FGDAE: a new machinery anomaly detection method towards complex operating conditions

Shen Yan, Haidong Shao, Zhishan Min, Jiangji Peng, Baoping Cai, Bin Liu

Research output: Contribution to journalArticlepeer-review

110 Citations (Scopus)
59 Downloads (Pure)

Abstract

Recent studies on machinery anomaly detection only based on normal data training models have yielded good results in improving operation reliability. However, most of the studies have problems such as limiting the detection task to a single operating condition and inadequate utilization of multi-channel information. To overcome the above deficiencies, this paper proposes a new machinery anomaly detection method called full graph dynamic autoencoder (FGDAE) towards complex operating conditions. First, a full connected graph (FCG) is developed to obtain the global structure information by establishing structural connections between every two channels. Subsequently, a graph adaptive autoencoder (GAAE) model is constructed to aggregate multi-perspective feature information between channels by adapting changes of the operating conditions and to reconstruct the information containing the essential features of normal data. Finally, a dynamic weight optimization (DWO) strategy is designed to guide the model learning the generalization features by flexibly adjusting the data reconstruction loss weights in each condition. The proposed method performs multi-condition anomaly detection under the challenge of training models with multi-condition unbalanced normal data and achieves better performance compared to other popular anomaly detection methods on the machinery datasets.
Original languageEnglish
Article number109319
Number of pages11
JournalReliability Engineering and System Safety
Volume236
Early online date20 Apr 2023
DOIs
Publication statusPublished - 31 Aug 2023

Keywords

  • machinery anomaly detection
  • complex operating conditions
  • multiple channels
  • full graph dynamic autoencoder
  • weight optimization strategy
  • graph convolution network

Fingerprint

Dive into the research topics of 'FGDAE: a new machinery anomaly detection method towards complex operating conditions'. Together they form a unique fingerprint.

Cite this