Aplicación de la metodología de patrón de puntos para el diagnóstico de fallo de anulación de funcionamiento de un cilindro en un motor de combustión interna

Abstract

This research work applies the symmetrized dot pattern (SDP) methodology to diagnose faults in internal combustion engine cylinders. Engine sounds are captured under different conditions, processed using the Fourier transform to generate polar graphs, and specific parameters are adjusted in the SDP algorithm. Since SDP images are difficult to interpret visually, convolutional neural networks (CNN) are used to classify the patterns, achieving an accuracy of 81.11% in diagnoses, demonstrating the effectiveness of the SDP technique.