東華大學圖書館 |

Modeling and Control for Grid-interactive Efficient Data Centers.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Modeling and Control for Grid-interactive Efficient Data Centers./
作者:	Fu, Yangyang.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2020,
面頁冊數:	242 p.
附註:	Source: Dissertations Abstracts International, Volume: 81-11, Section: B.
Contained By:	Dissertations Abstracts International81-11B.
標題:	Architectural engineering. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=27742568
ISBN:	9798645452162

Modeling and Control for Grid-interactive Efficient Data Centers.
Fu, Yangyang.

Modeling and Control for Grid-interactive Efficient Data Centers. - Ann Arbor : ProQuest Dissertations & Theses, 2020 - 242 p.

Source: Dissertations Abstracts International, Volume: 81-11, Section: B.

Thesis (Ph.D.)--University of Colorado at Boulder, 2020.

This item must not be sold to any third party vendors.

The electrical grid is facing multiple challenges: increasing peak electricity demand, high penetration of variable renewable electricity generation, and transmission and distribution (T&D) infrastructure constraints. These challenges stress the electrical grid by making it more and more difficult to balance supply and demand for different time scales under the constraints of current infrastructure. Demand-side entities such as data centers with flexible electrical loads can also be utilized to serve the balancing, and their contributions can be as viable as supply-side counterparts.Data centers have numerous opportunities to provide grid services considering their large capacities, flexible working environments and work loads, redundant design and operation, etc. Grid-interactive efficient data centers (GEDCs) have rich demand side resources, which have multi-scaled response time from milliseconds up to hours. The fast response resources such as servers can be used for providing frequency regulation (FR), as one of the ancillary services. Some electric markets in U.S., e.g., Pennsylvania New Jersey Maryland (PJM) Interconnection, allows demand sides to provide regulation services.However, due to the fact that either the programs or the markets are not mature especially from the data center's point of view, few of them are willing to participate. There are plenty of reasons, on top of which is the lack of a dynamic modeling and evaluation tool to assist the design and the control of GEDCs.An equation-based object-oriented testbed for GEDCs are proposed and built in the Modelica environment. The proposed testbed considers different physical systems (thermal, electrical, and electromagnetic, etc.) with different time-scaled dynamics involved. End-to-end models include the computer servers, quality of service (QoS), uninterruptible power system, renewable energy resources such as solar panels, typical cooling system for data centers etc. The case studies show that this testbed can be used to perform various analysis, including detailed analysis of energy efficiency and control performance in normal operation, as well as emergency operation.The proposed testbed is then validated using measurement data from an actual data center located in Massachusetts. Based on the validated testbed, energy efficiency measures such as cooling system retrofitting, and control improvement are then proposed and evaluated. The resulted energy savings can be as much as 24.2% for the cooling system alone while data center room is still maintained in an acceptable thermal environment in terms of temperature and relative humidity.To maximally harness the benefits from participating in energy market and regulation market, a real time optimization framework is proposed for GEDC without thermal energy storage systems (TESSs). This optimization framework first proposes a synergistic control strategy to enable FR service in GEDCs. The synergistic control strategy combines power management techniques at the server level with control of the chilled water supply temperature to track the regulation signal from the electrical market. A FR flexibility factor is also proposed to increase the IT capacity for FR. Then the performance of the control strategy is studied through numerical simulations using the proposed testbed. Simulation results show that with well-tuned control parameters, GEDCs can provide FR service in both regulation up and down. The performance of data centers in providing FR service is largely influenced by the regulation capacity bid, FR flexibility factor, workload condition, and cooling mode of the cooling system, and not significantly influenced by the time constant of chillers. The proposed synergistic control strategy can also provide an extra regulation capacity of 3% of the design power when chillers are activated, compared with a server-only control strategy.Finally, the proposed real-time multi-market optimization framework is investigated on the dynamic testbed with well-tuned parameters for FR service. Optimization results show that providing FR service over two days in January and July can save $23.6 and $115.8, respectively.However, without storage systems, GEDCs are difficult to limit their demands that contribute to a large portion of the utility bill. To enable the demand limiting, GEDCs with TESSs are investigated. First, a synergistic control strategy for FR service by adjusting the chiller capacity, storage charging rate and IT server CPU frequency is developed. Then, a model predictive control framework is proposed to minimize the operational costs of GEDCs with TESSs from participating in energy market and regulation market. Simulation results show that utilizing the TESS can not only reduce energy costs and demand charges but also harness FR revenues. The proposed multi-market optimization framework for the data center with TESS in two days can save the operational costs up to 8.8% ($1606.4) compared to the baseline data center with TESS. The savings are consisted of 0.2% ($38.7) of energy cost reduction, 6.5% ($1179.4) from demand reduction and 2.1% ($338.3) from regulation revenues. What's more, the proposed multi-market optimization framework for the data center with TESS can reduce the operation costs by $1793.2 in two days, saving 9.7% compared with a baseline data center without TESS.

