The Local Maxima Method for Enhancement of Time-Frequency Map

In this paper a new method of failure detection in rotating machinery is presented. It is based on a vibration time series analysis. A pure vibration signal is decomposed via the short-time Fourier transform (STFT) and new time series for each frequency bin are processed using novel approach called local maxima method. We search for local maxima because they appear in the signal if local damage in bearings or gearbox exists. Due to random character of obtained time series, each maximum occurrence must be checked for its significance. If there are time points for which the average number of local maxima is significantly higher than for the others, then the machine is suspected of being damaged. For healthy condition machinery, the vector of average number of maxima for each time point should not have outliers. The main attention is concentrated on the proper choice of required local maxima significance. The method is illustrated by analysis of very noisy both real and simulated signals. Also possible generalizations of this method are presented.