본문

서브메뉴

A Decentralized Privacy-Preserving Data-Driven Methodology for Energy Trading in the Smart Grid
A Decentralized Privacy-Preserving Data-Driven Methodology for Energy Trading in the Smart...
A Decentralized Privacy-Preserving Data-Driven Methodology for Energy Trading in the Smart Grid

상세정보

자료유형  
 학위논문 서양
최종처리일시  
20260209102906
ISBN  
9798265406552
DDC  
300
저자명  
Kampezidou, Styliani Ioanna.
서명/저자  
A Decentralized Privacy-Preserving Data-Driven Methodology for Energy Trading in the Smart Grid
발행사항  
[Sl] : Georgia Institute of Technology, 2023
발행사항  
Ann Arbor : ProQuest Dissertations & Theses, 2023
형태사항  
320 p
주기사항  
Source: Dissertations Abstracts International, Volume: 87-05, Section: A.
주기사항  
Advisor: Mavris, Dimitri N.;Romberg, Justin.
학위논문주기  
Thesis (Ph.D.)--Georgia Institute of Technology, 2023.
초록/해제  
요약To mitigate the Climate Change crisis, protect public health, reduce harmful pollutants, and deliver climate and public health benefits, the Clean Power Plan policy, followed by the 2030 electricity generation decarbonization and the 2050 net-zero economy goals, all instruct greenhouse emissions reduction. Over-investing in variable energy resources has led to grid reliability, such as the steep ramping in California's Independent System Operator (CAISO) service territory duck-curve-shaped net demand. It has been suggested by the North American Reliability Corporation (NERC) that the distributed load customers, i.e. commercial and residential buildings who are responsible for up to 32% of the total carbon dioxide emissions in the United States, are an ideal candidate not only for energy savings but also for grid reliability and flexibility improvement purposes due to the renewable generation expected to reach 90% of total electricity generation by 2030. Although many power utilities and building energy systems have implemented demand-response and demand-side management programs before, recent electricity market regulation changes have allowed for the distributed load to bid in the demand-response market for potentially higher monetary rewards via an energy broker, the aggregator. To ensure high aggregator rewards, as well as, a high impact on carbon dioxide reduction and grid reliability, former demand-response centralized aggregator solutions will need to be decentralized to support large numbers of customers and also attend to privacy-related concerns raised by the Department of Energy (DoE) and the U.S. National Institute for Standards and Technology (NIST).The proposed methodology for benefits extraction from the distributed load consists of two horizons and introduces additional flexibility which increases the day-ahead market rewards while preventing market operator-imposed penalties from real-time demand deviations. This methodology is scalable with the number of distributed load customers and privacy-preserving and is tested on the two proposed test cases with specific formulations which however can be changed in the future with the overall methodology remaining applicable. In the first horizon, the day-ahead, a scalable flexible market participation scheme is proposed for all the distributed load (prosumers) and the demand-response aggregator, who submit a cumulative demand profile bid to the market. In the second horizon, the real-time one, and under the assumption that a distributed load local energy management system has developed a day-ahead device schedule according to the total day-ahead demand profile, a mechanism for real-time demand corrections is proposed to avoid deviation penalties. Specifically, during the day-ahead realization, unpredictable device demands deviations due to user behavior, market, and weather phenomena, can potentially be corrected by real-time sensor-based suggestions to the local energy management system. This dissertation particularly focuses on HVAC devices that are responsible for about 50% of the total building's demand and proposes a non-personally identifiable sensor-based occupancy-informed HVAC controller input for real-time demand management suggestions in the absence of humans. The proposed methodology is developed with statistical learning and game theoretical concepts and is decentralized and privacy-concerned in the sense that certain behavior and preference-related customer information remains hidden from the aggregator during the day-ahead horizon, while only non-personally identifiable data is collected during the real-time horizon from environmental, non-personally identifiable measurements. The experimental processes are designed based on machine learning workflow structures and methods. Although the frameworks proposed in the two horizons contribute to energy savings and grid flexibility and reliability, the proposed scalable and privacy-preserving methodology that enables the solution and implementation of those frameworks is universal and can be applied to several other problem formulations that facilitate the electricity generation decarbonization and a net-zero economy.
일반주제명  
Load
일반주제명  
Integer programming
일반주제명  
Open source software
일반주제명  
Electricity generation
일반주제명  
Linear programming
일반주제명  
HVAC
일반주제명  
Engineering
일반주제명  
Energy
일반주제명  
Privacy
일반주제명  
Demand side management
일반주제명  
Distance learning
일반주제명  
Educational technology
기타저자  
Georgia Institute of Technology.
기본자료저록  
Dissertations Abstracts International. 87-05A.
전자적 위치 및 접속  
로그인 후 원문을 볼 수 있습니다.

