A Fault Diagnosis Model Based on Convolution Neural Network for Wind Turbine Rolling Bearing

2018 4th International Conference on Environmental Science and Material Application

生态环境科学;材料科学

Yang, Zhiling^1 ; Ma, Xiaoshan^1 ; Ma, Yuanchi^2

School of Energy and Mechanical Engineering, North China Electric Power University, Beijing, China^1

School of Renewable Energy, North China Electric Power University, Beijing, China^2

关键词: Case Western Reserve University;    Convolution neural network;    Convolutional neural network;    Expert experience;    Fault diagnosis model;    Feature engineerings;    Operating environment;    Wavelet denoising;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/252/3/032039/pdf
DOI  :  10.1088/1755-1315/252/3/032039

来源: IOP

【 摘 要 】

The rolling bearing has become the component with high failure rate in the wind turbine due to its poor operating environment. This paper proposes a fault diagnosis model based on convolutional neural network for the rolling bearing. The input of model is the short-time Fourier transformed spectrum of the vibration data of the rolling bearing, and the output is the codes of various fault types. The proposed model is verified by the bearing test data of Case Western Reserve University. In addition, the samples with various nosie levels and those after wavelet de-noising are input into the fault diagnosis model respectively, and the anti-noise performance of the proposed model is discussed. The result shows that the proposed model can automatically find fault features and identify various rolling bearings fault, avoiding the expert experience and feature engineering. This makes it more practical and generalizable in the fault diagnosis of rolling bearing of wind turbine.