An energy data-driven decision support system for high-performance manufacturing industries

Abstract

To maintain high performance in the looming economic recession, manufacturers are looking for ways to reduce plant downtime, maximise asset utilisation, and improve energy efficiency. In this paper, a unified decision support system (DSS) is proposed, which uses real-time energy measurements and process operational states to make effective decisions, enabling high-performance manufacturing. To reduce the number of required sensors and amount of logged data, our proposed DSS includes an intelligent framework which identifies the process operational states based on energy measurements. This process identification framework uses Haar transform and empirical Bayesian (EBayes) threshold to segment the power data and support vector machines (SVMs) to cluster the power segments into groups according to the underlying process operational states. To justify our proposed framework, comparative experiments with an existing framework are evaluated on two industrial applications, an injection moulding system and a stamping system. Experiment results show that our proposed framework is more effective in identifying the process operational states using the energy patterns.