EOLIOS's expertise in CFD simulation (Computational Fluid Dynamics) and in optimising cooling systems played a crucial role in solving the external thermal-airflow challenge of the data centers.
Our know-how made it possible to anticipate the rooftop thermal short-circuit phenomena and to validate the impact of the generators, thus guaranteeing optimal performance and efficient, sustainable thermal management for these complex infrastructures.
The essentials. Internal and external CFD study of an Italian data center (PAR5 & PAR6 sites). Internally, a risk of rack overheating (up to 35 °C, beyond the 28 °C setpoint) appears if two cooling units fail ; externally, rooftop plume dispersion is broadly under control, with localised short-circuiting under certain wind directions. Generators, a 3 m discharge stack and a digital twin validate reliable operation, including in degraded mode.
The impact of the thermal plumes outside the building is hard to predict because of the various variables that the design engineers and architects are unable to control. These variables include the wind speed, the air temperature, the wind direction (analysed via the local wind rose) and the topology of the surrounding structures.
However, these recirculation (or thermal short-circuit) phenomena have a direct impact on the performance of the cooling equipment positioned on the roof.
Re-ingestion, by a dry cooler, of part of the hot air discharged nearby (a generator plume or another exchanger). It raises the intake temperature and pushes the equipment out of its operating range.
EOLIOS supports you in studying the impact of these airflow issues to ensure optimal operation in all circumstances, even the most extreme (unfavourable wind scenarios or failures).
On this site, this includes the use of large dry-cooler networks and the optimisation of the hydraulic engineering (buffer tanks), aimed at guaranteeing sustainable heat removal and optimising operating costs. Flexibility and safety are at the heart of the design.
These data centers are designed with a redundant infrastructure (N+1 and N modes) to maintain cooling in the face of contingencies. This highly secure framework allows not only simplified maintenance but also continuous operation, even in the event of a power cut (blackout), thanks to the activation of the backup generators and UPS rooms.
Numerical anticipation plays a crucial role in the operational efficiency of these data centers. The ultra-detailed digital twin models the flows and virtually manages the critical operations, thus reducing the risks of thermal malfunctions linked to the outdoor environment. This modelling includes the analysis of the discharge stacks, the smoke dispersion and the impact of the built environment.
In short, thanks to their vast capacity, their rooftop cooling efficiency, their electrical redundancy and the intensive use of CFD simulation, data centers offer extremely robust solutions. By preventing localised overheating, these facilities become essential for companies seeking to support continuous, reliable global digital services.
Carrying out a CFD (Computational Fluid Dynamics) study is crucial for data centers in order to optimise the external airflow and the cooling, thus minimising the risks of hot-air recirculation while maximising thermal efficiency.
Cooling : the CFD simulations model the thermal plumes and the external airflows around the data center. This helps to validate the layout of the rooftop systems, the arrangement of the discharge stacks and the dry coolers to guarantee efficient cooling and avoid thermal short-circuiting.
Energy consumption : by analysing the winds and the external thermal dispersion, the CFD studies identify points of vigilance to preserve energy efficiency: heat downwash against the obstacles of the built environment, prevention of localised overheating on the roof.
Resilience management : the CFD models validate the behaviour of the installations during critical phases: optimisation of the buffer tanks and transient analysis during a blackout, facilitating the service continuity of the cooling resources.
Safety and backup : CFD simulates the degraded mode in the event of a power-grid failure, checking the robustness of the backup systems (N+1 and N modes). It validates in particular the correct dispersion of the flue gases from the generators.
By integrating these CFD analyses from the design stage, data centers can improve their operational efficiency, anticipate the impact of their direct weather environment and strengthen their safety, relying on a precise digital twin to consolidate the layout decisions.
At EOLIOS Engineering, we have fully grasped the importance of these external simulations for securing massive cooling systems. Thanks to our CFD expertise, we validate the overall operation of the site, reducing long-term risks and contributing to the sustainability of the data centers.
Optimising the data halls in data centers is essential, and CFD numerical simulation plays a major role in it. By modelling the airflows, the temperature distribution and the interaction between equipment, CFD makes it possible to design higher-performance cooling systems. It identifies the hot spots, improves air circulation and reduces energy costs by precisely adjusting the air conditioning and the layout of the racks. A strategic use of CFD ensures not only better operational performance and increased equipment reliability, but also contributes to reducing the carbon footprint of the data centers.
A data hall is an area dedicated to housing the IT equipment essential to processing and storing data. It is mainly made up of racks containing servers, storage units and network switches. To optimise cooling, the racks are often arranged in hot and cold aisles, separating the expelled hot air from the cool air intended to cool the equipment.
