A novel transfer learning fault diagnosis method based on Manifold Embedded Distribution Alignment with a little labeled
- PDF / 3,794,655 Bytes
- 15 Pages / 595.276 x 790.866 pts Page_size
- 13 Downloads / 165 Views
A novel transfer learning fault diagnosis method based on Manifold Embedded Distribution Alignment with a little labeled data Ke Zhao1 · Hongkai Jiang1
· Zhenghong Wu1 · Tengfei Lu1
Received: 22 January 2020 / Accepted: 19 August 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Accurate identification of rolling bearing faults is quite significant for the stable operation of mechanical systems. However, for practical diagnosis issues, it is difficult to obtain abundant labeled data due to the change of operating conditions and complex working environment, which puts forward higher requirements on the ability of the diagnosis methods. To tackle the mentioned problem, a novel transfer learning method based on a little labeled data is proposed, which uses bidirectional gated recurrent unit (BiGRU) and Manifold Embedded Distribution Alignment (MEDA). Firstly, frequency spectrum datasets are utilized to remove the redundant information of raw vibration signals. Secondly, the BiGRU network is constructed to generate auxiliary samples that are utilized as source domain. Finally, MEDA, as the most powerful non-deep transfer learning method, is applied to align the distribution of these auxiliary samples generated by BiGRU and the unlabeled samples from target domain. Experiment results indicate the excellent performance of the proposed method under a little labeled data. Keywords Transfer learning · Bidirectional gated recurrent unit · Manifold Embedded Distribution Alignment
Introduction Rolling bearing is one of the critical components in mechanical systems, so accurate identification of rolling bearing faults is quite significant for stable operation of mechanical systems (Jiang et al. 2013; Zhou et al. 2019; Wang et al. 2017). However, most of the rolling bearings are working under awful conditions, which include high temperature, high torque, high rotating speed, and etc. (Wu et al. 2017; Li et al. 2015). Rolling bearing is inevitable to malfunction taking these working conditions into consideration. Thus, it is of great significance to identify rolling bearing faults efficiently and accurately. Intelligent fault diagnosis has high application value for its ability to deal with large amounts of labeled data and accurately identify the rolling bearing faults (Shao et al. 2020). In 2019, Zheng et al. adopted extreme learning machine to classify the bearing fault conditions, and verified the effectiveness of the method through a large number of labeled data (Zheng et al. 2019). In 2018, Meng et al. proposed an
B 1
Hongkai Jiang [email protected] School of Aeronautics, Northwestern Polytechnical University, Xi’an 710072, China
auto encoder to diagnose the faults of rolling bearing, and a labeled dataset is used to evaluate the proposed method (Meng et al. 2018). In 2017, Lu et al. constructed a hierarchical convolutional neural network to identify the bearing health conditions, and sufficient experiments are conducted to demonstrate the performance of the proposed method (Lu et a
Data Loading...
