Optimal BESS Scheduling with Real-Time Assessment for Time-of-Use Customers Using Particle Swarm Optimization

This study presents an optimal day-ahead operational scheduling strategy for a battery energy storage system(BESS) to achieve savings for a time-of-use(TOU) customer. This optimization problem is solved using a Particle Swarm Optimization(PSO) algorithm with an evaluation of its sensitivity to photovoltaic(PV) and load forecast errors. A new initialization process based on two non-optimized algorithms is proposed to enhance PSO convergence during the optimal scheduling process. By charging the BESS with surplus PV generation or during off-peak periods and discharging it during peak times or when contracted demand limits are exceeded, an initial particle population is provided as an input to the PSO. The state of charge(SOC) is maintained within the normal range by stopping the BESS output selected by the PSO as needed. The PSO's performance is assessed in real-time operation using historical PV and load data. The proposed method is simulated using MATLAB software to verify its effectiveness showing 17% and 10% cost saving over one day and one month, respectively. Additionally, the PSO demonstrates robust performance for forecast error under 10%.