Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

TFN: An Interpretable Neural Network with Time-Frequency Transform Embedded for Intelligent Fault Diagnosis

Qian Chen, Xingjian Dong, Guowei Tu, Dong Wang, Baoxuan Zhao, Zhike Peng

TL;DR

A novel interpretable neural network termed as Time-Frequency Network (TFN), where the physically meaningful time-frequency transform (TFT) method is embedded into the traditional convolutional layer as an adaptive preprocessing layer.

Abstract

Convolutional Neural Networks (CNNs) are widely used in fault diagnosis of mechanical systems due to their powerful feature extraction and classification capabilities. However, the CNN is a typical black-box model, and the mechanism of CNN's decision-making are not clear, which limits its application in high-reliability-required fault diagnosis scenarios. To tackle this issue, we propose a novel interpretable neural network termed as Time-Frequency Network (TFN), where the physically meaningful time-frequency transform (TFT) method is embedded into the traditional convolutional layer as an adaptive preprocessing layer. This preprocessing layer named as time-frequency convolutional (TFconv) layer, is constrained by a well-designed kernel function to extract fault-related time-frequency information. It not only improves the diagnostic performance but also reveals the logical foundation of the CNN prediction in the frequency domain. Different TFT methods correspond to different kernel functions of the TFconv layer. In this study, four typical TFT methods are considered to formulate the TFNs and their effectiveness and interpretability are proved through three mechanical fault diagnosis experiments. Experimental results also show that the proposed TFconv layer can be easily generalized to other CNNs with different depths. The code of TFN is available on https://github.com/ChenQian0618/TFN.

TFN: An Interpretable Neural Network with Time-Frequency Transform Embedded for Intelligent Fault Diagnosis

TL;DR

A novel interpretable neural network termed as Time-Frequency Network (TFN), where the physically meaningful time-frequency transform (TFT) method is embedded into the traditional convolutional layer as an adaptive preprocessing layer.

Abstract

Convolutional Neural Networks (CNNs) are widely used in fault diagnosis of mechanical systems due to their powerful feature extraction and classification capabilities. However, the CNN is a typical black-box model, and the mechanism of CNN's decision-making are not clear, which limits its application in high-reliability-required fault diagnosis scenarios. To tackle this issue, we propose a novel interpretable neural network termed as Time-Frequency Network (TFN), where the physically meaningful time-frequency transform (TFT) method is embedded into the traditional convolutional layer as an adaptive preprocessing layer. This preprocessing layer named as time-frequency convolutional (TFconv) layer, is constrained by a well-designed kernel function to extract fault-related time-frequency information. It not only improves the diagnostic performance but also reveals the logical foundation of the CNN prediction in the frequency domain. Different TFT methods correspond to different kernel functions of the TFconv layer. In this study, four typical TFT methods are considered to formulate the TFNs and their effectiveness and interpretability are proved through three mechanical fault diagnosis experiments. Experimental results also show that the proposed TFconv layer can be easily generalized to other CNNs with different depths. The code of TFN is available on https://github.com/ChenQian0618/TFN.
Paper Structure (19 sections, 17 equations, 19 figures, 5 tables)

This paper contains 19 sections, 17 equations, 19 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Prilimilary
  3. Time-Frequency Transform
  4. CNN
  5. Methodology
  6. Structure of TFconv Layer
  7. Kernel Function of TFconv Layer
  8. Interpretability of TFconv Layer
  9. Fault Diagnosis Using TFN
  10. Experiment
  11. CWRU Public Bearing Dataset
  12. Planetary Gearbox Dataset
  13. Aerospace Bearing Dataset
  14. Discussion
  15. Comparison with Contrast Models
  16. ...and 4 more sections

Figures (19)

  • Figure 1: The process of inner product based TFT.
  • Figure 2: The process of traditional convolutional layer in CNN.
  • Figure 3: The process of TFconv layer.
  • Figure 4: The time-domain and frequency-domain diagrams of three kernel functions of TFconv layer. (a) STTF. (b) Chirplet. (c) Morlet Wavelet.
  • Figure 5: The comparison between C-FR and O-FR of initialized traditional convolutional layer and that of initialized TFconv layers. (a) C-FR of traditional convolutional layer. (b) C-FR of TFconv layer with STTF kernel. (c) C-FR of TFconv layer with Morlet wavelet kernel. (d) O-FR of traditional convolutional layer and TFconv layer with STFT kernel (solid line represents O-FR while dashed line represents C-FR).
  • ...and 14 more figures