Improving Operational Availability with a TPM-Integrated AI-Based Supply Management System in a Metalworking Company

Enciso-Ortiz P.; Leon-Zegarra R.; Gutiérrez-Falcón P.C. · Congreso Internacional de Innovacion y Tendencias en Ingenieria, CONIITI (2025) · DOI: 10.1109/coniiti68560.2025.11366179

Resumen

This study addresses the problem of low equipment availability at a metalworking company located in Callao, Peru, dedicated to the manufacture of air suspensions for heavy vehicles. In 2024, equipment availability was 82%, below the industry standard (90%), due to an increase in unplanned downtime and deficiencies in material supply management, a critical situation given that semi-trailer suspensions represent the main sales line. To address this problem, an intelligent supply management system was proposed and implemented, based on limited-memory artificial intelligence and integrated with the Total Productive Maintenance (TPM) methodology. This solution allowed real-time monitoring of machine status, improved forecasting of material requirements, and synchronized logistics activities, thus reducing operational downtime. As a result, availability was increased to 88%, stockout-related downtime was reduced from six to three days, and key maintenance indicators and key indicators such as Mean Time to Repair (MTTR) and Mean Time Between Failures (MTBF) showed marked improvement. The proposal was validated through a TPM pilot test and a simulation of the supply system using Arena software, version 16.

Instituciones: Universidad Peruana de Ciencias Aplicadas