Efficient Algorithm for Scalable Event-Based Demand Response Management in Microgrids

Demand response (DR) management has become one of the key enabling technologies for smart grids. Motivated by the increasing DR incentives offered by service operators, more customers are subscribing to various DR programs. However, with growing customer participation, the problem of determining the optimal loads to be curtailed in a microgrid (MG) during contingencies within a feasible time frame becomes computationally hard. This paper proposes an efficient approximation algorithm for event-based DR management in MGs. In event-based management, it is important to curtail loads as fast as possible to maintain the stability of an MG during the islanded mode in a scalable manner. A simple greedy approach is presented that can rapidly determine a close-to-optimal load curtailment scheme to maximize the aggregate customer utility in milliseconds for a large number of customers. This paper further derives a novel theoretical guarantee of the gap between the proposed efficient algorithm and the optimal solution (that may be computationally hard to obtain). The performance of algorithm is corroborated extensively by simulations with up to thousands of customers. For the sake of practicality, the proposed event-based DR management algorithm is applied to a feeder from the Canadian benchmark distribution system. The simulation results demonstrate that the proposed approach efficiently optimizes MG operation during islanded mode while maintaining appropriate voltage levels and network constrains.