Impact of forecasting error characteristics on battery sizing in hybrid power systems

Abstract

As we move into a decarbonised electricity network, hybrid renewable power plants (HRPP) will be coupled with batteries to compensate for the intermittent nature of the resource. However, a key question is how to correctly size the battery and for what purpose. We present a novel approach for battery sizing in an HRPP based on forecasting error characteristics. What distinguishes this work from previous studies is the unique approach for sizing the battery based on factors that directly impact the choice of size. We did this by using a rule-based strategy to determine battery energy and power capacities. From the simulation, we observed that the time and frequency components have significantly different impacts on determining battery energy and power capacities. In addition, three important measures for reducing the battery sizes are identified, including improving the forecasting accuracy, removing long-term frequency components, and making use of the combined effect of the forecast bias and battery characteristics. The required power capacity can range from around 1 kW to 36 MW and the required energy capacity from around 3 MWh to 15 GWh in different scenarios for a hybrid power plant co-locating a 15MW wind farm and a 15MW solar farm. The main key discovery is the unexpected outcome that the coupling of forecasting bias and battery efficiency can lead to higher battery capacity requirements for higher battery round trip efficiencies.