Intelligent Approach to Valve Clearance Diagnostic in Cars

Abstract

In many research centres all over the world work is nowadays being carried out aimed at compiling a method for diagnostic of machines technical condition. Non-destructive methods, including methods using vibroacoustic phenomena, have a special meaning. As we can read in the literature, vibroacoustic signals generated during the work of technical objects have necessary information connected with the technical condition of those objects. An appropriate transformation of registered signals is the main problem. This paper suggests using a discrete wavelet transform and energy and entropy, which provide the basis for a diagnostic system of valve clearance in a car combustion engine. Research aimed at building a diagnostic system using neural networks was conducted. This approach to diagnosis of the technical condition of combustion engine elements gives a possibility for the future enlargement of an on-board diagnostic system OBD, which is being used nowadays. © Springer-Verlag Berlin Heidelberg 2013.