Majlesi Journal of Mechatronic Systems Vol. 3, No. 4, December 2014 7 Probabilistic and Fuzzy Modeling of Charging and Discharging of Electric Vehicles in Presence of Uncertainties Applied to Unit Commitment Problem Saeed Nasiri 1 , Mohsen Parsa Moghadam 2 1- Department of Electrical Engineering,Tarbiat Modares University, Tehran, Iran. Email: Saeed.Nasiri@Modares.ac.ir 2- Department of Electrical Engineering,Tarbiat Modares University, Tehran, Iran Email: Parsa@modares.ac.ir Received: July 2014 Revised: August 2014 Accepted: September 2014 ABSTRACT Advent of electric vehicles in power systems has a variety of effects. The most important effect of electric vehicles in the power system is related to increase of demand due to charging process (G2V) and also power injection to grid (V2G). In this paper unit commitment problem has been solved considering electric vehicles. Since charging and discharging of electric vehicles depends on factors such as electricity price, fuel price, the rate of charging and discharging as well as the overall behavior of car owners, to cover these uncertainties, both probabilistic modeling and fuzzy sets theory are used. In probabilistic modeling of consumption and generation of electric vehicle in each hour, the number of Evs, the rate and start time of charging and discharging process, and the average distance traveled by each Ev is considered. Since fuzzy modeling can be used in cases that historical data is not available, in second part of this paper a fuzzy number has been assigned to determine the power generation or consumption related to Evs in each hour which is need to solve unit commitment problem. A 10-unit IEEE test system is considered for simulation with 50,000 gridable vehicles using a hybrid GA-ICA algorithm and then the results of probabilistic and fuzzy modeling are compared with each other. KEYWORDS: Electric Vehicles, Fuzzy Modeling, Generation Scheduling, Probabilistic Modeling, Unit Commitment. 1. INTRODUCTION According to increasing concerns over environmental issues such as reducing emission and less consumption of fossil resources, in recent years a great tendency has been emerged to increase the penetration of electric vehicles due to environmental as well as technical, economic and social advantages. Plug-in electric vehicles (PEVs) as mobile energy storage systems can improve the reliability, increase load factor, and reduce losses in electric network [1, 2]. Electric vehicles also reduce the need for small generation units during peak hours [3, 4]. Furthermore they reduce generation costs of fossil fuel units and also can participate in the reserve and frequency regulation markets [5, 6]. During recent years, there have been conducted some studies for generation scheduling in the presence of PEVs. In [7] the generation scheduling considering PEVs connected to the grid (V2G) has been studied aiming to minimize the operation and emission costs by using particle swarm optimization (PSO). In [8], generation scheduling has been implemented in the presence of PEVs aiming to minimize operation cost by Genetic and Lagrange Algorithm (GA-LA) in which also the charge of PEVs has been added to the constraints as a load. In this paper probabilistic and fuzzy theory has been used for modeling the uncertainty of electric vehicles by the goal of obtaining more actual model in unit commitment problem in the presence of PEVs aiming to minimize the operation costs of generating units and electric vehicles. The organization of this paper is as follows; the formulation of unit commitment problem in the presence of plug-in electric vehicles (UC-PEV) is given in section 3. After that probabilistic and fuzzy modeling of electric vehicles as a load and generation unit are investigated in section 4 and 5 respectively. Hybrid Genetic-Imperialist Competitive algorithm (GA-ICA) has been introduced in section 6 followed by numerical study in Section 7 and conclusion in Section 8.