Adaptive energy management in institutional building clusters (IBCs) : a data-driven approach

Despite low building density compared to urban areas, institutional building clusters (IBCs) hold substantial potential for enhancing sustainability. As higher education expands globally, IBCs often exceed their contracted electricity limits during peak demand, triggering penalties up to 100 times the regular rate. These charges can account for nearly a quarter of an institution’s annual energy costs, reaching millions for large universities. To reduce this burden, an adaptive energy management (AEM) framework is essential for improving energy efficiency and economic resilience during peak periods. To meet this imperative, this research developed a data-driven AEM framework to investigate building energy trends, determine buildings leading to peak demand scenarios, predict their future peak instances, and optimize peak demands considering onsite renewable energy generation and storage. The study initiates by scrutinizing state-of-the-art technologies and techniques influencing the energy and peak demand dynamics within IBCs. Further, a quantile-based data analysis technique was implemented to identify electrically inefficient buildings in IBCs. The results indicated that electrical fluctuations alone did not necessarily result in peak demands; a comprehensive analysis of both fluctuations and the overall impact of buildings is required for holistic comparison. Furthermore, considering climatic influences and interbuilding effects, a novel deep learning network was developed to simultaneously forecast future demand across multiple buildings. The network achieved a prediction accuracy improvement of 1.8% to 10.9% compared to existing methods. Lastly, an optimization strategy was implemented, conceptualizing IBC as an energy hub consisting of renewable energy and energy storage technologies to optimize peak demand scenarios. The findings indicate that, given current energy pricing, as well as equipment investment, operation, and maintenance costs, achieving a positive net present value requires an interest rate below 4% and solar efficiency exceeding 16%. Furthermore, the battery energy and power costs must remain below $CAD 280 and $CAD 420, respectively, to ensure economic feasibility while optimizing peak demand over the product's lifetime. The research outcomes carry profound implications, offering valuable insights and benefits for facility managers and policymakers to exercise sustainable operations in IBCs.