Abstract:In response to the complex and variable working conditions faced by rolling bearings during actual operation, as well as the issue of vibration signal acquisition being easily disrupted by noise interference, the study describes the multi-source domain anti-noise rolling bearing failure detection approach (MEDThresNet). The purpose of this model's design is to solve the challenges of a lack of corresponding sample data and noisy signals in bearing fault classification. Using multi-condition source domains, as opposed to a single working condition source domain data, might help gain information from diverse domains and minimise overreliance on data from a specific working condition source domain. This can significantly increase the model's generalisation and robustness, and fault identification accuracy. Convolutional modules with soft thresholding and attention mechanisms are applied in this network structure. Soft thresholding helps to suppress noise in the data during the training phase while keeping critical characteristics. The attention mechanism, on the other hand, allows the model to automatically focus on the critical areas of the defect information in the bearing vibration signals throughout the training phase, hence improving the network's performance and generalisation ability. Furthermore, the network aligns the joint distribution of source and target domain data across many particular levels using the Joint Maximum Mean Discrepancy approach to accomplish unsupervised domain adaptation. This allows the network to successfully transfer information learnt from the source domain data of the faulty bearing to the target domain of the faulty bearing, improving the model's generalisability on the target domain. This research tests the network on two datasets with varied working conditions, CWRU and Ottawa, and the findings demonstrate that the network is high robustness and accurate for multi-source domain transfer diagnosis in noisy environments.

Integrating Adaptive Input Length Selection Strategy and Unsupervised Transfer Learning for Bearing Fault Diagnosis under Noisy Conditions

A parallel ensemble optimization and transfer learning based intelligent fault diagnosis framework for bearings

A novel transfer learning network with adaptive input length selection and lightweight structure for bearing fault diagnosis

A Fault Diagnosis Method for Rolling Bearing Based on Deep Adversarial Transfer Learning with Transferability Measurement

An unsupervised transfer network with adaptive input and dynamic channel pruning for train axle bearing fault diagnosis

A Novel Meta-Learning Network With Adaptive Input Length and Attention Mechanism for Bearing Fault Diagnosis Under Different Working Conditions

Fault diagnosis of rolling bearing with variable working conditions in noisy environment based on dynamic soft threshold and attention mechanism

A New Bearing Fault Diagnosis Method Based on Deep Transfer Network and Supervised Joint Matching

An Improved Transfer Learning Method for Bearing Diagnosis under Variable Working Conditions Based on Dilated Convolution

A deformable CNN-DLSTM based transfer learning method for fault diagnosis of rolling bearing under multiple working conditions

Rolling Bearing Fault Diagnosis Method Based on Attention CNN and BiLSTM Network

Transfer learning-based channel attention enhancement network combined with Gramian angular domain field for fault diagnosis

An Intelligent Fault Diagnosis Approach Based on Transfer Learning from Laboratory Bearings to Locomotive Bearings

A hybrid deep-learning model for fault diagnosis of rolling bearings in strong noise environments

Vibration Images-Driven Fault Diagnosis Based on CNN and Transfer Learning of Rolling Bearing under Strong Noise

Deep transfer learning rolling bearing fault diagnosis method based on convolutional neural network feature fusion

A Transfer Learning Model for Rolling Bearing Fault Diagnosis Based on Texture Loss Strategy and Nuclear Norm Regularization

A Deep Convolutional Neural Network with New Training Methods for Bearing Fault Diagnosis under Noisy Environment and Different Working Load

Deep Dynamic Adaptive Transfer Network for Rolling Bearing Fault Diagnosis With Considering Cross-Machine Instance

Supervised Contrastive Learning with Multi-scale Attention Mechanism for Fault Diagnosis of Bearing under Variable Operating Conditions

A Strong Anti-Noise and Easily Deployable Bearing Fault Diagnosis Model Based on Time-Frequency Dual-Channel Transformer