Data Analytics Architecture for Energy Efficiency Optimization in Industrial Processes

Abstract

This paper proposes a novel control system that uses ontology models in combination with machine learning algorithms in order to increase the energy efficiency of the hot rolling mill process. The system architecture as well as the implemented modules, that will be developed, are going to be explained by the author. An OPC UA Server will be used as the data acquisition system and platform for a Semantic Ontology Engine. The overall energy usage can be calculated with empirical and physical formula that lead to a mathematical model of the process. This model will be used in combination with Deep Neural Networks to predict the energy demand aiming to implement a model predictive closed-loop control. Furthermore, a new communication interface for the existing control system can be obtained by combining ontologies with the derived information about the energy consumption of the process.