Hybrid optimal management of active and reactive power flow in a smart microgrid with photovoltaic generation

Abstract

This paper presents a new approach to the optimal power flow management for low-voltage urban microgrid (UMG) connected to the power grid (PG). The considered UMG consists of a photovoltaic generator, an electrochemical storage system, a micro-gas turbine (GT) and a residential loads. A new optimal energy management system (EMS) is designed to perform a day-ahead power flow management (off-line management) in the UMG. The proposed EMS aims at coordinating the active and the reactive power schedule for the distributed energy sources (DES). This EMS is based on 24 h ahead forecast data of the residential loads consumption, the photovoltaic power and the electricity tariffs. Indeed, the EMS predicts hour-by-hour the active/reactive power references for each DES, while ensuring the stability and the reliability of the UMG. Accordingly, the EMS has two objectives: (1) minimizing the production cost of the active/reactive power for each DES (economic criteria); (2) reducing the CO2 equivalent emissions by adjusting the GT operating point (environmental criteria). To meet these objectives, the optimization problem is stated as a discretized economic problem over a 24 h time horizon. This problem is formulated by a multi-objective function (expressed in cash flow) as well as the operating constraints of the DES. To solve the optimization problem, a hybrid optimization technique is adopted based on the dynamic programming (using the Bellman algorithm) combined with the linear programming (using the simplex algorithm). This resolution choice allows an exhaustive and effective exploration of the solutions search area in a relatively fast resolution time (CPU time). The effectiveness of the proposed management strategy is demonstrated using a comparison against a rules-based management. Indeed, a statistical comparison shows that the proposed management provides an optimization around 38% of the energy bill with respect to the rules-based one.