Optimising grid-forming inverters to prevent under-frequency load shedding with minimal energy storage

The increasing penetration of distributed photovoltaics decreases the available loads in distribution feeders. This makes the traditional under-frequency load shedding systems ineffective. As an alternative, large-scale energy storage systems (ESS) could be used to provide contingency frequency response. This study investigates the optimum sizing of the ESS to prevent under-frequency load shedding. The optimal size is determined for both droop and virtual synchronous generator control-based inverters, to maintain frequency within operating standards. The control parameters are treated as variables in the optimisation process, allowing the identification of the optimal set of control parameters that minimise the ESS size while preventing under - frequency load shedding. The power system is implemented in DIgSILENT PowerFactory, and the Hill climbing algorithm is used for optimisation. Given that a black box optimisation methodology is used, it is compatible with non-linear governor models and encrypted inverter-based models. Finally, a detailed analysis of the variations in the minimum active power rating for different control parameters is provided. A minimum energy storage power rating of 85 MVA is achieved when using virtual synchronous generator control. For droop control, the minimum storage capacity is 89 MVA. This analysis provides a deeper understanding of how different control parameters influence the required size of the energy storage system.