12 April 2024 in Measurement Science and Technology
Traditional fault diagnosis methods may not capture key information during feature extraction due to the large distribution difference under different working conditions, which can result in poor accuracy of the diagnostic model. To address this issue, a subdomain adaptive bearing fault identification method guided by the local maximal mean difference (LMMD) under the multiscale adaptive residual network is proposed in this paper. The bearing vibration signals are preprocessed by wavelet convolution and wide convolution to generate initial features. Then, the multi-scale adaptive residual network is used to adjust the feature weights of different scales and extract richer feature information. To reduce the intra-class distribution difference, the LMMD is employed. Additionally, local interclass maximum mean difference (LIMMD) is used to increase the inter-class difference, preventing misclassification of samples from different classes due to their close proximity and achieving sub-domain distribution alignment. The fault diagnosis performance of the domain distance metric model guided by LMMD and LIMMD under the multiscale adaptive residual network is verified through two different bearing model validation experiments.
20 November 2024 in IEEE Transactions on Instrumentation and Measurement
The challenge of insufficient datasets is a significant issue in rotating machinery prediction and health management, particularly due to the scarcity of high-quality samples, which hampers the training of deep learning-based fault diagnosis models. generative adversarial networks (GANs) can effectively address the problem of insufficient samples. Therefore, in this article, a subpixel fully connected GAN (SPFCGAN) is proposed based on GANs to generate high-quality samples. Initially, the collected low-resolution (LR) signals are input into the generator of this network structure to produce high-resolution (HR) features through the fully connected layer and elemental rearrangement. The generated signals, along with the collected original HR signals, are then fed into the discriminator. By comparing the difference between the generated signals and the real signals, the discriminator performs signal discrimination, thereby improving the quality of the generated signal. The feasibility of the proposed SPFCGAN is validated by bearing data and gear data for fault diagnosis analysis.