Sistema integrado de análise e predição de indicadores de desempenho de um processo industrial

Abstract

This paper presents the construction of an integrated system for analysis and prediction of indicators related to an industrial process, which aims to support decision making in the business environment, based on historical data. The developed system uses the concepts of Business Intelligence for the construction of interactive dashboards, and Machine Learning for modeling predictive algorithms capable of anticipating possible deviations in the finished product. For this, a real system was built using OPC, SQL Server and Power BI technologies for monitoring and analysis. For machine learning, we used the K-NN classification model, using the Python language.The validation of the created tool is presented, besides the results obtained from its use in the proposed industrial environment.