Anomaly detection in time series data using a combination of wavelets, neural networks and Hilbert transform

摘要

Real time detection of anomalies is crucial in structural health monitoring applications as it is used for early detection of structural damage and to identify abnormal operating conditions that can shorten the life of operating structures. A new signal processing algorithm for detecting anomalies in time series data is proposed in this study. The algorithm is expressed as a combination of wavelet analysis, neural networks and Hilbert transform in a sequential manner. The algorithm has been evaluated for a number of benchmark tests, commonly used in the literature, and has been found to perform robustly.