MARC

 008260203s2023        us                              c    eng  d
■001000017365977
■00520260209102906
■006m          o    d                
■007cr#unu||||||||
■020    ▼a9798265406552
■035    ▼a(MiAaPQ)AAI32315745
■035    ▼a(MiAaPQ)GeorgiaTech76734
■040    ▼aMiAaPQ▼cMiAaPQ
■0820  ▼a300
■1001  ▼aKampezidou,  Styliani  Ioanna.
■24512▼aA  Decentralized  Privacy-Preserving  Data-Driven  Methodology  for  Energy  Trading  in  the  Smart  Grid
■260    ▼a[Sl]▼bGeorgia  Institute  of  Technology▼c2023
■260  1▼aAnn  Arbor▼bProQuest  Dissertations  &  Theses▼c2023
■300    ▼a320  p
■500    ▼aSource:  Dissertations  Abstracts  International,  Volume:  87-05,  Section:  A.
■500    ▼aAdvisor:  Mavris,  Dimitri  N.;Romberg,  Justin.
■5021  ▼aThesis  (Ph.D.)--Georgia  Institute  of  Technology,  2023.
■520    ▼aTo  mitigate  the  Climate  Change  crisis,  protect  public  health,  reduce  harmful  pollutants,  and  deliver  climate  and  public  health  benefits,  the  Clean  Power  Plan  policy,  followed  by  the  2030  electricity  generation  decarbonization  and  the  2050  net-zero  economy  goals,  all  instruct  greenhouse  emissions  reduction.  Over-investing  in  variable  energy  resources  has  led  to  grid  reliability,  such  as  the  steep  ramping  in  California's  Independent  System  Operator  (CAISO)  service  territory  duck-curve-shaped  net  demand.  It  has  been  suggested  by  the  North  American  Reliability  Corporation  (NERC)  that  the  distributed  load  customers,  i.e.  commercial  and  residential  buildings  who  are  responsible  for  up  to  32%  of  the  total  carbon  dioxide  emissions  in  the  United  States,  are  an  ideal  candidate  not  only  for  energy  savings  but  also  for  grid  reliability  and  flexibility  improvement  purposes  due  to  the  renewable  generation  expected  to  reach  90%  of  total  electricity  generation  by  2030.  Although  many  power  utilities  and  building  energy  systems  have  implemented  demand-response  and  demand-side  management  programs  before,  recent  electricity  market  regulation  changes  have  allowed  for  the  distributed  load  to  bid  in  the  demand-response  market  for  potentially  higher  monetary  rewards  via  an  energy  broker,  the  aggregator.  To  ensure  high  aggregator  rewards,  as  well  as,  a  high  impact  on  carbon  dioxide  reduction  and  grid  reliability,  former  demand-response  centralized  aggregator  solutions  will  need  to  be  decentralized  to  support  large  numbers  of  customers  and  also  attend  to  privacy-related  concerns  raised  by  the  Department  of  Energy  (DoE)  and  the  U.S.  National  Institute  for  Standards  and  Technology  (NIST).The  proposed  methodology  for  benefits  extraction  from  the  distributed  load  consists  of  two  horizons  and  introduces  additional  flexibility  which  increases  the  day-ahead  market  rewards  while  preventing  market  operator-imposed  penalties  from  real-time  demand  deviations.  This  methodology  is  scalable  with  the  number  of  distributed  load  customers  and  privacy-preserving  and  is  tested  on  the  two  proposed  test  cases  with  specific  formulations  which  however  can  be  changed  in  the  future  with  the  overall  methodology  remaining  applicable.  