Search Results

Vibration signal measurement and analysis is an efficient non-intrusive method for engine diagnosis. In this paper, we propose a method based on wavelet transform for vibration signature extraction to detect, locate, and diagnose a range of common faults in valve train mechanism and combustion process. For this purpose, a flexible multibody dynamics model has been developed using ADAMS. The model includes the valve train system, cylinder head, and gas pressure effect. The vibration response of the cylinder head surface under simulated faults such as misfire, cylinder pressure variations, and excessive valve clearance is investigated. The spectral properties of vibrations caused by highly transient dynamics are analyzed using continuous wavelet transform. Wavelet decomposition is performed to identify and correlate dynamical response of the system with corresponding sources of excitation. Fault detection procedure is presented and discussed.