Symbolic time series analysis of ultrasonic data for early detection of fatigue damage

This paper presents a novel analytical tool for early detection of fatigue damage in polycrystalline alloys that are commonly used in mechanical structures. Time series data of ultrasonic sensors have been used for anomaly detection in the statistical behaviour of structural materials, where the analysis is based on the principles of symbolic dynamics and automata theory. The performance of the proposed method has been evaluated relative to existing pattern recognition tools, such as neural networks and principal component analysis, for detection of small changes in the statistical characteristics of the observed data sequences. This concept is experimentally validated on a special-purpose test apparatus for 7075-T6 aluminium alloy specimens, where the anomalies accrue from small fatigue crack growth.