Uma contribuição para o monitoramento de motores de combustão interna através da análise de sinais de pressão sonora e vibração

Abstract

Research in identification of models as well as parameters, are important and essential to make possible the use of mathematician-computational procedures in the development and improvement of industrial products. In this work the existing correlation between vibration and noise for the main defects in internai combustion engines, such as: valves chatter and noise tensors were studied. The methodology used was based on Signatures Analysis, with spectral and deconvolution techniques for the identification of noise sources. The studied techniques consisted of spectral analysis, Cepstral analysis and the Envelope Spectrum Technique. Through matching the responses of faulty engines with the responses of engines in good conditions, was possible to identify and to diagnosis the existing defects. Viability of the Spectral Analysis and the envelope spectrum technique to identify and to diagnosis faults in internai combustion engines was verified, and the Cepstral Analysis showed to be inefficient for the cases studied.