LAUSR

Abstract The objective of this work is to identify the presence of damages and to diagnose the damaged component in automotive gearboxes by comparing the vibration signals of the damaged and undamaged systems. The vibration signals were obtained in the gearboxes test bench taking into account ten samples approved by subjective method (based on human hearing) and three gearboxes damaged (two with bearing damaged and one with gear tooth damaged). The signals were obtained through five accelerometers positioned in the samples and each test comprises ten different steps with different coupling configurations of the gears. Different signal analysis techniques based on wavelet transform, mathematic morphology and energy (entropy) were used to verify the presence of damage in the systems. The presence of damage to the systems is verified directly by comparing the energy levels and entropy of the signals of the damaged and undamaged systems for the ten steps tests. A signal processing technique combining pattern spectrum and selective filtering in certain frequencies ranges was used for identification of component failures. This technique will, further on, be part of a diagnostic and quality control system that will be able to evaluate and identify the damaged component at the automotive gearbox.