Optimizing electrical and thermal energy storage systems for hour-ahead integrated demand response in industries

Industrial multi-energy microgrids (MEMGs) are increasingly challenged to meet electrical and thermal energy demands at minimal cost without disrupting operational processes. The current literature lacks strategies for optimizing electrical and thermal energy storage systems for integrated demand response (IDR), which can reduce industrial energy costs without disrupting operations. Moreover, including new technologies such as concentrated photovoltaic thermal (CPVT) panels and energy conversion devices is seldom explored in industrial MEMGs. This study addresses these gaps by developing an artificial intelligence (AI)-based optimization framework and applying it to enhance industrial MEMG performance. The AI contribution includes the formulation of a non-convex hour-ahead optimization problem and the use of a hybrid algorithm combining state-action-reward-state-action (SARSA) reinforcement learning with variants of differential evolution to solve the optimization problem. Additionally, a bidirectional long-short-term-memory networks-based deep stacked sequence-to-sequence autoencoder is employed for accurate hour-ahead forecasting of electrical and thermal energy from CPVT panels and hybrid structure of wind energy conversion systems-superconducting magnetic energy storage systems. The deployment of these AI methods optimize operation of MEMG configuration, specifically aimed at reducing energy bills and energy storage deterioration costs. The model is implemented in MATLAB, where it evaluates the profitability of time-of-use (TOU) and real-time pricing (RTP) tariffs for industrial users adopting hour-ahead IDR strategies. The results demonstrate the potential of hour-ahead IDR strategies in saving up to 92.4 % on energy bills and generating up to 48.2 % in revenue. • Optimization of an hour-ahead demand response for an industrial energy system. • Forecasting electrical and thermal energy for concentrated photovoltaic-thermal panels. • Improved architecture of an industrial energy system. • An hour-ahead demand response strategy provides savings of up to 92.4 %.