In  the  first  horizon,  the  day-ahead,  a  scalable  flexible  market  participation  scheme  is  proposed  for  all  the  distributed  load  (prosumers)  and  the  demand-response  aggregator,  who  submit  a  cumulative  demand  profile  bid  to  the  market.  In  the  second  horizon,  the  real-time  one,  and  under  the  assumption  that  a  distributed  load  local  energy  management  system  has  developed  a  day-ahead  device  schedule  according  to  the  total  day-ahead  demand  profile,  a  mechanism  for  real-time  demand  corrections  is  proposed  to  avoid  deviation  penalties.  Specifically,  during  the  day-ahead  realization,  unpredictable  device  demands  deviations  due  to  user  behavior,  market,  and  weather  phenomena,  can  potentially  be  corrected  by  real-time  sensor-based  suggestions  to  the  local  energy  management  system.  This  dissertation  particularly  focuses  on  HVAC  devices  that  are  responsible  for  about  50%  of  the  total  building's  demand  and  proposes  a  non-personally  identifiable  sensor-based  occupancy-informed  HVAC  controller  input  for  real-time  demand  management  suggestions  in  the  absence  of  humans.  The  proposed  methodology  is  developed  with  statistical  learning  and  game  theoretical  concepts  and  is  decentralized  and  privacy-concerned  in  the  sense  that  certain  behavior  and  preference-related  customer  information  remains  hidden  from  the  aggregator  during  the  day-ahead  horizon,  while  only  non-personally  identifiable  data  is  collected  during  the  real-time  horizon  from  environmental,  non-personally  identifiable  measurements.  The  experimental  processes  are  designed  based  on  machine  learning  workflow  structures  and  methods.  Although  the  frameworks  proposed  in  the  two  horizons  contribute  to  energy  savings  and  grid  flexibility  and  reliability,  the  proposed  scalable  and  privacy-preserving  methodology  that  enables  the  solution  and  implementation  of  those  frameworks  is  universal  and  can  be  applied  to  several  other  problem  formulations  that  facilitate  the  electricity  generation  decarbonization  and  a  net-zero  economy.
■590    ▼aSchool  code:  0078.
■650  4▼aLoad
■650  4▼aInteger  programming
■650  4▼aOpen  source  software
■650  4▼aElectricity  generation
■650  4▼aLinear  programming
■650  4▼aHVAC
■650  4▼aEngineering
■650  4▼aEnergy
■650  4▼aPrivacy
■650  4▼aDemand  side  management
■650  4▼aDistance  learning
■650  4▼aEducational  technology
■690    ▼a0791
■690    ▼a0537
■690    ▼a0800
■690    ▼a0710
■690    ▼a0454
■71020▼aGeorgia  Institute  of  Technology.
■7730  ▼tDissertations  Abstracts  International▼g87-05A.
■790    ▼a0078
■791    ▼aPh.D.
■792    ▼a2023
■793    ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17365977▼nKERIS▼z이  자료의  원문은  한국교육학술정보원에서  제공합니다.

미리보기

내보내기

chatGPT토론

Ai 추천 관련 도서


    신착도서 더보기
    최근 3년간 통계입니다.

    소장정보

    • 예약
    • 소재불명신고
    • 나의폴더
    • 우선정리요청
    • 비도서대출신청
    • 야간 도서대출신청
    소장자료
    등록번호 청구기호 소장처 대출가능여부 대출정보
    TF15938 전자도서 대출가능 마이폴더 부재도서신고 비도서대출신청 야간 도서대출신청

    * 대출중인 자료에 한하여 예약이 가능합니다. 예약을 원하시면 예약버튼을 클릭하십시오.

    해당 도서를 다른 이용자가 함께 대출한 도서

    관련 인기도서

    로그인 후 이용 가능합니다.