Racks organised in alternating corridors: the fronts draw cool air from a cold aisle, the rears reject hot air into a hot aisle. This separation limits the mixing of flows and improves cooling efficiency.
The cooling systems include precision air conditioners or water chillers, designed to maintain stable temperatures despite the heat generated by the servers. The electrical distribution relies on redundant systems, supported by uninterruptible power supplies (UPS) and backup generators, guaranteeing a continuous power supply in the event of an outage. Finally, smart sensors continuously monitor the temperature, the humidity and the performance of the equipment.
In this study, an overheating problem was identified in the left-hand part of the data hall, a crucial zone for the proper operation of the data center. This overheating occurs when two cooling systems fail simultaneously, and the racks concerned can reach 35 °C, well beyond the maximum limit of 28 °C. These conditions compromise the performance and reliability of the equipment, increasing the risk of failures that could affect data integrity and service continuity.
The configuration of the room and the installation of anti-intrusion grilles accentuate the problem by creating an uneven pressure distribution. An overpressure forms in the hot aisles on the left-hand side, preventing the efficient removal of hot air and causing a recirculation phenomenon, where the hot air returns into the system, accelerating the temperature rise.
The uneven distribution of the racks also contributes to the problem, as the failing cooling systems have to handle the thermal load of a distant island of racks, making this area particularly vulnerable. To solve this challenge, EOLIOS engineers designed several innovative solutions. Through in-depth dialogue with the client, they assessed the options and selected the most effective strategy, implemented in collaboration with all the stakeholders.
Optimising the cooling systems, such as the chillers and the generators, is essential to ensure the performance and reliability of the data centers. With high server densities that can exceed 10 kW per rack, thermal management becomes a major challenge. Uncontrolled overheating can cause hardware failures, reduce service availability and accelerate equipment wear.
Thus, thermal optimisation is not only about maintaining a stable temperature: it contributes to reducing the energy costs and the carbon footprint of the installations. Numerical simulations offer a detailed view of the temperatures and the thermal flows for different operating scenarios.
A precise 3D model of the infrastructure, such as the roof of a hyperscale data center fitted with ventilation enclosures, makes it possible to visualise all the critical systems and their interaction. The modelled structure includes the enclosure modules, the ventilation networks, the access walkways and the optimised arrangement of the cooling units. These models form the basis of the CFD simulations, essential for analysing and optimising the thermal performance.
Learn more: what are the causes and effects of a hot spot in a data center?The dry coolers play a central role in dissipating the heat generated by the servers. As AIR-WATER exchangers using outside air, they extract the heat from the cooling systems and ensure a constant flow of cool air. Their rooftop layout and their sizing must be carefully studied in order to avoid any hot-air recirculation phenomenon, liable to degrade the thermal performance.
An air-water exchanger that dissipates to the outside air the heat captured by the data center's water loop. Its efficiency depends directly on the temperature of the air drawn in, hence its sensitivity to short-circuiting.
The generators provide backup power in the event of a main-grid outage. Thanks to an automatic start-up, they take over the electrical load of the equipment and the cooling systems, ensuring service continuity and data protection. Their coordinated operation with the UPS guarantees a stable supply and high data center availability.
For each project, EOLIOS Engineering develops a detailed 3D model incorporating all the systems influencing the airflow, such as the dry coolers and the generators. The site plans, the 3D mock-ups and the equipment technical sheets make it possible to determine the essential characteristics such as the air flow rates and the dissipated power.
This digital twin makes it possible to precisely analyse the operation of the cooling systems and to identify the potential points of improvement. Thanks to their expertise in meshing and convergence, EOLIOS engineers guarantee reliable, robust CFD simulations. A fine, structured mesh precisely captures the variations in airflow and temperature, ensuring stable results representative of real conditions.
To simulate realistic conditions, a weather analysis based on the readings of the nearest station was integrated into the simulations. The outside temperature, the wind speed and the wind direction are critical variables that influence the thermal behaviour of the installations. The scenarios studied include different main wind directions and the impact on the generators. Two extraction models were compared:
Discharge at generator height.
Discharge via a 3-metre stack to limit hot-air short-circuiting and improve cooling efficiency.
Key takeaway. The 3 m stack proved to be the most favourable configuration: it moves the hot-air discharge away from the air intakes and clearly reduces the risk of rooftop thermal short-circuiting.
The weather analysis integrated into the numerical model made it possible to study the thermal dispersion according to various wind directions, relying on an annual wind rose to identify the most representative and most demanding configurations. The results show that heat dispersion takes place satisfactorily in the majority of the situations studied.
