Data-driven optimization of packed bed thermal energy storage heating system with encapsulated phase change material

Using phase change material (PCM) in domestic water tanks offers enhanced heat storage capacity and improved energy efficiency. Data-driven techniques such as artificial neural networks (ANN) show promise in real-time control and fully harnessing the advantages of thermal energy storage (TES). This study proposes an ANN model for a packed bed TES (PBTES) tank incorporating PCM, demonstrating exceptional accuracy and superior computational efficiency compared to numerical models. The ANN model is subsequently employed in a dynamic simulation of a space heating system, revealing that the PBTES tank outperforms conventional water tanks without PCM by storing an additional 31.3 % of heat. Under varying operating conditions, the PBTES tank achieves an operational cost reduction of up to 40.6 % compared to using only an electric boiler for space heating. Moreover, on the days with lower heating demand, PBTES tanks exhibit excellent peak-shaving capabilities. An optimized operating strategy is also developed with the assistance of the ANN model, showing the ability to reduce operational costs of the PBTES system by 17 %. These findings highlight the potential of data driven techniques for optimizing TES systems and maximizing economic benefits.