Abstract:Rotating machinery operates continuously for long periods of time under varying conditions in actual industrial environments. The number of fault samples increases with equipment operating time, whereas differences in data distribution are inevitable because of varying operating conditions. As a result, fault diagnosis of rotating machinery poses challenges related to the fault domain increments. In addition, the issue of catastrophic forgetting, where previously learned knowledge significantly affects the model's performance, remains a critical concern. Deep adaptive sparse residual network (DASRN), a new lifelong learning-based method, is proposed to overcome these challenges. The DASRN model is constructed based on an improved deep residual network. The model uses a task-aware dynamic masking strategy to adjust the retention and utilization of blank weights in the network, thereby balancing the model's memory and learning capabilities, mitigating catastrophic forgetting. Furthermore, throughout the entire training process of the model, a dynamic data removal strategy is applied to eliminate easily distinguishable samples. This approach adaptively allocates higher weights to more challenging samples, enhancing the model's learning efficiency while consolidating the sparsity within the model. Finally, the effectiveness of the proposed DASRN is validated on bearing datasets from the Chinese CRH380A high-speed train gearbox and integrated wheelset transmission test rig. Results demonstrate that DASRN provides an effective solution for catastrophic forgetting and domain incremental scenarios in rotating machinery fault diagnosis tasks.

An Adaptive Symmetric Loss in Dynamic Wide-Kernel ResNet for Rotating Machinery Fault Diagnosis Under Noisy Labels

A Robust Wavelet-integrated Residual Network for Fault Diagnosis of Machines with Adversarial Training.

Fault Diagnosis of Rotating Machinery Based on One-Dimensional Deep Residual Shrinkage Network with a Wide Convolution Layer

Fault Diagnosis of Rotating Machinery Based on the Improved Multidimensional Normalization ResNet

A Robust Intelligent Fault Diagnosis Method for Rotating Machinery under Noisy Labels

Transfer Residual Convolutional Neural Network for Rotating Machine Fault Diagnosis under Different Working Conditions

Iterative Error Self-Correction for Robust Fault Diagnosis of Mechanical Equipment with Noisy Label

Multi-Scale Adaptive-Routing Capsule Contrastive Network-Based Intelligent Fault Diagnosis Method for Rotating Machinery under Noisy Environment and Labels

Anti-noise transfer adversarial convolutions with adaptive threshold for rotating machine fault diagnosis

Fault Diagnosis of Rotating Machinery under Noisy Environment Conditions Based on a 1-D Convolutional Autoencoder and 1-D Convolutional Neural Network

Deep Adaptive Sparse Residual Networks: A Lifelong Learning Framework for Rotating Machinery Fault Diagnosis with Domain Increments

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

A fault diagnosis method for wind turbines gearbox based on adaptive loss weighted meta-ResNet under noisy labels

Multi-scale deep residual shrinkage networks with a hybrid attention mechanism for rolling bearing fault diagnosis

An Anti-Noise One-Dimension Convolutional Neural Network Learning Model Applying on Bearing Fault Diagnosis

Average Descent Rate Singular Value Decomposition and Two-Dimensional Residual Neural Network for Fault Diagnosis of Rotating Machinery

A New Deep Learning Model for Fault Diagnosis with Good Anti-Noise and Domain Adaptation Ability on Raw Vibration Signals

Enhanced Deep Residual Network with Multilevel Correlation Information for Fault Diagnosis of Rotating Machinery

Lightweight Deep Residual CNN for Fault Diagnosis of Rotating Machinery Based on Depthwise Separable Convolutions.

ANC-Net: A Novel Multi-Scale Active Noise Cancellation Network for Rotating Machinery Fault Diagnosis Based on Discrete Wavelet Transform

Mechanical equipment fault diagnosis method based on improved deep residual shrinkage network