Adaptive model predictive scheduling of flexible interconnected low-voltage distribution networks considering charging preferences of electric vehicles

Abstract

DC interconnection at the second side of distribution transformers helps achieve power sharing among nearby low-voltage distribution networks (LVDNs) and promote integration of intermittent inverter-based distributed generators (DGs). This paper proposes an adaptive model predictive scheduling method for flexible interconnected LVDNs considering charging preferences of electric vehicles (EVs). Firstly, the steady-state models of flexible resources including voltage source converters, energy storage systems along with AC and DC power flow models are established. Then, a model predictive control (MPC)-based rolling optimization model is formulated aiming to minimize the daily energy loss considering uncertainties of DGs, load and each charging station as a whole. To further explore the flexibility and dispatchability of each charging station, an adaptive MPC-based rolling optimization model is built considering three types of EVs with different charging preferences, i.e., uncontrollable EVs, charging-only EVs and vehicle-to-grid EVs. The scheduling window of the adaptive MPC-based scheduling is dynamically updated according to the maximum departure time of currently charging EVs to fulfill expected energy requirements of all EVs. Simulation results on a typical flexible LVDN show that the daily energy loss and total load fluctuation can be further reduced through real-time scheduling of controllable EVs in addition to existing flexible resources.