Control of mechatronics systems: Ball bearing fault diagnosis using machine learning techniques

Ball bearing fault is one of the main causes of induction motor failure. This paper investigates in the fault diagnosis of ball bearing of three phase induction motor using random forest algorithm and C4.5 decision tree. The bearing conditions are classified to four categories: normal, bearing with inner race fault, bearing with ball fault and bearing with outer race fault. The statistical features used for classification are extracted from mechanical vibration signal in time domain and frequency domain. Principal component analysis (PCA) and linear discriminent analysis (LDA) are used to reduce the dimension and complexity of the feature set. The classification accuracy of random forest algorithm and C4.5 decision tree are analyzed and compared. The experimental results show that the random forest algorithm not only works better than the C4.5 decision tree but also can classify the ball bearing condition effectively.

[1]  Mostefa Mesbah,et al.  Wavelet Decomposition for the Detection and Diagnosis of Faults in Rolling Element Bearings , 2009 .

[2]  K. I. Ramachandran,et al.  Automatic rule learning using decision tree for fuzzy classifier in fault diagnosis of roller bearing , 2007 .

[3]  B. Samanta,et al.  ARTIFICIAL NEURAL NETWORK BASED FAULT DIAGNOSTICS OF ROLLING ELEMENT BEARINGS USING TIME-DOMAIN FEATURES , 2003 .

[4]  Jin Chen,et al.  Decision tree and PCA-based fault diagnosis of rotating machinery , 2007 .

[5]  Dejie Yu,et al.  Application of EMD method and Hilbert spectrum to the fault diagnosis of roller bearings , 2005 .

[6]  Satish C. Sharma,et al.  Fault diagnosis of ball bearings using machine learning methods , 2011, Expert Syst. Appl..

[7]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[8]  Tielin Shi,et al.  A novel fault diagnosis method of bearing based on improved fuzzy ARTMAP and modified distance discriminant technique , 2009, Expert Syst. Appl..

[9]  Yaguo Lei,et al.  Application of a Novel Hybrid Intelligent Method to Compound Fault Diagnosis of Locomotive Roller Bearings , 2008 .

[10]  Zhiyuan Yang,et al.  The Application of Random Forest and Morphology Analysis to Fault Diagnosis on the Chain Box of Ships , 2010, 2010 Third International Symposium on Intelligent Information Technology and Security Informatics.

[11]  Yongyong He,et al.  Hidden Markov model-based fault diagnostics method in speed-up and speed-down process for rotating machinery , 2005 .

[12]  Ngoc-Tu Nguyen,et al.  Fault diagnosis of induction motor using decision tree with an optimal feature selection , 2007, 2007 7th Internatonal Conference on Power Electronics.

[13]  Jianping Xuan,et al.  Application of a modified fuzzy ARTMAP with feature-weight learning for the fault diagnosis of bearing , 2009, Expert Syst. Appl..

[14]  Mahmoud Omid,et al.  An Intelligent Combined Method Based on Power Spectral Density, Decision Trees and Fuzzy Logic for Hydraulic Pumps Fault Diagnosis , 2008 .

[15]  Fidel Ernesto Hernández Montero,et al.  The application of bispectrum on diagnosis of rolling element bearings: A theoretical approach , 2008 .

[16]  J. Ross Quinlan,et al.  C4.5: Programs for Machine Learning , 1992 .