As the demand for more data center space, power, and cooling continues to increase, the need for space management and equipment layout has become a critical design requirement. The increased load densities in the data center white space has led to an increased amount of support equipment, both inside and outside the data center building.
Inside the building
The mechanical cooling equipment such as the Computer Room Air Handling (CRAH) units, share space with electrical power equipment like the Power Distribution Unit (PDU) units in the CRAH gallery. They share space for several reasons which include but are not limited to PDU heat rejection contained in the gallery. The heat from the PDU is immediately rejected in the CRAH unit cooling system and this layout does not require a separate electrical room for the PDUs. Figure 1 below is a typical data center layout with CRAH units and PDUs sharing space in the CRAH Gallery.
Figure 1 Typical data center layout with hot aisle containment and CRAH galleries on the opposite side of the data center
Outside the building
Either in the yard, on the roof, or both, mechanical heat rejection systems such as cooling towers and chillers share space with electrical standby generators. A compact high-density layout inside the building results in a compact outdoor equipment layout. This leads to significant airflow management challenges and issues outside of the building.
Airflow patterns outside of the building are difficult to predict because of the different variables which design engineers and architects are not able to control. These variables include wind speed, air temperature and humidity, wind direction, and other activities surrounding the building. All these would impact the performance of the outdoor data center equipment.
The Computational Fluid Dynamics (CFD) analysis has become a critical tool for data center design for optimum yard or roof equipment layout. There are several CFD software vendors available on the market that are used for analysis and simulations. These simulations provide results that help data center owners and designers in the decision-making process in determining cost-effective layouts and performance. The CFD analysis when performed before the design is finalized and implemented helps in mitigating risks associated with design errors at the late stages of the design, which would result in significant and expensive change orders, construction delays or loss of compute capacity.
Most manufacturer’s equipment literature provides yard and roof equipment minimum placement clearance requirements. While this information is provided as guidelines, it is expected that designers will make a correct judgment when finalizing design layouts. A good example of this is a chiller manufacturer’s recommended minimum of 8-foot clearance from solid obstruction. However, these guidelines do not factor in ambient air conditions, wind speed, or height of the obstruction. Figure 2 below demonstrates how following a manufacturer’s recommended minimum clearance can result in undesired air-cooled chiller performance due to air recirculation.
Figure 2 Air Cooled Chiller Recirculation Demonstration
With the use of CFD analysis and simulation, the condition demonstrated on Figure 2 is identified and corrected before the design is finalized and construction is completed.
Another way the CFD analysis and simulation is a useful tool is by providing and understanding how multiple airflow patterns will interact with each other. Figure 3 below demonstrates the impact of generator exhaust air on the mechanical heat rejection equipment on the roof. In Figure 3 the hot exhaust air from the generator flue pipe and the radiator blown towards the building causes a stagnant air condition by the building. As a result of this condition the mechanical equipment inlet ambient air temperature range is outside of recommended manufacturer’s range which causes equipment not to perform as specified.
Figure 3 Generator Exhaust Air impact on Mechanical Equipment on the Roof
Conclusion
In conclusion, the Computational Fluid Dynamic analysis and simulation tools available in the market, have become a critical tool in design for new and existing data centers. CFD analysis is one of EYP Mission Critical Facilities’ most important and powerful tools used at every stage of the project to vet and verify that solutions proposed in our design will yield expected performance results. The expected performance results are subject to varying climatic and weather conditions at the data center location as typically published by DOE, NOAA, and ASHRAE.
About Author:
Gardson Githu, PE is a Senior Mechanical Engineer and Consultant at EYP Mission Critical Facilities.
Gardson’s experience focuses on the design and analysis of HVAC systems for commercial, industrial, and Data Center infrastructure facilities. His experience includes new facilities design, retrofit design, and mechanical systems analysis.
His project experience includes chilled water plants, thermal storage systems, fuel oil systems, and air handling systems. Gardson specialized in mechanical system energy optimization, data center risk site assessment and data center thermal mapping (computational fluid dynamic analysis).
He holds a Bachelor of Science degree in mechanical engineering from California State University Los Angeles, and a Master of Science degree in mechanical engineering with Themo-fluids option, from California State University Northridge.
He is a team member of the recently launched EYP Mission Critical Facilities and I3 Solutions Group Sustainability Initiative to offer a practical roadmap towards a Carbon Net-Zero data center by 2030.