An Energy Storage System (ESS) is one of the feasible resources for grid stability as a device or system that is able to store and supply electric energy at the required time. Moreover, Demand Response (DR) is highly considered as one of the promising ways to improve the system flexibility by accommodating variable generation source in electricity markets. The South Korea government are making huge efforts to bring more participation of DR resources based on attractive incentive and free competition in electricity market. They have initiated DR market to allow the participation of customer DR resources into electricity. However, the initial cost of ESS installation is still expensive so that its the economic benefit is not assured. Therefore, it is important to develop the optimal scheduling algorithm for ESS to maximize the profit.
In this manner, this thesis developed three ESS scheduling algorithms: the one is the optimal ESS scheduling algorithm with Photovoltaic (PV) generation, and the other two algorithms are multi-purpose use of ESS to both reduce peak demand and participate in the DR market. For this, 24-hour load forecasting model is developed by combining two load forecasting models: a very-short-term forecasting model by using the previous intra-hour information and a short-term forecasting model by using the previous day information. The both 24-hour load forecasting models use two meteorological factors including temperature and humidity. Furthermore, 24-hour PV generation forecasting model is developed by combining two PV generation forecasting models: a very-short-term forecasting model by using the weather forecasting data and a short-term forecasting model by previous intra-hour information. The both PV generation forecasting model uses ASHRAE Clear-Sky model for estimating the irradiance.
Then, the forecasted load and PV generations are input to the 24-hour optimal ESS scheduling algorithm. This algorithm has 3-objective function as follows: firstly, it is to maximize the customer’s revenue by self-consumption, minimize the peak load for reducing the base charge and, minimize the number of charging and discharging cycles.
The multi-purpose use of ESS scheduling algorithms also use the 24-hour load forecasting model and have similar processes with 1st algorithm. The 2nd algorithm is designed to maintain the operation for self-consumption maximum to maximize the profit and then, the remaining capacity of ESS is used to participate in DR market. The 3rd algorithm is designed to maintain its self-consumption minimum so that ESS can participate in DR market as much as possible. All the algorithms have same objectives to minimize the peak demand to reduce the base charge and to minimize the charging and discharging cycles to lengthen the life expectancy of ESS.
Then, DR market in South Korea is investigated to develop the operational strategy of ESS to participate in DR market by using the proposed algorithms. Based on these strategies, the operational strategy of ESS to participate in DR market is developed in this thesis.
The developed algorithms are validated using the actual industrial load profiles with ESS. Furthermore, the cost-benefit analysis is performed when ESS applying multi-purpose use of ESS scheduling algorithms when participate or not in DR market.
