Intelligent Charging System with Proactive and Reactive Scheduling for Electric Vehicles based on Demand Forecast
전력 사용량 예측 기반 지능형 전기자동차 충전 시스템의 프로액티브 및 리액티브 스케줄링 기법
- 주제(키워드) electric vehicle , electric vehicle charging system , charging scheduling , smart grid , proactive and reactive scheduling
- 발행기관 고려대학교 대학원
- 지도교수 조충호
- 발행년도 2015
- 학위수여년월 2015. 2
- 학위구분 박사
- 학과 일반대학원 컴퓨터정보학과
- 세부전공 데이터통신및네트워크 전공
- 원문페이지 122 p
- 실제URI http://www.dcollection.net/handler/korea/000000056719
- 본문언어 영어
- 제출원본 000045827892
초록/요약
This thesis presents a new architecture of Electric Vehicle (EV) charging systems and two novel EV charging scheduling schemes named Proactive First Come First Serve (P-FCFS) and Proactive Cost Based Scheduling (P-CBS) with foresting the amount of available charging capacity in a residential complex. In the proposed architecture, a Charging Management Server (CMS) manages the EV authentication, communications with home servers, a set of chargers, and so on, while an energy management server (EMS) estimates the amount of available charging capacity for EV charging systems. Thus, the EMS uses a forecast algorithm that includes data preprocessing, correlation analysis, i.e. Pearson’s correlation coefficient, and forecast modeling, i.e., moving average (MA), autoregressive (AR) and simple linear regression models, for the amount of available charging capacity. For the realistic environment, we use real electric use data collected by an apartment in Dong-tan and temperature data from the Korea Meteorological Administration. Finally, we obtain the charging capacity for the EV charging systems using the hourly charging capacity from the forecast results and amount of electricity used in a target apartment. Based on the estimated available charging capacity for EV charging systems from the EMS, the CMS schedules the EVs using the proposed P-FCFS and P-CBS schemes that have priorities for arrival time and low charging cost, respectively. The proposed P-FCFS and P-CBS schemes consist of two stages, the proactive and the reactive scheduling stages. That is, in the first stage, i.e., the proactive scheduling stage, when EVs arrive and request charging, the CMS uses a proactive scheduling algorithm to generate an initial and feasible schedule with consideration of the estimated available charging capacity. In the second stage, i.e., the reactive scheduling stage, the CMS periodically checks the real-time available charging capacity to compare it with the estimated available charging capacity. Then, the CMS adaptively updates the initial schedule using a reactive scheduling algorithm if the real-time available charging capacity is different from the estimated available charging capacity. We evaluate the system’s performance using a Monte Carlo simulation in terms of the mean charging ratio, unsatisfied charging probability, failure probability, dropping probability, mean unit cost, utilization of charging capacity, and so on. From the simulation results, we show that the P-FCFS and P-CBS schemes not only show the full charging time when the EVs arrive but also outperform another scheme named FCFS that has no forecast procedure. It is shown that the maximum number of EVs with 100% of the mean charging ratio is 93 which adopted P-FCFS with Γ=3. Further, the P-CBS scheme reduced the cost 18% and 33.6% when Γ=2 and 3, respectively when the mean charging probability equals 100% for all algorithms. In addition, through the simulation results, we analyzed the number of required chargers according to the number of EVs in apartments. Therefore, this thesis provides guidelines for the design of efficient and reliable EV charging systems that can be used to maximize the energy efficiency with consideration of the amount of available charging capacity.
more목차
I. INTRODUCTION 1
I.1. OBJECTIVES 2
I.2. ORGANIZATION 3
II. RELATED WORK 4
II.1. Trend of the Electric Vehicle 4
II.2. Infrastructure of EV Charging system for the residential complex 6
II.3. EV charging scheduling 8
II.4. Data analysis of battery model, driving pattern 10
III. PROPOSED EV CHARGING SYSTEM 13
III.1. Target System 13
III.2. EV charging procedure 14
III.3. System model 18
III.4. Performance Measure 23
IV. ESTIMATION OF AVAILABLE CHARGING CAPACITY BY DEMAND FORECAST 27
IV.1. Forecast as a step in the process of estimation of the EACC 27
IV.2. Description of data 28
IV.3. Data preprocessing 29
IV.4. Correlation analysis 30
IV.5. Forecast model 39
IV.6. Estimation of EACC 47
V. PROACTIVE AND REACTIVE CHARGING SCHEDULING METHODS 50
V.1. Proposed Electric Vehicle Charging Scheme 50
V.2. Main procedure for overall EV charging 51
V.3. FCFS Charging 53
V.4. P-FCFS Scheduling 57
V.5. P-CBS 66
V.6. Simulation Results 73
VI. CONCLUSIONS 85
APPENDIX 88
REFERENCES 98