ISBN: 9798645452162Subjects--Topical Terms:

3174102
Architectural engineering.
Subjects--Index Terms:

Demand response

Modeling and Control for Grid-interactive Efficient Data Centers.
LDR:06602nmm a2200361 4500 001 2271290
005 20201007134651.5
008 220629s2020 ||||||||||||||||| ||eng d
020 $a 9798645452162
035 $a (MiAaPQ)AAI27742568
035 $a AAI27742568
040 $a MiAaPQ $c MiAaPQ
100 1 $a Fu, Yangyang. $3 3548701
245 1 0 $a Modeling and Control for Grid-interactive Efficient Data Centers.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2020
300 $a 242 p.
500 $a Source: Dissertations Abstracts International, Volume: 81-11, Section: B.
500 $a Advisor: Zuo, Wangda.
502 $a Thesis (Ph.D.)--University of Colorado at Boulder, 2020.
506 $a This item must not be sold to any third party vendors.
520 $a The electrical grid is facing multiple challenges: increasing peak electricity demand, high penetration of variable renewable electricity generation, and transmission and distribution (T&D) infrastructure constraints. These challenges stress the electrical grid by making it more and more difficult to balance supply and demand for different time scales under the constraints of current infrastructure. Demand-side entities such as data centers with flexible electrical loads can also be utilized to serve the balancing, and their contributions can be as viable as supply-side counterparts.Data centers have numerous opportunities to provide grid services considering their large capacities, flexible working environments and work loads, redundant design and operation, etc. Grid-interactive efficient data centers (GEDCs) have rich demand side resources, which have multi-scaled response time from milliseconds up to hours. The fast response resources such as servers can be used for providing frequency regulation (FR), as one of the ancillary services. Some electric markets in U.S., e.g., Pennsylvania New Jersey Maryland (PJM) Interconnection, allows demand sides to provide regulation services.However, due to the fact that either the programs or the markets are not mature especially from the data center's point of view, few of them are willing to participate. There are plenty of reasons, on top of which is the lack of a dynamic modeling and evaluation tool to assist the design and the control of GEDCs.An equation-based object-oriented testbed for GEDCs are proposed and built in the Modelica environment. The proposed testbed considers different physical systems (thermal, electrical, and electromagnetic, etc.) with different time-scaled dynamics involved. End-to-end models include the computer servers, quality of service (QoS), uninterruptible power system, renewable energy resources such as solar panels, typical cooling system for data centers etc. The case studies show that this testbed can be used to perform various analysis, including detailed analysis of energy efficiency and control performance in normal operation, as well as emergency operation.The proposed testbed is then validated using measurement data from an actual data center located in Massachusetts. Based on the validated testbed, energy efficiency measures such as cooling system retrofitting, and control improvement are then proposed and evaluated. The resulted energy savings can be as much as 24.2% for the cooling system alone while data center room is still maintained in an acceptable thermal environment in terms of temperature and relative humidity.To maximally harness the benefits from participating in energy market and regulation market, a real time optimization framework is proposed for GEDC without thermal energy storage systems (TESSs). This optimization framework first proposes a synergistic control strategy to enable FR service in GEDCs. The synergistic control strategy combines power management techniques at the server level with control of the chilled water supply temperature to track the regulation signal from the electrical market. A FR flexibility factor is also proposed to increase the IT capacity for FR. Then the performance of the control strategy is studied through numerical simulations using the proposed testbed. Simulation results show that with well-tuned control parameters, GEDCs can provide FR service in both regulation up and down. The performance of data centers in providing FR service is largely influenced by the regulation capacity bid, FR flexibility factor, workload condition, and cooling mode of the cooling system, and not significantly influenced by the time constant of chillers. The proposed synergistic control strategy can also provide an extra regulation capacity of 3% of the design power when chillers are activated, compared with a server-only control strategy.Finally, the proposed real-time multi-market optimization framework is investigated on the dynamic testbed with well-tuned parameters for FR service. Optimization results show that providing FR service over two days in January and July can save $23.6 and $115.8, respectively.However, without storage systems, GEDCs are difficult to limit their demands that contribute to a large portion of the utility bill. To enable the demand limiting, GEDCs with TESSs are investigated. First, a synergistic control strategy for FR service by adjusting the chiller capacity, storage charging rate and IT server CPU frequency is developed. Then, a model predictive control framework is proposed to minimize the operational costs of GEDCs with TESSs from participating in energy market and regulation market. Simulation results show that utilizing the TESS can not only reduce energy costs and demand charges but also harness FR revenues. The proposed multi-market optimization framework for the data center with TESS in two days can save the operational costs up to 8.8% ($1606.4) compared to the baseline data center with TESS. The savings are consisted of 0.2% ($38.7) of energy cost reduction, 6.5% ($1179.4) from demand reduction and 2.1% ($338.3) from regulation revenues. What's more, the proposed multi-market optimization framework for the data center with TESS can reduce the operation costs by $1793.2 in two days, saving 9.7% compared with a baseline data center without TESS.
590 $a School code: 0051.
650 4 $a Architectural engineering. $3 3174102
653 $a Demand response
653 $a Equation-based modeling and control
653 $a Frequency regulation
653 $a Grid-interactive efficient Buildings
653 $a Model predictive control
653 $a Multi-market optimization
690 $a 0462
710 2 $a University of Colorado at Boulder. $b Civil, Environmental, and Architectural Engineering. $3 3350108
773 0 $t Dissertations Abstracts International $g 81-11B.
790 $a 0051
791 $a Ph.D.
792 $a 2020
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=27742568