Fault diagnosis method of rolling bearings under varying working conditions based on deep feature transfer

  • PDF / 1,257,711 Bytes
  • 9 Pages / 595.22 x 842 pts (A4) Page_size
  • 108 Downloads / 246 Views

DOWNLOAD

REPORT


DOI 10.1007/s12206-020-1003-9

Journal of Mechanical Science and Technology 34 (11) 2020 Original Article DOI 10.1007/s12206-020-1003-9 Keywords: · Deep feature transfer · Fault diagnosis · Rolling bearing · Varying working condition

Correspondence to: Yujing Wang [email protected]

Citation: Kang, S., Qiao, C., Wang, Y., Wang, Q., Hu, M., Mikulovich, V. I. (2020). Fault diagnosis method of rolling bearings under varying working conditions based on deep feature transfer. Journal of Mechanical Science and Technology 34 (11) (2020) 4383~4391. http://doi.org/10.1007/s12206-020-1003-9

Received December 16th, 2019 Revised

March 7th, 2020

Accepted August 17th, 2020

Fault diagnosis method of rolling bearings under varying working conditions based on deep feature transfer Shouqiang Kang1, Chunyang Qiao1, Yujing Wang1, Qingyan Wang1, Mingwu Hu1 and V. I. Mikulovich2 1

School of Electrical and Electronic Engineering, Harbin University of Science and Technology, Harbin 2 150080, China, Belarusian State University, Minsk 220030, Belarus

Abstract

Rolling bearing vibration data and their labels are difficult or impossible to obtain under varying working conditions. Thus, multistate identification of different fault positions and degradation degrees has relatively low accuracy. This paper proposes a fault diagnosis method based on deep feature transfer. Sparse denoising autoencoder extracts deep features of the frequency-domain amplitude sequences of rolling bearing vibration signals, and the features are used to compose feature sample sets of the source and target domains. Joint geometrical and statistical alignment adaptively processes feature samples of the source and target domains, and this way reduces the distribution divergence and subspace transform shift of the inter-domain feature samples. The classification is achieved using softmax. Experimental results show that the feature visualization effect by visualization algorithm t-SNE using the proposed method is better than those of other methods in this paper. A higher accuracy can also be achieved for rolling bearing fault diagnosis under varying working conditions.

† Recommended by Editor No-cheol Park

1. Introduction

© The Korean Society of Mechanical Engineers and Springer-Verlag GmbH Germany, part of Springer Nature 2020

Normal operation of rolling bearings, which are an important component of rotary machine, is important in ensuring the performance of production equipment [1]. In practice, the working conditions of rolling bearings change constantly, and this condition directly affects the bearing vibration characteristics [2]. The traditional fault diagnosis methods under constant working conditions is prone to misdiagnosis or missed diagnosis under varying working conditions [3, 4] Therefore, accurately identifying the running state of rolling bearings under varying working conditions is important for the healthy operation of mechanical equipment. In recent years, rolling bearing fault identification under different working conditions and operational st