Introduction

International Standard Serial Number:

ISSN 1001-4551

Sponsor:

Zhejiang University;

Zhejiang Machinery and Electrical Group

Edited by:

Editorial of Journal of Mechanical & Electrical Engineering

Chief Editor:

ZHAO Qun

Vice Chief Editor:

TANG ren-zhong,

LUO Xiang-yang

Tel:

86-571-87041360,87239525

Fax:

86-571-87239571

Add:

No.9 Gaoguannong,Daxue Road,Hangzhou,China

P.C:

310009

E-mail:

meem_contribute@163.com

Fault diagnosis method of wind turbine gearbox based on ResNet-ELM and transfer learning
Published:2023-09-20 author:SUN Ye, ZHANG Ze-ming, LIU Xiao-yue. Browse: 1107 Check PDF documents
Fault diagnosis of wind turbine gearbox based on ResNet-ELM and transfer learning


SUN Ye1, ZHANG Ze-ming2, LIU Xiao-yue2

(1.China Academy of Information and Communications, Beijing 100083, China; 

2.School of Electrical Engineering, North China University of Science and Technology,Tangshan 063210, China)


Abstract: Aiming at the problems that the traditional fault diagnosis methods had poor diagnosis effect and poor generalization ability in the wind turbine gearbox under variable operating conditions and mixed fault conditions, a wind turbine gearbox fault diagnosis method(TL-RN-ELM) based on deep residual network (ResNet)-extreme learning machine(ELM) and transfer learning(TL) was proposed. Firstly, the principles of continuous wavelet transform (CWT), convolutional neural network (CNN), ResNet, TL and ELM were introduced. Then, the TL-RN-ELM fault diagnosis model of wind turbine gearbox was established. Finally, the bearing data set and gearbox data set were used to validate the proposed method. Data acquisition and processing were carried out from the CWRU bearing data set and the SEU gearbox data set. The original one-dimensional vibration signal was converted into twodimensional wavelet time-frequency image using CWT, and the ResNet18 model was trained using CWRU bearing data set to generate a pre-training model. The data in the pre-training model was migrated to the SEU gearbox dataset, the module was fine-tuned, features were extracted and input into the ELM classifier, and then the classification results were compared with the other three types of models. The experiment results show that the average accuracy of TL-RN-ELM can reach 98.79% for small sample migration fault diagnosis from bearing to bearing, bearing to gear and mixed fault. Comparing with other methods, the average accuracy rate is increased by 4.73%~9.6%. The research results show that this method has good diagnostic effect and generalization ability.

Key words: gear transmission; transfer learning(TL); deep residual network(ResNet); extreme learning machine(ELM); convolutional neural network(CNN); continuous wavelet transform(CWT); model pre-training; model drift; wavelet time-frequency diagram


Advanced Search
Online Office System
  • Chinese Core Periodicals
  • Chinese Sci-tech Core Periodicals
  • SA, INSPEC Indexed
  • CSA: T Indexed
  • UPD:Indexed


2010 Zhejiang Information Institute of Mechinery Industry

Technical Support：Hangzhou Bory science and technology

You are 1895221 visit this site