Algorithm Design and Regulation Strategies for EV-charging Scheduling

電動汽車充電調度的算法設計及管理策略

Student thesis: Doctoral Thesis

View graph of relations

Author(s)

Related Research Unit(s)

Detail(s)

Awarding Institution
Supervisors/Advisors
Award date14 Sep 2017

Abstract

Electric vehicle (EV) is now a traveling means attracting plenty of users, investors and countries. Apart from the well-known features of environmental friendliness and high energy efficiency, the widespread deployment of EVs, nevertheless, will render an increasing energy demand from power grids. This can inevitably impose some undesirable challenges such as peak load amplification, voltage drops, stability issues, etc. In recent years, the V2G (Vehicle-to-Grid) technology has also become an important concern in power systems. To facilitate the penetration of EVs into the current power networks, a feasible and economical method is to employ smart scheduling strategies for EV charging, through which the goal of intelligent load regulation as well as economic benefits can both be reached. Meanwhile, owing to the operational characteristics differences between EV and power system, efficient and robust charging algorithms are essential to achieve such smart EV charging scheduling. Therefore, in this research, studies involving the algorithm and regulation strategies for scheduling of EV charging have been conducted.

The thesis covers three main topics. The first topic addresses on the algorithm design for EV charging. To begin with, the characteristics of EV charging as well as the constraints from both EV batteries and power systems have been investigated. Considering the V2G scenario and different charging power rates, various charging and discharging modes can occur in power system. To support an advanced and economic EV charging process under diverse EV charging modes, a comprehensive Intelligent Scatter Search (ISS) algorithm evolved from the framework of basic scatter search has been designed. The Filter-SQP (sequential quadratic programming) and mixed-integer SQP techniques are employed as local solvers and the computation time is significantly reduced. Moreover, a hybrid method comprised with ISS and GA (Genetic Algorithm) approach has been further developed to handle large scale EV charging.

The second topic aims at assisting the operation of power distribution networks with the penetration of EVs. The massive EV charging can induce problems including voltage drops and peak amplification. Besides, the charging pattern of EVs also poses challenges to power system restoration under sudden line failures. Therefore, an EV schedule-control based strategy has been designed to address these issues in order to achieve voltage regulation and load shifting under both normal operation and line failure scenario. To be specific, a framework consists of two agents has been developed, wherein a two-stage voltage control agent schedules EVs to perform load shifting and voltage regulation under normal condition, and a fault control agent deals with line failure scenario to recover power supply for out-of-service basic and EV loads. A three-level queue table mechanism has been designed to collaboratively perform EV scheduling. The influence of EV charging locations on the voltage variations of other nodes is considered and alleviated through a voltage sensitivity analysis method. Moreover, graph theory is employed to complete the network reconfiguration process to deal line failure situation. The effectiveness of the scheme has been verified with the simulation results based on a modified IEEE 30 nodes test feeder.

The third topic focuses on developing a pricing policy that can guide the participation of EVs to perform demand response. As an effective demand response approach, a reasonable and intelligent pricing strategy is desirable to motivate EV users to naturally regulate their charging behavior and thereby benefit the power system operation. Among the existing pricing method reported in literature, however, the discharging price is generally set in accordance with the spot electricity price. Nevertheless, when modulating the price tariffs, both the charging and discharging sides must be cohesively equilibrated in order to ensure operator’s benefit and guarantee the enthusiasm of V2G service. As such, a pricing policy for EV discharging service which fairly associates the V2G tariff with system load condition, maximum power limit, and price rate for user load, has been developed. The price setting follows a hierarchical optimization procedure between the operator and end users where the former attempts to maximize its profit whilst the latter aims to balance respective energy cost and comfort. The optimization for the operator is achieved via GA whereas the energy management of each user is decomposed as a single power optimization problem. Simulation results have proven the benefit to liberate the discharging price from the user load tariff, and verified the effectiveness of the proposed strategy.