Bi-Level Energy Dispatch Optimisation for Peak Load Shaving in Microgrids with Battery Storage

Peak load periods in electrical networks result in significant power losses, straining generation capacity, and increasing operational costs. Strategically optimising electricity supply from battery storage-based distributed generation (DG) during these peak times can mitigate these losses and enhance grid efficiency through peak load shaving (PLS). However, electrical loads vary throughout the day and night. Thus, identifying peak and off-peak loads would be the most challenging task to complete before scheduling operations for batteries. This study presents a bi-level energy optimisation framework to identify the peak and off-peak loads and plan the energy dispatching operations for battery storage. The bi-level energy optimisation framework is developed in such a way that during the first level, peak load times (PLT), off-peak load times (OPLT), and no operation times (NOT) from the daily time-varying load profiles are identified. During the second level, scheduling of energy supply to and from the batteries is performed using a seven stage battery dispatch controller. Considering the reduction in power losses as a primary objective function, the genetic algorithm (GA) is used to solve the optimisation problem for three time-varying load profiles (industrial, residential, and commercial) using an IEEE 33 bus microgrids network. The numerical results for three case studies have revealed that the proposed method could significantly shave the peak loads by 23.3% in industrial, 18.89% in residential and 10.99% in commercial loads. Due to the shaving of peak loads, the reductions in power loss of 5.73%, 5.44%, and 2.45% in each load profile are also noticed. To further validate the efficacy, the results from the optimisation framework were compared with results using fixed values. The comparison showed that the proposed optimisation approach could achieve maximum peak shaving in microgrids. The PLS, reduction in daily power losses, improvement in load factors, and enhancements in bus voltage profiles for each load confirm that the proposed optimisation approach can be helpful in the future planning of PLS in microgrids.