Title :
Gearbox fault prognosis based on CHMM and SVM
Author :
Kang, Jianshe ; Zhang, Xinghui ; Zhao, Jianmin ; Cao, Duanchao
Author_Institution :
Mech. Eng. Coll., Shijiazhuang, China
Abstract :
A new gearbox fault prognosis scheme based on continuous hidden Markov model (CHMM) and support vector machine (SVM) is developed. Based on the features which are the energies of intrinsic mode functions (IMFs) decomposed by empirical mode decomposition (EMD) extracted from normal gearbox vibration signal, a CHMM is trained to model the normal condition. The logarithm of the probability of this CHMM is then used to detect any defects and assess their severity. Then, SVM is used to predict the value of new feature which is the logarithm of the probability. Experimental data collected from a gearbox degradation test is used to verify the efficacy of the new scheme.
Keywords :
fault diagnosis; gears; hidden Markov models; mechanical engineering computing; probability; signal processing; support vector machines; vibrations; SVM; continuous hidden Markov model; defect detection; empirical mode decomposition; gearbox degradation test; gearbox fault prognosis; gearbox vibration signal; intrinsic mode function; probability; support vector machine; Feature extraction; Frequency domain analysis; Hidden Markov models; Kernel; Support vector machines; Vectors; Vibrations; CHMM; EMD; SVM; fault prognosis; gearbox;
Conference_Titel :
Quality, Reliability, Risk, Maintenance, and Safety Engineering (ICQR2MSE), 2012 International Conference on
Conference_Location :
Chengdu
Print_ISBN :
978-1-4673-0786-4
DOI :
10.1109/ICQR2MSE.2012.6246327