Optimal supply air temperature ranges of various air-side economizers for a modular data center in South Korea

Abstract

Data centers account for 1.5% of world total electricity consumption, and a variety of studies have been conducted to save data center energy. Although many researchers have investigated the use of various economizers for data centers, little study has been conducted on the economizers’ performance on a modular data center in South Korea climate. The main purpose of this study is to present optimal supply air temperature ranges of various air-side economizers for a modular data center in South Korea. In this study, the optimization process is established by considering server thermal characteristics, thermal environmental ranges, and type of economizers. In the process, the simulation cases are established for the different economizer types, supply air temperature ranges, and heat exchanger sizes. The energy performance of each case is investigated through energy simulation. Based on the simulations, the optimal supply air temperature ranges of various air-side economizers for a modular data center in South Korea are suggested. This study is composed of 6 chapters, and each chapter is organized as follows: In chapter 1, the research background, objectives, and thesis outline are described. The current status of data center energy use and the motivations of this study are presented. Then, the specific goals and the structure of this study is described. In chapter 2, the general overview of a modular data center cooling system is presented. After presenting a conventional system configuration of cooling system in a modular data center, the control method of cooling system is presented. In addition, the indoor thermal guidelines related to the data centers are reviewed. Finally, this chapter is ended by presenting the method to calculate annual cooling load of a modular data center. In chapter 3, three types of economizers (direct type and indirect types with a heat pipe or an indirect evaporative cooler) used in this study are presented. From the system configuration of each economizers, its annual operation strategy, control flow chart, and pros and cons compared to other types are described. In chapter 4, the details of energy simulation are described. It starts with the energy optimization process and then presents models of a modular data center, servers, and cooling systems. In chapter 5, the simulation results are presented. The annual energy consumption of the conventional system and economizers are presented, and then the energy optimization results of each economizer are presented. Finally, based on the simulation results, the optimal supply air temperature ranges of each economizer are summarized by locations. In chapter 6, the conclusions drawn from this study are summarized: 1) More precise cooling load estimation is feasible in a modular data center through the process established in this study. In the process, the step-by-step methods to calculate data center cooling loads by component are presented. While this process is hard to be implemented in general data center in design phase because of the design changes in space use and equipment selection, it is useful in a modular data center due to its modular architecture and less design changes. 2) In the data center equipped with containment architecture and pressure controlled computer room air handler (CRAH) fans (usually in modular data center), the server thermal characteristics should be considered in data center simulations. Unlike the CRAH inlet and outlet temperature difference controlled data centers (i.e., open aisle architecture), the modular data center’s CRAH airflow and server airflow are similar. 3) Indirect air-side economizer with indirect evaporative cooler is usually more energy efficient in a modular data center located in South Korea. This is mainly attributed to the humidity level limitations of direct air-side economizers. In addition, the indirect air-side economizer does not introduce outdoor air which possibly contains contaminants. 4) The optimal supply air temperature range of various air-side economizers in South Korea is 18-23°C. In addition, the best energy performances are shown in Cheorwon, Daegwallyeong, Seoul, Sokcho, and Uljin. Although the results shown in this study are limited to the modular data centers in South Korea, this study gives clear implications; the higher supply air temperature might increase the total energy consumption in a data center, and thus the optimal supply air temperature ranges need to be estimated in design phase.