CFD for datacentre cooling optimization

Datacentres are a growth area as the demand in the global market for data storage and processing resource increases. However, all this power hungry equipment generates a significant amount of heat which has to be carefully managed to ensure it runs properly.  Sustainability is also an important factor to consider with more and more information now being moved into centralised datacentres, putting a strain on energy costs and increasing carbon footprint. It's also vital that a data centres systems are resilient to cooling system failure in order to avoid thermal runaway and prevent damage to IT equipment. Therefore, energy efficiency and heat management have become increasingly important in data centre design.

Datacentre thermal modelling using Computational Fluid Dynamics (CFD) is becoming more important for data centre infrastructure management and efficient cooling design. It is used as a design and verification tool in support of decision-making, troubleshooting, design and optimisation of existing and new data centre sites. It is used to forecast data centre performance by quantifying the airflow and temperatures that accompany physical changes to the layout of equipment and power loads, as well as exploring design resilience should cooling systems be taken offline for maintenance or experience an unplanned failure.

This presentation outlines how a  STAR-CCM+® methodology can be used to perform datacentre cooling studies, similar to the specialist datacentre modelling packages marketed at users with limited CFD experience. It provides an overview of typical ASHRAE design criteria and describes how various IT components are represented for the CFD analysis. A semi-automation method using macros will be also be outlined for both model creation and the control feedback process.

Presenter(s): 
WSP Group
Rama Pathakota
Session Time Slot(s): 
Time: 
07/03/2017 - 2:50pm-07/03/2017 - 3:15pm
Room: 
Room I
Presentation Image: 
Session Track: 

Session Tracks (STAR Global Conference 2017)