Optimization of hot and cold aisle rack/cabinet configuration in a data center using CFD analysis

OVERVIEW OF HOT AND COLD AISLE CABINET DATA CENTRE:

In the current fast-paced technological environment, effectively managing data centers has become an essential element for organizations aiming to enhance performance and minimize operational expenses. A key aspect of achieving this goal involves adopting hot and cold aisle containment strategies within server racks. Hot and cold aisle racks are the configurations used in data centers to optimize airflow and temperature control. The hot aisle contains equipment that expels hot air, while the cold aisle draws in cool air to maintain ideal operating conditions for servers and other hardware. By segregating airflow and strategically managing temperature variations, hot and cold aisle configurations play a pivotal role in maintaining the ideal operating conditions for IT equipment and improves overall efficiency. Carefully positioning of servers and racks in hot aisle is the key to ensuring proper airflow management. When setting up cold aisle racks in a data center, airflow direction and enclosure design are crucial factors to consider. Additionally, positioning of close control air conditioner (CCU) should facilitate proper airflow management to maximize cooling efficiency. Computational Fluid Dynamics (CFD) analysis offers a wide range of advantages, including the enhancement of energy efficiency and optimization of cooling systems. It also aids in reducing downtime, improving equipment performance, and achieving significant cost savings.

CLOSE CONTROL AIR CONDITIONER ( CCU)

CCU, also known as a close control air conditioner, a computer room air conditioner, or a server room air conditioner, is designed to regulate temperature and humidity to a precise fixed point. It comprises of an Indoor evaporator module and a condensing module. Additionally, the system includes a humidifier, a reheater, and a safety thermostat. The microprocessor is responsible for managing the entire unit and has the necessary capabilities to interact with other units to ensure environmental conditions are met.

3D CAD MODEL DEVELOPMENT

The 3D CAD model of the data center is created using PTC Creo CAD software, utilizing the layout drawings provided by our client. The model is meticulously developed to meet the requirements for conducting CFD analysis.

CFD Mesh

Due to the large dimensions of the CAD model, it has been divided into four parts and a CFD mesh has been generated for each individual part. These meshed models are then assembled, ensuring that proper interfaces are provided according to the best practices in the CFD industry. The overall mesh count amounts to 24 million cells, comprising a combination of Polyhedral and Hexahedral elements (Poly Hex core mesh). The quality of the mesh has been maintained in accordance with the best practices in the CFD industry.

CFD Modelling

The flow is modelled as 3D, steady state, viscous and Incompressible. Turbulence closure is implemented using the K-Omega SST Turbulence model. Fan boundaries are applied at specific locations. Conjugate heat transfer effects are taken into consideration in the CFD analysis, while heat dissipation is simulated using a volumetric heat load approach.

HEAT LOAD

The total thermal load of the data center is approximately 950 kW, which is handled by 10 Closed Control Unit (CCU) air-conditioning units. These units are designed to maintain the required temperature and airflow conditions within the data center by removing the heat generated by IT equipment and other internal sources. Of the 10 installed CCUs, 2 units are maintained as redundant units to provide cooling-system reliability and operational flexibility. This arrangement ensures that the data center can continue to meet its cooling requirements during maintenance or in the event of a unit failure, subject to the available capacity of the remaining operating units.

TEMPERATURE FIELD

The data center has a total thermal load of approximately 950 kW, which is managed by ten Closed Control Units (CCUs) to maintain the required operating temperatures and ensure effective heat removal. Of the ten installed units, two are designated as redundant to provide operational reliability and cooling-system resilience in the event of equipment failure or maintenance. To evaluate the effectiveness of the cooling arrangement, a Computational Fluid Dynamics (CFD) analysis was carried out for five distinct configurations, considering different combinations of redundant CCUs being taken out of operation. The simulations were used to assess the airflow distribution and temperature fields throughout the data center, with particular emphasis on identifying potential hot spots, uneven cooling, and areas susceptible to inadequate heat removal. Based on a comparative assessment of the temperature distributions across the five configurations, recommendations were provided to the client identifying the redundancy combinations that should be avoided to minimize the risk of localized overheating and maintain reliable thermal conditions within the data center.

The results of the CFD analysis indicate that the existing cooling system is adequate to handle the data centre’s total thermal load of approximately 950 kW, while maintaining acceptable temperature distribution and airflow conditions throughout the facility. The analysis also confirms that the cooling infrastructure provides a sufficient level of redundancy to support reliable operation during planned maintenance or the unavailability of selected CCUs. However, the comparative assessment of the different operating configurations highlights that certain combinations of redundant units may adversely affect temperature distribution and increase the risk of localized hot spots. Accordingly, the recommendations provided to the client identify the configurations to be avoided, helping ensure effective cooling performance, operational reliability, and thermal stability under the evaluated conditions.