low minority class fault identification rates, limited feature representation, and a lack of adaptive reward mechanisms, making it difficult to effectively identify minor faults in dynamic environments.

Method used

A method based on attention perception and decision coupling is adopted. Three-channel images are generated through Gram angle field and Markov transfer field transformation. Combined with a deep reinforcement learning framework of channel attention mechanism and asymmetric reward function, minority class faults are autonomously diagnosed.

Benefits of technology

Without relying on manual resampling or prior threshold settings, it significantly improves the ability to identify a very small number of fault modes, providing a highly sensitive and robust intelligent diagnostic solution suitable for complex industrial scenarios.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to the technical field of industrial intelligent monitoring and fault diagnosis, in particular to an unbalanced fault diagnosis method based on attention perception and decision coupling. According to the technical scheme, the method comprises the following steps: obtaining a one-dimensional vibration time sequence observation signal of target industrial equipment in an operation state, and preprocessing the one-dimensional vibration time sequence observation signal; and respectively inputting the preprocessed time sequence signals into the Grubby angle field sine transformation, the Grubby angle field differential transformation and the Markov transfer field for parallel coding. On the premise of not depending on manual resampling or priori threshold setting, weak fault fingerprints submerged in strong background noise can be autonomously captured, the upper limit of recognition of a very few types of key fault modes is remarkably improved while the diagnosis reliability of most types of samples is ensured, and the recognition efficiency is improved. An intelligent diagnosis solution which is high in sensitivity and robustness and does not need manual intervention is provided for complex industrial scenes such as wind power, aviation, rail transit and the like which have high reliability requirements and are difficult in fault data acquisition.
Need to check novelty before this filing date? Find Prior Art