However, the analysis highlighted localised zones where thermal short-circuit phenomena can appear, mainly linked to the high density of the technical equipment on the roof. Under certain configurations, the wind favours the return of hot air towards the air intakes of the cooling systems. Although these temperature rises remain limited, localised and without critical impact on the overall operation of the site, they constitute important points of vigilance in an optimisation approach.
The study also made it possible to analyse the impact of the generators, the discharge stacks and the associated electrical installations during critical scenarios. Checking the robustness of the installations in degraded modes (N+1 and N) is essential, particularly during a blackout- or ride-through-type transient analysis.
Unlike sites requiring heavy physical modifications, the 3D simulations show that the thermal plumes from the generators do not significantly disturb the operation of the rooftop cooling systems. The layout of the stacks and the overall design allow an efficient removal of the hot flows. The results confirm that there is no notable contamination of the air drawn in by the sensitive equipment.
In modern data centers, the cooling systems play a central role, and the air drawn in must remain cool enough to guarantee their efficiency. To optimise this performance, the external thermal-airflow study also looked at the hydraulic engineering, notably the optimisation of the buffer tanks and the stratification phenomena.
The analysis confirms that the overall thermal behaviour of the site is under control. However, because of the recirculation phenomena identified on the roof, EOLIOS engineers recommend particular vigilance regarding certain wind configurations. Relying on the hot-spot mapping and the digital twin (Cloud Viewer), the project benefits from an anticipative approach that secures the reliability, performance and sustainability of the installations.
Drawing on its expertise in numerical simulation, and more particularly in external data center airflow, EOLIOS was able to validate the design choices and propose suitable recommendations to monitor the rooftop thermal short-circuit phenomena. Instead of imposing costly structural modifications, the study confirmed the relevance and overall consistency of the current layout, including for the generators.
The analysis of the different wind configurations made it possible to identify the specific conditions favouring the downwash of hot air towards the air intakes. The chosen solutions point towards increased vigilance during these weather scenarios, rigorously simulated in the 3D model. These observations confirm that the temperature rises remain limited, localised and without critical impact on the operation of the systems.
Thanks to this study, EOLIOS secured the design of the rooftop airflow systems from the design stage. This anticipative optimisation reduces the risks of thermal malfunctions and the performance losses linked to unfavourable outdoor conditions. The continuous use of the digital twin, notably via the Cloud Viewer, will support the decision-makers throughout the site's lifecycle.
Know-how: building the digital twin of your data centerInternal overheating, rooftop short-circuiting, generators and the digital twin: the answers to the questions from operators and design teams.
To secure the cooling of an Italian data center by CFD simulation, internally (data halls, overheating risk in the event of a double failure) and externally (dry-cooler and generator plumes on the roof), as on our External & internal CFD, Hyperscale project.
The re-ingestion, by the dry coolers, of part of the hot air discharged nearby. It raises the intake temperature and degrades the cooling. Under certain wind directions, the high density of equipment on the roof favours this phenomenon.
Discharging via a 3-metre stack moves the hot air away from the air intakes and proved the most favourable configuration to limit short-circuiting, compared with a discharge at generator height.
Yes. Simulations in degraded N+1 and N modes, including blackout-type transients, show that the generator plumes do not contaminate the air drawn in by the sensitive rooftop equipment.
It models the flows and virtually replays the critical operations. Combined with hot-spot mapping, it enables an anticipative approach throughout the site's lifecycle.
Explore our expertise, projects and technical papers to go further than the FAQ.
The study carried out by EOLIOS Engineering focuses on the thermal optimisation of hyperscale data centers through CFD simulations: improvement of air circulation, efficiency of the cooling systems, reduction of energy consumption and of the carbon footprint. Overheating and short-circuiting identified, with enclosures and digital twins in support.
Smoke controlSmoke-control engineering in a data center
HyperscaleExternal & internal CFD – Hyperscale Data Center
OptimisationCFD optimisation – Data Center
InternalData Center – DC28 – Internal
Technical roomsTechnical rooms – Data Center
CoolingCooling optimisation – Data Center
ExternalData Centers – DC15.1 & DC15.2 – External
ExternalData Center – PA 22 – External
GeneratorPressure-loss study – Generator – Data center
ExternalData Center – Paris
FireData Center – NOVEC gas
InternalData center – DC17 – Internal
ExternalData center – D14 – External
ExternalData center – DC25 & DC26 – External
InternalData Center – DC10 – Internal
InternalData center – DC25 – Internal
ExternalData Center – DC25 & DC26 – External
The simplest way is to talk it through together. Our engineers reply with an initial technical read — not a sales pitch.