An energy management system for industrial manufacturing: A hybrid approach with demand response

This paper presents a novel matheuristic approach for a high-level energy management system (EMS) integrated with demand response, aiming to optimise energy costs and enhance renewable energy utilisation in industrial manufacturing. The primary research objective is to develop a scalable solution procedure capable of tackling the complex, NP-hard problem of energy-aware production scheduling. The system employs a hybrid approach, integrating a Mixed Integer Linear Program (MILP) within the Genetic Algorithm’s (GA) fitness function for job scheduling and minimising total energy costs while maximising renewable energy penetration and guaranteeing production constraints. The EMS is applied to a factory microgrid scenario, considering energy production from wind turbines, photovoltaic panels, and combined heat and power (CHP) plants, alongside battery energy storage systems (BESS). The manufacturing process possesses a number of realistic features, including several stages with parallel unrelated machines with different energy-consumption states, batching in some machines, or setup times, among others. The proposed solution for this case study achieves a 32% reduction in energy costs compared to baseline operation, which requires only seconds of computational effort, demonstrating its effectiveness and scalability for demand-responsive manufacturing environments. Methodology is validated using real production data, providing insights into the potential of this approach to improve both economic and environmental performance in the industrial sector.