Two-stage approach for optimal dispatch of distributed energy resources in distribution networks considering virtual power plant concept

Abstract

SUMMARY With increasing levels of distributed generation (DG) penetration, especially variable output ones, new challenges have appeared in electric power systems. The outputs of these generators are variable and determined largely by nature of the power sources such as wind, wave, and sunlight. One solution to decrease problems caused by the variable output of these resources is to aggregate them into a virtual power plant (VPP). The VPP is an energy management system tasked to aggregate the capacity of some DGs, storage facilities, and dispatchable loads (DL) for the purpose of energy trading and/or providing system support services. The economically optimal control of distributed energy resources (DERs) included in the VPP is complicated by uncertainties, nonlinearities, and inter-temporal constraints. Therefore, a novel two-stage dispatch framework is presented here for the integration of the stochastic and the controllable DERs in the distribution network through the VPP concept. The first stage deals with the VPP optimal bidding model (VOBM). A probabilistic price-based unit commitment using point estimate method (PEM) is used in the VOBM to allow the VPP to decide on the unit commitment of its DERs and the optimal bids to the day-ahead market. The second stage deals with active distribution network management. The distribution system operator uses input from the VOBM to manage any local network constraints and to determine the characteristics of the entire local network at the grid supply points, while seeking to minimize the hourly operation cost. The proposed optimization algorithm is applied to an 18-bus system. Copyright © 2012 John Wiley & Sons, Ltd.