Providing multiple services on multiple time scales with energy storage

点击翻译按钮获取中文摘要

Energy storage is an important source of flexibility in electric power systems. Its uses include balancing renewable variability, reducing peak power consumption to avoid using expensive peaker plants, improving system efficiency by shifting demand from day to night, and providing regulation to maintain stability on fast time scales. A current challenge is choosing how much of each service a storage device should provide. For example, if a battery commits a portion of its capacity in a day-ahead market for spinning reserve, then that capacity will not be available for providing regulation in real time. Although the cost of technologies such as batteries, flywheels, and supercapacitors is dropping, storage remains expensive. This makes it critical to derive maximum value from installed storage assets, and the full value of a storage device can only be realized if it provides multiple services. In the proposed research, we will design algorithms that enable energy storage to optimally provide multiple services simultaneously. We will also design strategies for storage to bid its capacity in multiple, distinct markets. We will model the problem as a dynamic program. Dynamic programming has been previously used to obtain control policies for energy storage providing individual services. We will extend this line of work to the case of storage providing multiple services, seeking analytical policies where possible and using approximate techniques in other cases.