Dispatch Deviation Reduction-oriented Virtual Power Plant Aggregation

Qiyuan Li

doi:10.1109/tsg.2026.3742828

Dispatch Deviation Reduction-oriented Virtual Power Plant Aggregation

2026-10-13

0

PRE

AI

DOI：10.1109/tsg.2026.3742828

Original

Save

Abstract

En 中文

Aggregating massive DERs as a VPP has been encouraged for their participation in transmission-level operations. This necessitates characterizing the feasible set of VPP power transfers. The major studies focus on better geometrical approximations of the feasible set. However, a better geometric accuracy may not indicate a smaller dispatch deviation. Consequently, this paper innovatively develops a dispatch deviation reduction-oriented VPP aggregation method which contains two interactive evaluation and update modules. The evaluation module formulates a mixed-integer linear programming problem upon the centralized transmission-distribution coordination problem to quantify the worst-case dispatch deviation due to VPP aggregation. Furthermore, the update module employs Danskin’s theorem and perturbation-based sensitivity analysis to compute the gradient which updates the approximate set parameters towards a reduced deviation. To ensure our computational efficiency in daily operations under different load scenarios, a deep neural network-based strategy is incorporated into the two modules so that the module computations can be displaced with the end-to-end mapping from load scenarios to the approximate set parameters which are enforced for dispatch deviation reduction. The proposed method is verified by numerical experiments in one 9-bus and the IEEE 136-bus test systems.

Keywords:

Virtual power plants

distributed energy resources

dispatch deviation

aggregation

deep neural networks

Dispatch Deviation Reduction-oriented Virtual Power Plant Aggregation

Abstract

Journal

Organization

Cited Papers