CFD engineering of data centers
Optimizing the climate performance of data centers using CFD simulation
- Urbanization of hot and cold alleys
- Validation of maximum air temperature at equipment inlet
- Recommendations on ice water temperatures
- Study of critical failure scenarios
- Identification of bypass and recirculation air flows
- Improved air distribution and cooling efficiency
- Precise dimensioning of grating layout
- Optimizing placement and control of air handling systems
- Study of the setpoints of air treatment systems.
What are the uses of CFD simulation for data centers?
Understanding the interaction between different air conditioning systems in data centers
CFD modeling is typically used during the design phase for sizing analysis of facilities. Computer simulation provides information on the relationship between the operation of air conditioning systems and variations in the thermal load of IT equipment. With this information, IT and site personnel can optimize airflow efficiency and maximize cooling capacity.
Data centers and server rooms house a wide variety of equipment and systems. An abundance of subsystems, such as rows of storage networking equipment, network switching equipment, and high-density servers, results in a multitude of heat sources and air conditioning systems. Consequently, understanding how these different systems interact and affect air distribution and pressure within a room is very challenging.
Our computer modeling methods, and in particular computational fluid dynamics (CFD) tools, allow us to analyze these options in detail and to graphically represent the temperature distribution , the flow propagation velocity , and the atmospheric pressure in the different spaces
CFD modeling is not a simple solution to a problem, but a pragmatic application that can help optimize the overall understanding of the project or explore how it is possible to reduce the operating costs of an existing facility.
On the other hand, it allows validation by simulation of the climatic sizing of a room composed of high-density servers in an extreme scenario .
The following are the advantages that can be obtained from the use of CFD modeling in the design of a server room, or when analyzing a data center in operation.
Testing hypotheses about the operation of cooling systems (CRAH) in the server room
- Crisis management – Will the backup air conditioners be able to handle the load if one of the main air conditioners fails?
- The company plans to virtualize applications; what happens if you add a new density array?
- Will the data center operate more efficiently if the IT equipment is placed differently?
- Will the data center operate more efficiently if the cooling system is changed?
- What happens if the temperature of the air entering the computer equipment increases?
- What happens if the temperature of the chilled water supplied to the CRAH is increased?
- Is my system even viable?
Digital twin – Data center
Our integrated research approach
Preliminary design: Crisis scenario study
When an EOLIOS engineer with expertise in data center design works with your team, they can resolve data center design challenges . Furthermore, our mastery of climate control for this type of facility guarantees results and prevents potential problems for the client.
Most importantly, by simulating crisis scenarios , unknown variables and the consequences of problems (power outage from the supplier) can be avoided from the outset.
By using CFD modeling in the design phase, power and cooling systems can be optimized before construction and before investment in a project. CFD modeling allows site personnel and IT specialists to validate system functionality under all crisis conditions, thus mitigating the risk of malfunctions during operation.
We are studying the restart scenarios of the safety equipment , the loop inertia , and the temperature rise to calculate the maximum temperatures at all points in the space and to identify the most unfavorable computer servers.
Why perform a CFD simulation of a data center?
CFD modeling: an essential tool for optimal server room design
The initial investment in CFD modeling pays off many times over . The CFD model allows for simple and easy-to-understand 3D images for all project stakeholders, which can be analyzed by considering both static images and dynamic videos.
If necessary, the information is protected, because our files can only be read on media players for which a special driver provided by EOLIOS is installed.
These color-coded 3D maps of server rooms show heat distribution, air velocity, and pressure within the space, allowing us to identify areas that need refinement or a complete redesign . During the design phase, we typically optimize the placement of floor grilles to ensure consistent airflow across all servers.
“The airflow through a raised technical floor grille can vary by more than 50% depending on its positioning, the room configuration, and its distance from the air conditioning systems.”
Once the macroscopic modifications have been made (room layout, hot and cold aisles), the design documentation can be optimized by making the final adjustments (specific sizing of the diffusion grilles, sizing of the CRAH). Finally, the system’s reliability is verified using transient simulations, recreating cooling system failure scenarios.
Conclusions and recommendations can be drawn from the information gleaned from transient scenarios to provide a safe and reliable operating environment for your IT investment. Thus, EOLIOS’ CFD modeling services offer significant added value for investors and integrators, as well as for end users, who gain a clear picture of the risks their data center faces.
Why run an external CFD simulation of a data center?
Define power losses under extreme conditions
Data centers are sites where heat extraction density is very high. The sharing of systems, the stacking of server rooms, and the continuous increase in the power of server racks lead to extremely high heat dissipation on the roof. Since the dissipation surface area is limited by the building’s dimensions, this results in a very high concentration of rooftop air conditioning systems, leading to a significant risk of power loss, or even cascading system failures, during extreme weather conditions.
In very high heat, or during power outage scenarios on the site, plumes of superheated air generated by the cooling systems can generate a sharp drop in cooling power, or even a cascading shutdown of the roof systems leading to a generalized failure of the site.
External CFD studies of data centers allow for the study of power loss risks and the optimization of the layout of rooftop cooling systems .
Studies of the risks of pollution in premises
In parallel, it is possible to verify that the fumes (NOx) released from the generator sets are not captured by rooftop air handling units, which would lead to air pollution in the offices.
Find out more:
Our specific solutions for data centers :
CFD engineering of data centers
Find out more
Thermal storage study
Find out more
External CFD simulation for data center
Find out more
Data Center Fire Simulation
Find out more
Thermal study of technical premises
Find out more
Urban heat island impact study for data centers
Find out more
Audit, diagnosis and 3D modeling of existing data centers
Find out more
Energy optimization and PUE calculation for data centers
Find out more
Commissioning: study and CFD modeling of load benches
Find out more
Designing your data center’s digital twin
Find out more
Data center engineering
Initially created in France, EOLIOS Ingénierie is the benchmark thermal and aeraulic simulation consultancy for data centers in Europe and worldwide. The company supports operators, designers and project owners at every stage in the life of a data center: design, optimization, renovation or extension.
By combining cutting-edge scientific expertise, state-of-the-art simulation tools and in-depth knowledge of the data center ecosystem, EOLIOS Ingénierie is a trusted partner for ensuring the availability, security and energy performance of IT infrastructures, while anticipating sustainability issues.
EOLIOS
PARIS – LONDON – MILAN – MUNICH – MADRID – WARSAW – CASABLANCA – UNITED ARAB EMIRATES – USA
