Highly Accurate Machine Fault Diagnosis Using Deep Transfer Learning

By Siyu Shao, Stephen Marcus McAleer, Ruqiang Yan, Pierre Baldi. IEEE Transactions on Industrial Informatics, 2019

1,371 citations recorded by the source corpus.

We develop a novel deep learning framework to achieve highly accurate machine fault diagnosis using transfer learning to enable and accelerate the training of deep neural network. Compared with existing methods, the proposed method is faster to train and more accurate. First, original sensor data are converted to images by conducting a Wavelet transformation to obtain time-frequency distributions. Next, a pretrained network is used to extract lower level features. The labeled time-frequency images are then used to fine-tune the higher levels of the neural network architecture. This paper creates a machine fault diagnosis pipeline and experiments are carried out to verify the effectiveness and generalization of the pipeline on three main mechanical datasets including induction motors, gearboxes, and bearings with sizes of 6000, 9000, and 5000 time series samples, respectively. We achieve state-of-the-art results on each dataset, with most datasets showing test accuracy near 100%, and in the gearbox dataset, we achieve significant improvement from 94.8% to 99.64%. We created a repository including these datasets located at mlmechanics.ics.uci.edu.

Subfields: Artificial Intelligence, Control and Systems Engineering

View Paper: https://www.semanticscholar.org/paper/22adb2413901b74128f2a02584dafa77afbd8d60

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Long Wen
2. Liang Gao
3. Xinyu Li
4. Chunhui Zhao
5. Liangjun Feng
6. Ming Zhao
7. Jing Lin
8. Shibin Wang

In Control and Systems Engineering

1. Liang Gao
2. Xinyu Li
3. Long Wen
4. Yihao Xue
5. Mengjie Huang
6. Rui Yang
7. Zidong Wang
8. Roberto Ferrero

Citations in the corpus

• A transfer convolutional neural network for fault diagnosis based on ResNet-50 (2019, methodology): “Shao et al. [39] proposed a new deep transfer learning method by using VGG16 and frozen the first three convolution blocks as feature extractor.”
• A comprehensive review on convolutional neural network in machine fault diagnosis (2020, methodology): “transferred to new tasks with proper fine-tuning. In addition, they designed three transfer learning strategies to investigate the feature transferability in the different network levels. Shao et al. [201] employed the WT to convert raw signals into images and developed a deep transfer framework for”
• Intelligent fault diagnosis of machinery using digital twin-assisted deep transfer learning (2021, methodology): “Feature extraction involves extracting the features from the collected data through different domain analysis activities such as those in time domain, frequency domain, and time-frequency domain [6].”
• Potential, challenges and future directions for deep learning in prognostics and health management applications (2020, background): “These works (Shao et al., 2018; Cao et al., 2018) convert raw signals into graphical images as input, and then fine tune ImageNet pre-trained image classification models for their own fault diagnosis tasks.”
• Applications of Unsupervised Deep Transfer Learning to Intelligent Fault Diagnosis: A Survey and Comparative Study (2019, methodology): “33 and Appendix A called PU-Types.”
• Fault Description Based Attribute Transfer for Zero-Sample Industrial Fault Diagnosis (2021, background): “To overcome the trouble in collecting samples for certain faults, an acceptable way is to apply the knowledge learned from some easily obtained or historical faults (training faults) to those faults which are difficult or expensive to collect (target faults) [13], [14].”
• Intelligent Fault Diagnosis of Rotor-Bearing System Under Varying Working Conditions With Modified Transfer Convolutional Neural Network and Thermal Images (2020, background): “In addition, it is challenging to train an excellent CNN from scratch only using a few samples [24].”
• Understanding and Learning Discriminant Features based on Multiattention 1DCNN for Wheelset Bearing Fault Diagnosis (2020, methodology): “For example, the input layer of the 2DCNN in [27] is 224 × 224 × 3.”
• Deep Transfer Learning for Bearing Fault Diagnosis: A Systematic Review Since 2016 (2023, methodology): “[67], frequency-domain analysis [68], [69], [70], [71], and time–frequency-domain analysis [72], [73], [74], [75], [76],”
• Deep Learning Algorithms for Rotating Machinery Intelligent Diagnosis: An Open Source Benchmark Study (2020, methodology): “SEU gearbox datasets were provided by Southeast University [45, 46].”