Title :
Application of HSMM on NC machine´s state recognition
Author :
Qiang, Huang ; Zhihua, Ding ; Xiao, Zhang
Author_Institution :
Sch. of Mech. & Mater. Eng., Jiujiang Univ., Jiujiang, China
Abstract :
It is significant to identify the running-states of NC machines for ensuring the machining accuracy and running stability. Vibration diagnosis is an on-line prognostics and diagnosis technique by picking-up the frequency characters of the vibration signal on NC machine. In the paper, combining with the wavelet noise reduction and character extraction with varying scales, the Hidden Semi-Markov model is built by the example of headstock bearing abrasion to recognize the running-states effectively. According to experiment and simulation researches, it indicates that the veracity of identification is 96.7% in the 120 test samples after training the HSMM with 80 training samples. This fault diagnosis method is satisfied for the engineering demand, and it can be applied for vibration analysis for other complex machineries.
Keywords :
abrasion; fault diagnosis; hidden Markov models; machine bearings; machining; mechanical engineering computing; numerical control; vibrations; HSMM application; NC machine state recognition; character extraction; diagnosis technique; headstock bearing abrasion; hidden semi Markov model; noise reduction; vibration diagnosis; Background noise; Character recognition; Computer numerical control; Fault diagnosis; Frequency; Noise reduction; Signal analysis; Stability; Vibrations; Wavelet transforms; Hidden Semi-Markov Model; NC machine; state recognition;
Conference_Titel :
E-Health Networking, Digital Ecosystems and Technologies (EDT), 2010 International Conference on
Print_ISBN :
978-1-4244-5514-0
DOI :
10.1109/EDT.2010.5496609
