A Method of Fault Detection on Diesel Engine Based on EMD-Fractal Dimension and Fuzzy C-Mean Clustering Algorithm

摘要

For the non-stationary characteristics of vibration signal and fuzzy characteristics of feature parameter, a method based on EMD-fractal dimension and FCM is proposed for feature extraction and pattern recognition of diesel engine mechanical fault. Firstly decompose vibration signal by EMD, choose IMFs can reflect fault characteristic information better according to the correlation factor, and compute fractal dimension of the selected IMFs as feature vector, which is used as input of FCM after standardization. The optimized classified matrix and clustering centers are obtained. By calculating the nearness degree between the unknown-fault samples and the known-fault ones, the fault pattern is identified at last. The experimental results express that this method can diagnose faults of the crank-shaft bearing of diesel effectively.