Sistema de manutenção preditiva de falhas em válvulas em um processo industrial utilizando inteligência artificial.

Abstract

This paper presents the development of an artificial intelligence platform for the valve failure prediction in an industrial process, with the objective of avoiding stop the production line due to unforeseen breaks, improve the operator's routine and reduce maintenance costs. The system was developed using maintenance, automation, business intelligence and machine learning concepts to develop a prediction system. The project was developed using tools such as Node-Red, for data collection, InfluxDB for data storage and the framework Flask for the creation of a web page, responsible for dashboard visualization. For the failure prediction, machine learning models such as linear regression and K-Means were used. The results obtained with the creation of the tool and also the prediction are presented.