Avoiding PV-Induced Overvoltage through Grid-Connected Batteries Using Model Predictive Control

The problem of temporary overvoltage when integrating large renewable power plants into the existing grid is considered. A grid-connected battery energy storage system with an advanced predictive control algorithm is proposed to reduce the overvoltage in time periods of excessive renewable production. An approximative grid model is developed using system identification methods on regular operation data. A model predictive control algorithm utilizing predictions of load and generation determines the optimal operation strategy of the battery. A comparison in simulation between a reference case with PID control and the proposed algorithm shows a large reduction of the time spent in overvoltage conditions, especially in the case of consecutive days of high renewables production. The results suggest that energy storages combined with a predictive control strategy can effectively alleviate the overvoltage problem in renewables integration. Especially in the case when available storage capacity is comparatively low, a uniform overvoltage reduction can be realized. The method can potentially increase the PV host capacity of current grids.