Seminar 56 - Using Optimization for Airflow Management in Data Centers and Operating Rooms

Track: Modeling Throughout the Building Life Cycle
Sponsor: 4.10 Indoor Environmental Modeling, 9.9 Mission Critical Facilities, Data Centers, Technology Spaces and Electronic Equipment
Chair: James W. VanGilder, P.E., Member, Schneider Electric, Andover, MA
The value of CFD modeling for indoor environment applications is well established. This seminar uniquely focuses on minimizing energy while ensuring the health of electronic and human occupants – in data center and operating-room applications. "Optimization" is the primary unifying theme with two presentations focusing on the familiar manual "design-stage" optimization carried out by the CFD modeler, one including a formal optimization study, and another proposing a formal optimization engine for ongoing operational control and energy savings in data centers.

1. Optimization Study of Stanchion Layout and Flow Partitioning to Achieve Uniform Airflow through Perforated Tilesin Data Centers
Cheng-Xian (Charlie) Lin, Ph.D., Member, Florida International University, Miami, FL
Airflow non-uniformity normally causes uneven cooling for computing servers in data centers. To tackle the airflow variation from perforated tiles, we perform an optimization study, which considers these two most important parameters, i.e., stanchion layout and flow partitioning (flow resistance). The response surface methodology based on a radial basis function is used to reduce the run time for producing a large set of genetic generations during an optimization process in which the tile airflow variation is minimized. As a result, guidance on the design of stanchions and perforated tile type selection is provided.

2. Optimizing Supply Airflow Location in Data Centers Using CFD
Ramin Rezaei, Associate Member, Southland Industries, Dulles, VA
Optimizing airflow management in data centers saves money by reducing the energy usage and avoiding hot spots. One of the main design criteria for data centers is the location of supply airflow. The three widely used data center cooling architectures are "underfloor", "overhead" and "sidewall" supply. In this presentation, the difference between the flow and temperature distributions amongst these 3 scenarios is investigated using CFD . Detailed three-dimensional analysis helps find optimum airflow temperature and flowrate.

3. Improving Data Center Efficiency with Active Airflow Control
James W. VanGilder, P.E., Member1 and Christopher Healey, Ph.D.2, (1)Schneider Electric, Andover, MA, (2)Schneider-Electric, Andover, MA
The efficient control of cooling for data centers is an issue of broad economic importance due to the significant energy consumption of data centers. Many solutions attempt to optimize the control of the cooling equipment with temperature, pressure, or airflow sensors. This presentation shows how simulation-based approaches combined with power-consumption models can reduce cooling energy consumption in data centers. It also provides a real-life case study to demonstrate how energy-saving cooling set points can be found using calibrated simulations and smooth metamodels of the system.

4. Optimizing Air Change Rates in an Operating Room Using CFD
Mehran Salehi, Ph.D., Associate Member, Southland Industries, Dulles, VA
Indoor air quality in hospital operating rooms is of a great concern for both patients and medical personnel, thus mandating the use of efficient HVAC systems and active gas scavenging systems. In this study, the velocity, temperature and particle distribution of an actual operating room is studied at different ACH values. Using CFD it is possible to determine the optimized ACH value and regions with high pollution concentration and ensure that the sterile zone is supplied with sufficient air to remove particles.

Presented: Wednesday, January 24, 2018, 8:00 AM-9:30 AM
Run Time: 90 min.

This is a zip file that consists of PowerPoint slides synchronized with the audio-recording of the speaker (recorded presentation), PDF files of the slides, and audio only (mp3) for each presentation.