Data Center Series 01 - External CFD

Over the years, one of the most challenging but yet fun technical tasks we had was working on data center computational fluid dynamics (CFD). So we wanted to start a data center series to talk about some basics, workflow, and design considerations.

Let’s kick it off with something often overlooked but super important: external CFD.

What is a data center?

Think of it like a library, but instead of storing books, they store and handle digital data, a lot of it. We live in a digital world and the data centers are the infrastructure to support various activities.

What are the key considerations when building a data center?

The reason why data centers are called mission critical is because ideally they should be reliably staying up and connected all time. So everything is designed to achieve that goal or at least minimize down time. This guiding principle leads to rigorous requirements in site, power, cooling, security, connectivity, redundancy, and operations.

What could be on the exterior of a data center site?

The building which will host those data modules and various equipment.

Generators, these kick in if the power goes out, usually sitting in big weatherproof boxes with their own fuel tanks. Those generators can be gigantic and release a significant amount of heat when turned on.

Power distribution equipment like transformers and switchgear yards.

Cooling systems like chillers, cooling towers, pumps, etc. As air cooled systems become more popular, the chillers yard can be quite extensive.

Security infrastructure including fencing, barriers, security guardhouse, etc.

What are the risks?

Data centers consume a significant amount of power and they generate tremendous amounts of heat. The heat needs to be removed effectively otherwise various equipment can derate or fail.

On the external side, there are two major concerns.

1) Chillers Yard

Exhaust Recirculation: Hot air from chiller heat rejection is drawn back into intakes, leading to derating or failure.

Cross Contamination: Generator exhaust (radiator and combustion) can reach chiller intakes, impacting performance.

Enclosure Effects: Canopies, dunnage, or screen walls increase pressure drop, causing chillers to derate or fail.

2) Generators Yard

Exhaust Recirculation: Hot air from radiators or stacks can re-enter intakes, reducing cooling efficiency and raising engine temperatures.

Overheating Risk: Restricted airflow can cause generators or nearby equipment to overheat, shut down, or derate.

Heat Trapping: Walls or canopies can trap heat, especially in hot or low-wind conditions.

Insufficient Makeup Air: Poor ventilation can create negative pressure, reducing combustion efficiency or starving engines of air.

Why should a data center developer evaluate the risk?

It’s often required by Service Level Agreement (SLA) and incompliance with SLA can have serious consequences. For the developer, you can’t manage the risk if you don’t evaluate it. It’s a complex challenge - but with the right analysis, it can be understood and effectively mitigated.

How to evaluate the risk?

Computational Fluid Dynamics (CFD) is a powerful tool to visualize and quantify airflow and thermal interactions in generator and chiller yards. By simulating critical conditions, CFD identifies risks such as hot air recirculation, exhaust cross-contamination, and pressure buildup. This allows engineers to proactively evaluate performance impacts, optimize equipment layout, and design mitigation strategies - long before problems surface in the field.

What does the process look like?

Step 1. Weather Study

Analyze site-specific weather conditions, focusing on prevailing wind direction, wind speed, and ambient temperature.

Step 2. Geometry

Build the model geometry, including but not limited to buildings, barriers, chillers, generators, transformers, dunnage, canopies, platforms, and other major structures.

Step 3. Equipment Details

Define equipment parameters per specifications. This may include chiller airflow rates, free areas, heat rejection values, and generator radiator and combustion intakes, exhausts, and stacks.

Step 4. Domain

Establish the computational domain size based on the site footprint and the height of the tallest object.

Step 5. Simulation

Run simulations under various conditions, as required by the design team.

Step 6. Analyze

Review simulation results to identify and assess potential thermal and airflow risks.

Step 7. Solutions

Collaborate with the design team to propose and evaluate solutions that mitigate the identified risks.

If you like this blog, follow along for more!

Contact allenm@clearbrook.energy if you have more questions.