Energy storage sizing for voltage control in LV networks under uncertainty on PV generation

Energy storage systems may represent a viable solution to tackle over- and undervoltages arising in low voltage networks due to the increasing penetration of low carbon technologies. An algorithm for siting and sizing energy storage systems in radial low voltage networks was proposed in previous work. Siting exploits the voltage sensitivity matrix of the network, while sizing is performed by solving a multi-period optimal power flow problem. In this paper, we discuss the sizing step of the aforementioned algorithm for a low voltage network featuring overvoltages due to the high penetration of photovoltaic generation. Since the considered decision problem is affected by uncertainty on photovoltaic generation, a scenario-based approach, coupled with suitable scenario reduction techniques, is analyzed. While this approach is useful to keep the computational burden affordable, we investigate whether the energy storage system sizes found after scenario reduction can guarantee the solution of overvoltages with a priori defined confidence level. This is done by testing the overall procedure on a real Italian low voltage network provided by the main Italian distribution system operator.