An Imbalanced Fault Diagnosis Method Based on Residual Wasserstein Generative Adversarial Network

  • Li Li Tang
  • , Yuan Xu
  • , Wei Ke
  • , Ming Qing Zhang
  • , Chong Xing Ji

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In actual industrial processes, operations generally remain in a normal state for an extended period, with faults occurring less frequently. Since the amount of data representing faults is small, there is an imbalance between normal samples and fault samples, which affects the accuracy of fault diagnosis. To solve this problem, this paper proposes an imbalanced fault diagnosis method based on the Residual Wasserstein Generative Adversarial Network(ResWGAN). Firstly, shortcut connections are added to the hidden layers of both the generator and the discriminator to form a residual network, replacing the fully connected layers. Secondly, ResWGAN is trained to generate virtual fault samples and construct a balanced data set. Thirdly, a one-dimensional convolutional neural network (1D CNN) is used to build a classifier to achieve fault diagnosis. Finally, the experimental results on the Tennessee Eastman process show that the proposed method has a high diagnostic accuracy.

Original languageEnglish
Title of host publicationProceedings of 2025 IEEE 14th Data Driven Control and Learning Systems Conference, DDCLS 2025
EditorsMingxuan Sun, Ronghu Chi
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages80-85
Number of pages6
ISBN (Electronic)9798350357318
DOIs
Publication statusPublished - 2025
Event14th IEEE Data Driven Control and Learning Systems Conference, DDCLS 2025 - Wuxi, China
Duration: 9 May 202511 May 2025

Publication series

NameProceedings of 2025 IEEE 14th Data Driven Control and Learning Systems Conference, DDCLS 2025

Conference

Conference14th IEEE Data Driven Control and Learning Systems Conference, DDCLS 2025
Country/TerritoryChina
CityWuxi
Period9/05/2511/05/25

Keywords

  • Fault Diagnosis
  • Imbalanced Data
  • Residual Network
  • Tennessee Eastman Process
  • Wasserstein Generative Adversarial Networks

Fingerprint

Dive into the research topics of 'An Imbalanced Fault Diagnosis Method Based on Residual Wasserstein Generative Adversarial Network'. Together they form a unique fingerprint.

Cite this