Graduation Semester and Year

2023

Language

English

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Aerospace Engineering

Department

Mechanical and Aerospace Engineering

First Advisor

Dereje Agonafer

Abstract

ABSTRACT: A data center is a specifically constructed establishment designed to accommodate computing, storage, and transmission equipment, thereby facilitating essential business operations spanning diverse sectors of the economy. This facility serves as the designated physical infrastructure for housing servers, networking components, and data storage systems. In the context of traditional data centers, a substantial portion of energy, specifically around 45% to 55%, is directed towards supporting information technology (IT) operations. Additionally, a considerable allocation, varying from 30% to 40%, is attributed to the facilitation of cooling processes. The scope of this dissertation study comprises nine distinct chapters, each contributing to addressing the challenges to establish energy-efficient and optimized data centers. The research commences by focusing on energy-efficient control strategies for hyperscale data centers through scaled-down Computational Fluid Dynamics (CFD) modeling. This approach involves validating the model against real data center datasets and conducting simulations tailored to specific needs. Leveraging machine learning techniques, this investigation aims to anticipate control setpoints for optimizing airflow and chilled water control at the data center room level. As the research progresses, the study addresses the limitations of CRAH units in sustaining high-density rack setups. Consequently, the feasibility and energy implications of implementing Rear Door Heat Exchangers (RDHx) are explored. This includes modeling a single rack with RDHx, assessing both passive and active modes. Moving forward, the study expands to assess the energy efficiency of RDHx across data center deployment. Furthermore, in response to the challenges posed by hybrid cooled data centers, an investigation delves into the heat capture ratio for servers equipped with cold plates. The study quantifies heat capture by different mediums, considering varying temperatures and flow rates. It also delves into the impact of immersing hybrid cooled servers under diverse ambient conditions. The following research also evaluates strategies to optimize heat sinks for immersion cooling environments, considering a wide range of fluids and design parameters. This includes simulation-driven optimization of cold plates' configurations, enabling improved performance for processors characterized by high Thermal Design Power (TDP). Moreover, the study delves into the enhancement of air-cooled heat sinks for single-phase immersion cooling setups, considering fluid thermo-physical properties for optimal design. Through these interconnected chapters, the dissertation endeavors to provide a comprehensive understanding of energy-efficient and optimized data centers, encompassing a spectrum of strategies and technologies aimed at addressing the challenges associated with increasing power densities in modern chip architectures.

Keywords

Electronic cooling, Data centers, Thermal management

Disciplines

Aerospace Engineering | Engineering | Mechanical Engineering

Comments

Degree granted by The University of Texas at Arlington

31783-2.zip (4295 kB)

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.