An Efficient Reduction Method for Available Time Intervals of Large-Scale Demand-Side Resources

Incorporating demand-side resources (DRs) via virtual power plants into power system dispatch has become increasingly prevalent to increase power balancing flexibility. Nonetheless, diverse available time intervals for DRs impose substantial computational burdens for aggregation and may lead to infeasible dispatch after disaggregation. This work proposes an efficient reduction method for aggregating DRs with diverse time intervals and demand requirements while ensuring feasible dispatch. The method determines how to reduce the time interval ranges of all DRs to alleviate dispatch complexity. We reformulate an equivalent ILP problem with an available time interval matrix to reduce the problem scale. To further alleviate the computational burden, a fast approximation algorithm that combines interval clustering and weight selection is designed for large-scale DR aggregation. Numerical tests show that the proposed method maintains larger aggregated feasible regions to provide more flexibility than other inner-approximation methods, while maintaining feasible dispatch. Overall, the reformulated problem and fast approximation algorithm effectively reduce the computational time with a limited reduction in dispatch optimality.