|
|
Publications
Gao, D., Huang, K., Zhu, Y.S., Yan, K., Ren, ZJ. and Guedes Soares, C. (2024), Semi-supervised small sample fault diagnosis under a wide range of speed variation conditions based on uncertainty analysis, Reliability Engineering and System Safety, Vol. 242, 109746.
The generalization capability of a model is one of the most significant concerns in variable speed and few-shot-bearing fault diagnosis. The original vibration signal is not only subject to different responses due to different fault components, but also to changes in fault-related frequency or frequency band due to different rotational speeds. The wide variety of fault types and a large range of rotational speed variations make the fault-related features in the signal difficult to be mined, and therefore the generalization capability of the model is limited. To address this problem, a method for fault diagnosis through feature perturbation and decision fusion is proposed in this paper based on a semi-supervised training approach. First, a signal processor is used to extend the original vibration signal into several different perspectives to increase the diversity of features; then, graph convolution and different training methods are introduced to further perturb the features from the model perspective; finally, the behaviour of weak classifiers is analyzed from a statistical perspective, and the final fault types of the samples are determined by conditional probability solving. To verify the effectiveness of the proposed method, variable speed experiments containing 20 kinds of speeds were conducted based on the Bearing Prognostics Simulator test bench. The constructed dataset verifies the validity of the proposed method.
If you did not manage to obtain a copy of this paper: Request a copy of this article
For information about all CENTEC publications you can download: Download the Complete List of CENTEC Publications
|