Fault Diagnosis of Gearbox Based on Wavelet Transforms and Neural Networks

摘要：

In this paper, an actual system based on wavelet transform and artificial neural networks was established to diagnose different types of fault in a gearbox. As a key step, biorthogonal wavelet was used to denoise in feature extraction of signals because of its properties of compact support, high vanishing moment and symmetry. Consequently, a multi-layer perceptron network were designed to diagnose the fault status with feature vectors as inputs. In order to improve the network learning speed and stability, Levenberg-Marquardt algorithm was used to train the network. The present classification accuracy indicates the effectiveness of gearbox failure diagnosis.

关键词： fault diagnosis gearbox biorthogonal wavelet neural network

作者: Guangbin ZHANG Yunjian GE Yongjiu LIU

作者单位: Department of Automation, University of Science & Technology of China, Hefei, 230026, China Institut Institute of Intelligent Machines, Chinese Academy of Sciences, Hefei, 230031, China Department of Automation, University of Science & Technology of China, Hefei, 230026, China Institut

会议类型: 国际会议

会议名称: 2011 International Conference on Machanical Engineering,Materials and Energy(2011年机械工程、材料与能源国际会议 ICMEME 2011)

会议地点: 大连

会议语种:英文

页码: 919-922

在线出版日期: 2011-10-19（万方平台首次上网日期，不代表论文的发表时间）