Flexible region aggregation of adjustable loads via an adaptive convex hull strategy

Increasing interest has been directed toward the potential of heterogeneous flexible loads to mitigate the challenges associated with the increasing variability and uncertainty of renewable generation. Evaluating the aggregated flexible region of load clusters managed by load aggregators is the crucial basis of power system scheduling for the system operator. This is because the aggregation result affects the quality of the scheduling schemes. A stringent computation based on the Minkowski sum is NP-hard, whereas existing approximation methods that use a special type of polytope exhibit limited adaptability when aggregating heterogeneous loads. This study proposes a stringent internal approximation method based on the convex hull of multiple layers of maximum volume boxes and embeds it into a day-ahead scheduling optimization model. The numerical results indicate that the aggregation accuracy can be improved compared with methods based on one type of special polytope, including boxes, zonotopes, and homothets. Hence, the reliability and economy of the power system scheduling can be enhanced.