Scheduling distributed energy storage units to provide multiple services under forecast error

When energy storage units (ESUs) within the distribution grid, such as batteries, provide local services such as supporting the integration of photovoltaics, peak shaving, and infrastructure upgrade deferral, they are inactive or only partially used most of time. Moreover, they are often not profitable because of their high investment costs. Their unused capacities could be used to provide power system services, such as frequency control, allowing them to generate additional revenues. However, individual units might not be available to provide system services over the entire frequency control contract duration, since they must also provide their local services. This paper shows how a set of distributed ESUs can simultaneously provide local services individually and system services in aggregate. Using a model predictive control approach, a central scheduler dynamically allocates parts of the energy and power capacities of each ESU to either the local or grid service with the objective of maximizing the profit of the aggregation. A key contribution of this paper is the development of an algorithm that handles both resource aggregation and optimal provision of multiple services. We find that multitasking can almost double an ESU's profits as compared with a single-service approach, and that the benefits from aggregation increase with the frequency control contract duration and with the forecast error.