A risk assessment or a powerful tool for the optimisation of data centre design: what should the role of CFD be in data centre design?

In the field of data centre design, striking the right balance between risk management, resilience, and efficiency is paramount. While sticking to proven designs can minimise risks, it can limit innovation and potential optimisation opportunities. With changing market dynamics and growing environmental concerns, there's an urgent need to maximise energy efficiency and reduce carbon footprints. Computational Fluid Dynamics provides the potential for a detailed and accurate insight into the operation of the data centre, throughout the design process. This is incredibly useful to highlight any risks within the design, but it also allows for multiple design options to be tested at an early design stage, highlighting potential opportunities for lowering energy consumption and carbon emission. These strategies can include reducing storey heights, optimising Hot Aisle Containment (HAC) required and reducing the clearances between data racks leading to a smaller data hall footprint.

At Bryden Wood, we advocate for a broader application of CFD in optimising and innovating data centre design, aiming for a more sustainable future within the industry.

Data Hall Design Optimisation

CFD Analysis of a Switch Room

CFD is traditionally used within data halls and Electrical Plantrooms to assess IT or Electrical Plants (e.g. UPS), both in normal running (N+X) and failure scenarios (N) and to ensure optimal cooling distribution and equipment performance. By integrating CFD early in the design process we can explore design options available to improve optimisation from both an economic and carbon standpoint.

Embodied carbon for a recent DC project.

MEP systems, especially cooling and power distribution, account for a considerable proportion of a data centre’s embodied carbon. By leveraging CFD, we can effectively minimise this impact by optimising cooling efficiency. Structures contribute to approximately 10% of the overall embodied carbon on a DC project, making it important to explore options that enhance design sustainability through material reduction.

Examples of typical options we analyse to optimise the design:

• Different equipment selections – We test different options to find a balance between cost, energy usage and performance.
• Cooling unit positioning – Our analysis ensures an even distribution of cool air is provided to maximise cooling efficiency, in both normal and failure scenarios.
• Quantity of cooling units – We analyse the effect of removing cooling units, to reduce cost and energy usage, and remove excessive, redundant cooling.
• Architectural features – We analyse the effects of altering the hot return ceiling void depths, extent of hot aisle containment, aisle pitches and reducing FFLs. This drives the reduction of the economic and carbon cost of the building and reduce building footprint.
• Structural Options – Various structural options can be considered to minimise material usage and reduce building height, carbon, and cost, whilst not compromising cooling performance.

CFD can provide evidence during the design process to minimise design risk and help optimise data hall designs further. This allows the design to be improved across all disciplines. Below is an example of a structural study, looking at how structural options affect the air pressures within the data hall.

Pressure Plots from a recent Structural Optimisation CFD Analysis Study

 

External CFD

External CFD Model

External CFD analysis plays a crucial role in data centre design by assessing the risk of hot air being entrained by cooling equipment. We can identify potential design risks and understand prevalence and impact. At Bryden Wood, our team utilises CFD analysis from pre-planning to detailed design stages, leveraging its insights to navigate the dynamic nature of site massing and reconcile technical requirements with local regulations. This helps optimise the process and allows architects to test options that may reduce the building footprint and massing.

External CFD is also a great tool for site-agnostic design. The same design can be tested across a range of weather conditions and orientations, with the risks and benefits of each highlighted. This insight at the early design stage is extremely valuable, for example the impact on plant and building layouts can be optimised when considering the prevailing wind direction.

It is important to highlight that we don’t only simulate building operation, but also interpret analysis results comprehensively, focusing on the average uplift across all the heat rejection plant, generators and building ventilation intakes rather than solely on the worst-affected equipment. This approach allows us to engage in dialogue with manufacturers to understand how any uplift will affect the performance of the equipment.

With this information, the amount of energy used for cooling can be minimised, reducing both embodied and operational carbon. 

Short Break Scenarios

Short break scenarios are simulations used to ascertain the behaviour of a data hall during a switch to standby power generation following a failure of mains power, employing transient analysis within Computational Fluid Dynamics (CFD).

This process has two main components. Firstly, modelling the data hall in 3D with the restart times of all equipment inputted into the model. And secondly, constructing a 'Flow Network' to represent the chilled water system, which interfaces with the 3D model to provide mutual feedback during the simulation. Through this comprehensive analysis, we can evaluate the performance of both the data hall and chilled water system at every stage of the outage and subsequent restart, thereby validating the resilience of the design and providing a level of assurance.

This analysis allows for the optimisation of design elements, such as UPS, to ensure inclusion only where necessary, leading to significant cost savings and a reduction in embodied carbon. Accurate determination of buffer vessel sizing through this analysis helps prevent oversizing, thereby saving costs and plant space, and minimising the need for structural reinforcement.

The optimisation of the chilled water system design ensures even distribution throughout the data hall. This optimisation process is crucial, as it prevents the tendency to oversize equipment or specify unnecessary components to meet resilience requirements, providing assurance that the building will meet requirements and allowing for greater potential for economic and carbon cost savings.

Conclusion

As the demand for data centre capacity continues to grow and challenges in electrical supply capacity become more pronounced alongside regulatory pressures, there is an increasing need for data centre designs to be optimised.

The integration of CFD into the early design process empowers stakeholders to push boundaries, optimise performance, and embrace sustainable practices, all whilst ensuring the functionality and resilience that is so crucial to the performance of these buildings.

Learn more about our approach to sustainable design.