Optimal scheduling of battery storage systems and thermal power plants for supply-demand balance

This paper focuses on the day-ahead scheduling problem of generating power for thermal power plants and charging/discharging battery energy storage systems based on interval predictions of photovoltaic power. Our previous approach to this problem used the Jacobian of a solution with respect to the variation in demand. However, this study was limited in the sense that the output capacity constraints of thermal power plants were not taken into account. To overcome this problem, we introduce a virtual thermal plant and apply several properties of an M-matrix. To provide guidelines for the amount of power that should be generated, we determine the exact regulating capacity for each thermal power plant and a storage battery so as to maintain the supply-demand balance. The efficiency of the proposed method is verified numerically. We show that a brute-force Monte Carlo method cannot estimate the exact regulating capacity of the thermal power plants when the batteries effectively balance the supply and demand. Moreover, it is found that, unless their deterioration cost decreases significantly, storage batteries cannot be effectively utilized.