Robust linear architecture for active/reactive power scheduling of EV integrated smart distribution networks

• Published in: Electric power systems research Vol. 155; pp. 8 - 20
• Main Authors: Pirouzi, Sasan, Aghaei, Jamshid, Vahidinasab, Vahid, Niknam, Taher, Khodaei, Amin
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.02.2018; Elsevier Science Ltd
• Subjects: Active and reactive power management; Active-reactive power; Architecture; Benders decomposition; Distribution management; Electric vehicles (EVs); Flow equations; Linear programming; Mathematical models; Networks; Nonlinear programming; Operating costs; Power consumption; Power flow; Power management; Robust optimization; Robustness; Schedules; Studies; Uncertainty; Active and reactive power management; Linear programming; Robust optimization; Electric vehicles (EVs); Benders decomposition
• ISSN: 0378-7796; 1873-2046
• DOI: 10.1016/j.epsr.2017.09.021

[Display omitted] •Distribution power management using electric vehicles is proposed.•Robust optimization model has been adopted.•The optimization problem has been linearized.•Uncertain behavior of loads, prices and EVs are modeled. This paper develops a robust bundled active and reactive power management of EV integrated smart distribution networks. To model the problem, at first, the deterministic formulation of the problem is expressed as a non-linear programing (NLP), which minimizes the difference between the energy cost and the revenue of EVs’ (parking lot’s) reactive power exchange with the network as the objective function, subject to the AC power flow equations, system operation limits and EVs’ characteristics as the problem constraints. Then, while the NLP optimization reveals local optima, the NLP model is converted into a linear programming (LP) model using linearized AC power flow equations. The system uncertainties including active and reactive loads, electrical energy and reactive power prices as well as EVs’ charging/discharging schedules are modeled in the proposed linear model. Accordingly, the robust model is implemented and it considers one scenario, namely the most-conservative scenario of the objective function in the main problem. To decrease the calculation time, Benders decomposition (BD) approach is used to speed up the total processing time. The proposed robust linear architecture is tested on three distribution test networks to demonstrate its efficiency and performance. The results show that the NLP model can be substituted with the high-speed LP model. Moreover, the computation speed is improved by using the BD method. In addition, the capacity of the injected power of EVs is reduced in the most-conservative scenario in comparison with the deterministic model’s scenario, while the consumed power of loads and EVs have been increased in this scenario. The proposed robust architecture against uncertainties is shown to yield a more robust solutions at the expense of higher operation cost.