Multi-domain sequential signature analysis for machinery intelligent diagnosis

Feature extraction plays an important role in machinery fault diagnosis and prognosis. The features extracted from time, frequency and time-frequency domains are widely investigated to describe the properties of overall signal from different perspectives, seldom considering the sequential characteristic of time-series signal in which the fault information may be embedded. This paper investigates a novel approach combing modified Symbolic Aggregate approXimation (SAX) framework and Kernel Principal Component Analysis (KPCA) to extract fault information by analyzing sequential pattern in time-series signal for fault diagnosis. SAX reduces the dimensionality of raw data by transforming the original real valued time series into a discrete one with analyzing signal sequential characteristic and then multiple features are fused by KPCA for fault classification. The proposed approach has high computation efficiency and feature extraction accuracy. Experimental studies on reciprocating compressor valve demonstrate that the presented approach outperforms the methods of SAX-entropy using support vector machine for classification.