Coordinated reactive power control of hydro generators in the Nordic grid

Hydrogenerators are well known for their voltage control capacity to maintain the Nordic Grid's dynamic security. New methods for coordination between the hydropower fleet and the grid is an unavoidable energy-transitioning strategy. This strategy challenges the current role of the old-fashioned hydrogenerators and demands enhanced coordination beyond the currently implemented technology. This paper proposes a new method to coordinate such multi-machine systems to deliver the required reactive power needed by the grid operator with reduced power losses in the system. Reinforcement Learning (RL) is investigated as a tool for enhancing a reactive power control scheme for multi-machine reactive power support. An in-depth simulation study of a 132 kV radial system from the Norwegian power system is presented, and the overall efficiency improvements are demonstrated on a real-world data set from Nordpool.