ISE Graduate Student Colloquium

Title: Energy storage operational modeling to maximize arbitrage value and improve reliability

Presenter: Hyeongjun Kim

Advisor:  Ramteen Sioshansi

Committee:  Antonio Conejo and Huanxing Yang

An energy storage device is widely used to accommodate an uncertainty of generation capacity, create financial benefit, and reduce costs. This research focuses on arbitrage trading, generation capacity adequacy, and frequency regulation of the storage. First, storage can create an arbitrage value using price patterns by charging and discharging energy. When the price is low in the morning, storage buys electricity from the market, and charges it in a storage unit. When the price peaks in early evening or late afternoon, it sells stored energy to the market and earns money. As the price of the real-time electricity market is uncertain, two-stage stochastic programming is used to model this feature of the storage device. Then, the value of stochastic solution is estimated from the objective function value obtained from this result. The second part of the research explores generation capacity adequacy when a power shortage may occur. Arbitrage will decrease when it dispatches energy to recover from a power shortage; thus, the storage will be reluctant to contribute to generation capacity. To prevent the storage device from neglecting a power shortage, a penalty mechanism is applied to the model. Stochastic dynamic programming is used in this study to determine an optimal decision policy. Then, capacity value is estimated using a decision policy made from this dynamic programming model. The capacity value provides information about how big of a storage device is needed to improve the power grid’s system reliability, as well as how to apply a penalty mechanism. Frequency regulation, which helps the power system to maintain frequency at a stable level, is added to the third part of the research. In the frequency regulation market, certain amount of power is reserved for this service, and it will be called by operator when it is needed. The ratio between the capacity reserved and called to provide energy, and the electricity prices are uncertain, so stochastic dynamic programming model is also used for this study. Capacity value is also estimated for this part, and it will help the storage operator to decide on the size of energy storage capacity and policymakers determine how to apply the penalty scheme.