Application of Bayes Classifier and Entropy of Vibration Signals to Diagnose Damage of Head Gasket in Internal Combustion Engine of a Car

Currently applied diagnostic systems of technical condition use more and more advanced methods of gathering and processing input data. Many scientific research centres all over the world deal with problems connected with this topic. Up till now no such unified set of guidelines has been constructed which would allow for construction of properly functioning diagnostic system no matter which object is chosen. That is why there is a constant need to test the possibilities of application of existing methods and their modification to match given diagnostic cases. In this article the results of tests conducted with the use of Bayes classifier for diagnostic purposes are presented. The purpose of the tests was diagnosis of technical condition of head gasket in internal combustion engine of a car. The source of information about technical condition was the vibration signal measured in various measurement points. In order to describe the character of changes occurring in vibration signal the measurement in form of entropy was marked for decomposed signal with the use of discrete wavelet transform (DWT).