Aplicação de aprendizado de máquina para predição de falhas em equipamentos

Abstract

Equipment failure is a negative phenomenon that occurs in industrial systems, caused by a variety of factors including wear, aging, and adverse operating conditions. These failures can result in equipment downtime, product defects, and potential damage to infrastructure. In na industrial context, equipment failure can lead to significant losses for companies, compromising product quality and operational efficiency. In light of this, this work aims to develop a failure prediction method using machine learning, given its predictive capability through real-time data analysis. For this purpose, a variety of machine learning techniques were gathered to solve the problem. Data processing, training, refinement, and calibration of the models were carried out using the Python programming language and auxiliary libraries such as pandas and scikit-learn. The results show that machine learning can significantly improve the accuracy of predictive maintenance, achieving a detection efficiency of up to 96.2%.