抄録
The fault diagnosis by acoustic signals on reciprocal engine is more difficult than independent mechanical elements, because almost reciprocal engine is composed of great number of structural bodies and mechanical elements. Reciprocal engine find it hard to identify of fault elements to exist a number of noise sources within the engine. This paper experimentally examines the possibility of diagnose fault conditions about a four stroke outboard marine engine based on generated acoustic signals measured by two microphones. Artificially assumed fault condition in the outboard engine is concerning intake and exhaust valve spring. Identification of above the fault conditions has been carried out based on cross correlation function of close two points signals and paralleled ARX model (Auto Regressive with Extra Input) at diagnosed location within the engine. It is possible to diagnose above fault condition by the proposal SN ratio improvement method.
