Author

Abhishek Guhe

Graduation Semester and Year

2016

Language

English

Document Type

Thesis

Degree Name

Master of Science in Mechanical Engineering

Department

Mechanical and Aerospace Engineering

First Advisor

Dereje Agonafer

Abstract

The skyrocketing growth in data centers, facilities that house information technology (IT) equipment for storage, management and distribution of data while striving for 24/7/365 operation with 100% up-time, has accounted for 1.3% of global energy use. A significant portion of data center energy is dedicated to removing the heat generated by IT equipment to maintain safe operating conditions and optimum computing performance. Energy efficient cooling of data center is of vital importance. One of the options for significantly cutting the cooling cost is the use of air side economization (ASE), Indirect evaporative cooling (IEC) and Direct evaporative cooling (DEC) without using expensive chilled-water systems or air-cooled CRAC units. The topology of a test bed modular data center (MDC) under consideration consists of an Information Technology (IT) module supported with a DEC and IEC module. MDC is a dynamic and complex environment with multiple mechanical and electrical control systems aimed at maintaining continuous operation of the data center. Highly nonlinear correlations, large number of constraints and multiple operating configurations make data center control a challenging research problem. In this study, a neural network model that adapts to the actual data center conditions using historical operating data and predicts the optimum configuration to reduce energy consumption is evaluated. In addition, the neural network model has the ability to learn from real time data collected from various data center sensors. When combined with a cooling unit, a predictive model that is fast and accurate in finding an optimal operating point for the modular data center unit can be implemented. In 2011, the Technical Committee (TC) 9.9 under American Society of Heating Refrigeration and Air-Conditioning Engineers (ASHRAE) expanded the operating envelope for data center thermal management in its Thermal guidelines for Data Processing environments, thus making it possible to operate IT equipment at higher server inlet temperatures and humidity and also switch to Indirect/Direct evaporative (I/DEC) and free cooling mode for increased number of hours per year. This study includes control strategies for operating I/DEC in tandem and also individually to achieve the target conditions for data center environment with minimum fluctuations in temperature and humidity. Staging of DEC for segmented cooling will provide flexibility in efficiently controlling temperature and humidity with significant water saving capability. Predictive cooling using weather forecast will counteract the start time delay of cooling modules avoiding ramping of unit due to unexpected weather conditions. The results show potential energy savings achievable through the proper implementation of control strategies and artificial neural network in operation of ASE, DEC and IEC of data center.

Keywords

Air-side economization, Evaporative cooling, Artificial neural networks, Data center cooling

Disciplines

Aerospace Engineering | Engineering | Mechanical Engineering

Comments

Degree granted by The University of Texas at Arlington

26935-2.zip (2606 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.