Machine Learning for Predictive Maintenance of Industrial Machines

Resumo

This work presents a novel approach to constructing a predictive model using machine learning techniques for the early identification of machine faults for industrial maintenance purposes. The topic is justified by the importance of predictive maintenance in increasing operational efficiency, reducing costs, and improving the reliability of industrial equipment. The early identification of failures makes it possible to carry out preventive interventions, avoiding negative consequences and optimizing the use of industrial resources. Traditional maintenance techniques, such as corrective and preventive maintenance based on fixed intervals, cannot adequately predict unexpected events that could lead to equipment failure. The main objective is to develop and validate a predictive model capable of detecting patterns that indicate possible failures in industrial machinery before they occur. To this end, a hybrid algorithm combining fuzzy logic and machine learning will be applied compared to three traditional models widely used in the literature. The study includes a comparative analysis of the results obtained, evaluating the performance, accuracy, and effectiveness of each model. The advantages and limitations of the developed model will also be discussed, as well as suggestions for future research to improve predictive systems in the sector continuously. This article is divided into sections, starting with the introduction, with a general contextualization of the problem, the arguments, the justifications, the objective, and the structure of the work. The following section presents research into related work. Then the methodology used to develop the fault prediction, and identification models will be presented, followed by the experimental results, and finally the conclusions.