Approach to Detection of Anomalies of Process Signals Using Classification and Wavelet Transforms
Damir A. Murzagulov, Alexander V. Zyamatin, Павел Михайлович Острасть · 2018 International Russian Automation Conference (RusAutoCon) · 2018
The present-day IT infrastructure development level of industrial enterprises makes it possible to collect and store process information, thereby opening the possibilities to use intelligent data analysis systems. This paper overviews the problem of detecting the anomalies in process signals to improve safety and economic efficiency of the industrial enterprises. The combination of the machine learning and wavelet transform methods are proposed to detect anomalies. We have overviewed the nature of process signals to be taken into consideration as well as the advantages of wavelet analysis against the known methods of digital signals processing.