Multiple Misfire Identification by a Wavelet-Based Analysis of Crankshaft Speed Fluctuation

The analysis of the crankshaft speed fluctuation is one of the most investigated and used techniques for the detection and isolation of the misfire events in an internal combustion engine. During the past, lots of methods for single misfire detection were presented in literature but the multiple misfire analysis still represents an open challenge. This paper describes a new analysis technique based on the wavelet approach that allows for both the extraction of the frequency components related to a misfire event, and its localization in the time domain. The detection process is performed by analyzing the crankshaft angular velocity measurement, in order to underline decelerations due to misfire events, and investigate the post-misfire pattern in a suitable domain (wavelet domain). Moreover, the proposed algorithm allows for an easy isolation of the cylinder responsible for the misfire. Experimental results support the validity of the described approach