Self-scheduling of electric vehicles in an intelligent parking lot using stochastic optimization

Abstract

Electric vehicles charging and discharging management as well as large scale intermittent renewable power generation management are known as the two most important challenges in the future distribution system operation and control. Proper integration of these energy sources may introduce a solution for overcoming to challenges. In this paper, a stochastic charging and discharging scheduling method is proposed for large number of electric vehicles parked in an intelligent parking lot where intelligent parking lots are potentially introduced as aggregators allowing electric vehicles interact with the utilities. A self-scheduling model for an intelligent parking lot equipped with photovoltaic system and distributed generators is presented in this paper in which practical constraints, solar radiation uncertainty, spinning reserve requirements and electric vehicles owner satisfaction are considered. The results show that the proposed parking lot energy management system satisfies both financial and technical goals. Moreover, electric vehicle owners could earn profit by discharging their vehicles as well as having desired state of charge in the departure time.

Introduction

EVs are an important component of an electric power network in the near future. The widespread adoption of EVs may introduce a solution to the world fossil fuel shortage as well as the air pollution crisis [1]. The emission reduction aim is achieved by proper and optimum utilization of the EVs as energy storages and loads in the power system integrated with RESs [2], [3], [4]. Beyond these advantages, connection of EVs into the power network may bring up some technical drawbacks that need to be addressed properly. With the widespread adoption of EVs, the power system may face significant challenges due to the huge electricity demand of these loads [5], [6]. For example, if 30% of conventional vehicles in the US were replaced by EVs, the total charging load would be 140 GW, which accounts for 18% of the US summer peak load of 780 GW [7].

EVs utilize battery as an energy storage system in order to provide a power supply for their electric-drive motors. When EVs are plugged into a power outlet, can operate in two modes: charging or G2V mode, and discharging or V2G mode. In the former, the EV is regarded as a load to the utility, while in the latter EV could supply energy to the grid by discharging the stored energy in its battery. Therefore, EV in the view of the utility grid is considered as a probable load or generation unit [8]. With V2G capability, the state of charge of an EV׳s battery can go up or down, depending on the revenues and grid׳s demands. Through V2G, EV owners can make revenue while their cars are parked; it can provide valuable economic incentives for EV owners. On the other hand, utilities significantly benefit from V2G due to increase in system flexibility and reliability as well as using energy storage for intermittent RESs such as wind and solar.

An EV is designed for transportation, so the main duty of battery storage in the EV is to provide sufficient power for the vehicle to drive. In order to maximize customer satisfaction and minimize disturbances of the grid, EVs parking lots are a good solution for handling the EVs energy management challenges. Parking lots will be appropriate places for implementing the V2G strategy as EVs are parked several hours per day in them [9], [10]. A vehicle may spend 23 h each day parked [11] and also 90% of vehicles are parked even during peak traffic hours [12].

In [13], an estimation of distribution algorithm to schedule large number of EVs charging in a parking lot has been proposed. The method optimizes the energy allocation to the EVs in the real-time while considering various constraints associated with EV battery and utility limits. The paper has only proposed the charging method of EVs and the V2G option was not taken into account. The authors in [14] proposed a simulated annealing approach and heuristic technical validation of the obtained solutions to solve the energy resources scheduling. A case study considering 1000 EVs connected to a distribution network managed by a virtual power plant has been presented. The EVs scheduling schemes proposed in [15], [16] only dealt with the battery charging without considering V2G capability. The V2G scheduling models proposed in [17], [18] tried to optimize the charging and discharging powers to minimize the cost. In charging and discharging scheduling, the scheduler tries to optimize the bidirectional energy flows between the grid and EV׳s Battery. In [19], an optimization problem of scheduling EV charging with energy storage for the day-ahead and real-time markets has been proposed. Also, a communication protocol for interactions among different entities including the aggregator, the power grid, the energy storage, and EVs was considered. Some recent literatures [20], [21], [22], [23] discussed about the charging points equipped with PV panels. The solar power can be considered as a valuable energy source for charging EVs. Parking lots equipped with PV panels can provide cheap and green energy for EVs and in this way, reduce emission from transportation sector. On the other hand, the internal control system of EVs has also attracted increasing research efforts because of its considerable advantages in terms of vehicle motion control, energy optimization, and vehicle structural arrangement. More details on the control issue of EVs has been discussed in [24], [25], [26].

In the EVs management model, different types of objective functions have been presented in the literatures. For example, the objective could be to minimize the cost and air pollutant emission for a sustainable integrated electricity and transportation infrastructure by maximum utilization of RESs using EVs [27]. If the aggregated EV batteries are considered as a potential energy storage system, another objective could be taken into account to maximize the capability of the aggregated batteries in order to mitigate the unpredictable fluctuations of renewable energy [28]. A novel objective function maximized the average SOC for all vehicles at the next time step [13].

In this paper, an IPL with PV system on the roof, DGs and a bidirectional utility grid connection is presented for stochastic charging and discharging scheduling of 500 EVs. The grid connection is considered to satisfy any charging demand greater than the PV and DGs output and to supply energy to the grid during peak hours. An energy management system for the PV based parking lot is proposed here in which the PV generation uncertainty and V2G capability of EVs are considered. Moreover, the proposed model considers system constraints and customer׳s preferences. The contributions of the proposed method are highlighted as follows:

• Include and aggregate intermittent PV generation with EVs charging and discharging scheduling.

• Evaluate EVs role in providing reserve as well as energy.

• Consider the EVs owners preferences in EVs energy management program.

The rest of the paper is organized as follows: in Section 2, the proposed system components are introduced. Section 3 presents the problem formulation; including the resources and EVs constraints. A case study and analysis of the results are shown in Section 4. Finally, concluding remarks are presented in Section 5.