With the increasing of electric vehicle (EV) adoption in recent years, the impact of EV charging activities to the power grid becomes more and more significant. In this article, an optimal scheduling algorithm which combines smart EV charging and V2G gird service is developed to integrate EVs into power grid as distributed energy resources, with improved system cost performance. Specifically, an optimization problem is formulated and solved at each EV charging station according to control signal from aggregated control center and user charging behavior prediction by mean estimation and linear regression. The control center collects distributed optimization results and updates the control signal, periodically. The iteration continues until it converges to optimal scheduling. Experimental result shows this algorithm helps fill the valley and shave the peak in electric load profiles within a microgrid, while the energy demand of individual driver can be satisfied.
Global warming concerns make Electric Vehicle (EV) and Plug-in Hybrid Electric Vehicle (PHEV) more and more popular in recent years. It is estimated that by 2020, EVs will comprise about 4 percent of the 100 million light-duty vehicles [1]. Four million EVs in the US will produce 40 GW of potentially available power if each EV is grid-integrated with an average of 10 kW of available capacity. Currently there are lack of policies to regulate EV charging. All charging activities are spontaneous without coordination and considerations of other grid circumstances. Un-coordinated EV charging behaviors will impact the power grid and degrade power quality when the number of charging sessions in local distribution network reaches a certain level [2], resulting in voltage fluctuation, power loss and blackout, etc. Smart charging can solve these issues by optimizing the EV charging schedule and energy allocation [3]- [5]. Significant studies have been conducted to incorporate bi-directional power flow and grid services support from EVs, upgrading traditional dumb consumptions to controllable distributed energy resources (DERs) [6]. If one can combine EV charging and V2G service into a smart bi-directional charging strategy, EVs will become valuable assets to the power grid. However, distributed EV batteries must be aggregated in a large scale so that they can participate in the wholesale market to provide grid services. Such aggregation would require huge computational resources if it's performed by traditional centralized control scheme. Smart bi-directional charging scheduling depends highly on EV driver behaviors. To facilitate the integration of EVs as DERs into power grid, the following challenges should be addressed: 1) A bi-directional charging algorithm incorporates both smart charging and V2G service; 2) A decentralize control strategy which lowers computational cost with large-scale integration; 3) EV user behavior modeling and prediction.
Previous research has been done to patricianly solve EV charging coordination problems. But none of them provides a comprehensive solution which addresses all the three aforementioned challenges based on real-world implementation. Gan et al. [5] use optimal control to optimally allocate EV charging time and energy, but the algorithm requires users to provide charging schedule, which may degrade customers’ satisfaction by requiring inputs frequently. Wang et al. [7] provide two algorithms to aggregate EV charging with consideration of random user behavior model and renewable generation in the scheduling. Authors in [8] have developed smart charging strategy according to time-ofuse (TOU) price from day-ahead predictions. These smart charging strategies can flatten the grid load profile by valleyfilling but cannot do much during peak load. V2G can be performed to relieve the stress on the grid [9]- [11]. Methods for increasing smart charging and V2G scalability and interoperability are developed in [10], but large scale implementation strategy based on real-world data is rarely investigated.
In this paper we proposed a smart bi-directional charging scheduling algorithm that incorporates EV smart charging and V2G service for large scale implementation. An aggregated control center sends control signal to all EV charging stations in its network. Each distributed charging station performs an optimization locally according to this control signal to update This work has been sponsored in part by grants from the California Energy Commission (CEC) entitled “Demonstration of PEV Smart Charging and Storage Supporting Grid Operational Needs”. Sponsor Award number: EPC-14-056 EV bi-directional charging strategy and reports it to control center. Subsequently, control center collects strategies from networked charging stations and then updates its original control signal. The iteration ends when algorithm converges to an optimal charging strategy for all plug-in EVs in the network. Computational cost is shared by all charging stations which greatly lowers computing burden and time in control center. In each iteration, the information update is asynchronous, there is no need for all charging stations update their strategy. Also there is no strict rule for control center to update control signal every iteration. These features are critical for a successful large scale implementation where prompt updating is impractical. User charging behavior and preference prediction is performed by mean estimation and linear regression using real-world EV usage data. Predicted EV user charging schedule and energy demand serve as input parameters for the proposed algorithm. The contributions of this paper can be summarized as: 1) A distributed optimal bidirectional charging control algorithm implementation with real-world data; 2) Integrate EVs as DERs into power grid to flatten load curves, correct voltage deviation, conduct voltage regulation, etc.; 3) A driver behavioral model based on mean estimatio
This content is AI-processed based on open access ArXiv data.
