A machine-learning based data-oriented pipeline for Prognosis and Health Management Systems

Marcos Leandro Hoffmann Souza, Cristiano André da Costa, Gabriel de Oliveira Ramos

Abstract: Highlights•Presents a data-driven pipeline machine learning based to support decision making.•Focuses on the Prognosis and Health Management System of a petrochemical process.•Uses process control and health monitoring data from critical equipment.•Employs CNN, XGBoost e SHAP to detect anomalies/highlight the important variables.•Analyses System Remaining Useful Life comparing different Deep Learning techniques.