A receding horizon control approach for re-dispatching stochastic heterogeneous resources accounting for grid and battery losses

摘要

In this paper, we propose a re-dispatch scheme for radial distribution grids hosting stochastic Distributed Energy Resources (DERs) and controllable batteries. At each re-dispatch round, the proposed scheme computes a new dispatch plan that modifies and extends the existing one. To do so, it uses the CoDistFlow algorithm and applies a receding horizon control principle, while accounting for hard time computation constraints that impact on the instantaneous update of a dispatch plan. CoDistFlow handles stochastic DERs and prosumers uncertainties via scenario-based optimization and the non-convexity of the AC Optimal Power Flow by iteratively solving suitably defined convex problems until convergence. We perform numerical evaluations based on real-data, obtained from a real Swiss grid. We show that, with our proposed re-dispatch scheme, the daily dispatch tracking error can decrease more than 80%, even for small battery capacities, and if re-dispatch is frequent enough, it can be eliminated. Finally, we show that re-dispatch should be performed as often as the market allows and the performance continues to improve.