Maestría en Ingeniería · 2026

Desarrollo de modelo predictivo basado en machine learning para anticipar fallas en el probador de suspensión de la línea de livianos del Centro de Diagnóstico Automotor Certi Express Pereira S.A.S.

Predictive maintenance has become a fundamental mechanism for ensuring the reliability, availability, and efficiency of industrial equipment. In the Colombian automotive sector, Automotive Diagnostic Centers (CDAs) play a crucial role in road safety by ensuring that vehicles meet the technical requirements of national regulations. However, traditional corrective and preventive maintenance methods generate downtime and additional costs that affect the continuous operation of inspection equipment. This research develops a machine learning-based solution to analyze the results of adhesion tests obtained using the suspension tester of the light vehicle line at the Certi Express Pereira S.A.S. Automotive Diagnostic Center. 2,360 test records were collected and, after a cleaning, normalization, and variable analysis process, a predictive model was built capable of detecting patterns associated with potential failures in both the measuring equipment and the inspected vehicles. The results show that the application of machine learning models in automotive diagnostic environments significantly improves the early detection of anomalies, allows for optimized maintenance planning, and contributes to reducing operating costs and downtime.

Palabras clave

Tesauro de su biblioteca