Análise de um plano para implementação de manutenção preditiva baseada na análise de vibração em máquinas-ferramentas nas oficinas do DEMA da UEM

Abstract

This work presents a study focused on the analysis and proposal of a predictive maintenance plan based on vibration analysis, with specific application to the workshops of the Department of Mechanical Engineering at Eduardo Mondlane University. The research is grounded in the increasing demand for efficient maintenance strategies in academic environments, aiming to ensure the availability and reliability of machine tools used in laboratory practices. The study is structured in four stages: a theoretical review of predictive maintenance and vibration analysis; development of a Python-based computational tool capable of processing vibration signals and identifying failures; formulation of a conceptual predictive maintenance plan; and evaluation of its implementation feasibility. Failures signals such as unbalance, misalignment, mechanical looseness, and bearing defects were simulated, which allowed validating the effectiveness of the proposed tool. In conclusion, the study presents a replicable model for academic workshops, combining maintenance-engineering principles with practical low-cost technological solutions.