Abstract:Fault diagnosis is significant to guarantee the safety and reliability of the machinery. Local faults will make the collected vibration signals deviate from the normal signals. To extract fault-related features from vibration signals, many deep-learning-based methods have been proposed for machinery fault diagnosis recently. However, due to the extreme conditions (e.g., high background noise and limited labeled training samples), it is a challenging task to implement feature extraction from the collected signals with nonlinear and nonstationary characteristics. To implement feature extraction and noise reduction of vibration signals, this article proposes a novel network, that is, deep morphological shrinkage convolutional autoencoder (DMSCAE) for gearbox fault diagnosis considering the insufficient labeled training samples. First, a morphological convolutional autoencoder is proposed for noise filtering and feature extraction. Second, a multibranch structure with different structural elements (SEs) is used in the morphological layer to extract impulsive components. Finally, a soft thresholding-based shrinkage is employed to filter ineffective features, where an adaptative method is developed to adjust the threshold automatically in the backpropagation procedure. The experiments on two gearbox fault diagnosis tests are conducted to verify the performance of DMSCAE. The results indicate that DMSCAE obtains a better performance for fault diagnosis than other DNNs, for example, efficient channel attention network (ECANet) and self-calibrated convolutional network (SCNet).

What problem does this paper attempt to address?

The paper aims to address several key issues in gearbox fault diagnosis: 1. **Nonlinear and Non-stationary Signal Feature Extraction**: Extracting fault-related features from collected vibration signals becomes highly challenging due to the operation of mechanical equipment under extreme conditions (such as high background noise and limited labeled training samples). These vibration signals typically exhibit nonlinear and non-stationary characteristics. 2. **Feature Learning with Limited Labeled Samples**: In practical industrial applications, faulty components are replaced promptly once defects are detected, making it difficult to collect sufficient labeled samples. This results in poor generalization performance of existing deep learning models. 3. **Multi-scale Information Extraction**: Due to the dynamic nature of vibration signals, a single network branch cannot effectively extract multi-scale information. To address the above challenges, the paper proposes a novel deep learning network—Deep Morphological Shrinkage Convolutional Autoencoder (DMSCAE), with the main contributions including: 1. **Unsupervised Learning Network**: A morphological convolutional autoencoder is proposed for simultaneous denoising and feature learning of vibration signals without the need for labeled samples. This enables effective feature learning even with insufficient labeled samples. 2. **Multi-branch Morphological Layer**: A multi-branch morphological layer is constructed in DMSCAE, using different structural elements (SEs) to achieve multi-scale feature learning and denoising of non-stationary signals. 3. **Soft-thresholding-based Shrinkage Layer**: A soft-thresholding-based shrinkage method is introduced to highlight impulsive information. The soft threshold can be automatically updated according to the characteristics of the vibration signals. Experimental validation shows that DMSCAE performs excellently in gearbox fault diagnosis tasks, outperforming other deep neural networks (such as ECANet and SCNet).

Deep Morphological Shrinkage Convolutional Autoencoder-Based Feature Learning of Vibration Signals for Gearbox Fault Diagnosis

Fault Feature of Gearbox Vibration Signals Based on Morphological Filter Dynamic Convolution Autoencoder

An Enhanced Convolutional Neural Network with Enlarged Receptive Fields for Fault Diagnosis of Planetary Gearboxes

Adaptive Densely Connected Convolutional Auto-Encoder-Based Feature Learning of Gearbox Vibration Signals

Deep Sparse Representation Network for Feature Learning of Vibration Signals and Its Application in Gearbox Fault Diagnosis

One-Dimensional Residual Convolutional Autoencoder Based Feature Learning for Gearbox Fault Diagnosis

A Convolutional Neural Network Based Feature Learning and Fault Diagnosis Method for the Condition Monitoring of Gearbox

Multiscale Dynamic Fusion Global Sparse Network for Gearbox Fault Diagnosis

Evolutionary Manifold Regularized Stacked Denoising Autoencoders for Gearbox Fault Diagnosis

An Automatic and Robust Features Learning Method for Rotating Machinery Fault Diagnosis Based on Contractive Autoencoder.

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

Gearbox fault diagnosis method based on deep convolutional neural network vibration signal image recognition

A neural network constructed by deep learning technique and its application to intelligent fault diagnosis of machines

Application of Deep Neural Network in Gearbox Compound Fault Diagnosis

Fault Diagnosis of Gearbox Based on Deep Residual Shrinkage Network in Noise Environment

Adaptive Morphological Feature Extraction and Support Vector Regressive Classification for Gearbox Fault Diagnosis.

Discriminative Sparse Autoencoder for Gearbox Fault Diagnosis Toward Complex Vibration Signals

Establishment of a deep learning network based on feature extraction and its application in gearbox fault diagnosis

Fault Diagnosis Method for Planetary Gearbox Based on Intrinsic Feature Extraction and Attention Mechanism

A New Deep Transfer Learning Network Based on Convolutional Auto-Encoder for Mechanical Fault Diagnosis

End-To-End Convolutional Neural Network Model For Gear Fault Diagnosis Based On Sound Signals