Increasing the efficiency of a virtual battery storage providing frequency containment reserve power by applying a clustering algorithm

We consider a virtual battery energy storage consisting of individual storage systems used for both, maximization of household PV-power self-consumption and provision of frequency containment reserve power. A central controller manages the frequency containment reserve power provision by setting benchmarks for each individual storage every 15 minutes. The individual storage systems provide frequency containment reserve power autonomously based on the frequency deviation to its nominal value and the given benchmark. A uniform frequency containment reserve power dispersion to all individual storage systems causes marginal power requests most of the time, which lead to high losses due to the bad efficiency of the rectifiers in low part-load operation. Therefore, this paper suggests an algorithm for only using subsets of the virtual battery energy storage with a certain amount of individual systems for frequency containment reserve power provision (“clustering”). The algorithm aims at improving the efficiency and is able to reduce both the charging and the discharging losses by more than 10 percentage points. In addition it balances the states of charge of the individual storages and increases the self-consumption of PV-power of the households. These results are derived from detailed simulations of the system.