Digital Acoustic Signal Processing Methods for Diagnosing Electromechanical Systems

Abstract

An effective means of preventing accidents, identifying critical operating modes, and diagnosing equipment failures of electromechanical systems are the methods of functional diagnostics. A number of problems of diagnostics of electromechanical complexes at the present time can be realized by acoustic methods, by analyzing the signals received from working assemblies. The actual scientific task is the formation of a procedure for analyzing acoustic signals generated by working equipment of electromechanical complexes, based on the use of modern methods of digital time series processing in real time. The article analyzes the acoustic signals obtained as a result of an experiment when operating electromechanical equipment. At the first stage of processing, the signals are passed through a low-pass filter and a band-pass filter. The spectra of the amplitudes of the signals before and after filtering, as well as the dynamics of signals in the phase space, are studied. For signals before and after processing, the autoregressive moving average models were calculated and their standard deviations were analyzed. The application of the procedure for analyzing acoustic signals and standard methods for their digital processing will allow real-time decision-making support systems to be implemented with automatic detection (diagnosis at the rate of measurement of diagnostic signals) of machinery equipment malfunctions, their degree of danger and the formation of a list of compensating measures.