Short-Time Analysis of Combustion Engine Vibroacoustic Signals through Pattern Recognition Techniques

The paper presents a method of analysis of vibroacoustic signals taken from Diesel engines based on short-time analysis. The main aim of the research is to distinguish between normal engine operation and that of misfire. The approach utilizes a time window sliding along a signal in time and observing the changes of given parameters. The set of parameter values creates a multidimensional parameter space where one can observe the time evolution. For recognition and detection of different system states a clustering in the parameter space is suggested. The first results show the possibility of distinguishing some different clusters within the parameter space which may correspond to different engine states.