Métodos não invasivos para detecção e isolamento de falhas em motores de combustão interna baseados em dimensões fractais e análise multiresolução wavelet

Abstract

The present work presents a totally noninvasive system for detection and isolation of internal combustion engine faults through the sound signal processing. An acquisition system was developed whose data is transmitted to a smartphone in which the signal is processed and the user has access to the information. A study of the chaotic behavior of the vehicle was performed and the feasibility of using the fractal dimensions as a tool to diagnose engine misfire and alternator belt problems was verified. An artificial neural network is used for fault classification using fractal dimension data extracted from the sounds emitted by the running engine. For comparison purposes, a strategy based on wavelet multiresolution analysis was also implemented. The proposed solution enables low-cost, non-contact vehicle diagnostics without the need for sensor installation and in real time. The system and method were validated through experimental tests with a success rate of approximately 99% for the faults under consideration.