Internal CFD study of a data hall: analysis of the airflows, the temperature distribution and hot-spot detection to optimise server cooling.
EOLIOS Engineering provided an overview of the thermo-aeraulic conditions of the various phenomena taking place in the data center room, in relation to the IT processes and the climate-control systems.
Key point. Internal CFD study of a data center room in Spain. The challenge: mastering the thermo-aeraulic phenomena induced by the hot and cold aisles and assessing the efficiency of the air mixing that cools the servers. EOLIOS builds a highly detailed digital twin (air volumes, server lines, lighting, material conductivity, all the HVAC systems and diffusion grilles), simulates various recycler-shutdown cases and runs a transient study during a power outage to protect the hardware.
The challenge of such a project is mastering the particular thermo-aeraulic phenomena induced by the presence of hot and cold aisles, as well as assessing the efficiency of the air mixing intended to cool the servers. The CFD studies made it possible to visualise these phenomena for various recycler-shutdown cases, leading to the optimisation of the HVAC treatment systems.
The digital twin of the data center reproduces the air volumes, all the server lines, the lighting and the material conductivity. All the HVAC systems are modelled, and the diffusion grilles are modelled precisely to faithfully reproduce the airflow in the different aisles. Such a detailed model allows highly accurate calculations and the rapid identification of potential thermal problems.
A data center air-handling unit that draws in the hot air rejected by the servers, cools it and returns it towards the cold aisles. Studying its various shutdown cases conditions the resilience of the cooling.
The entire set of servers could be modelled in CFD. The recyclers, the lighting, the heat-extraction fans, the equipment systems and the electrical power cabinets are precisely integrated into the study.
The resolution fineness of the CFD reproduces a complex temperature distribution that is nonetheless very close to the real conditions of use. The velocities and temperature distributions at the recycler outlets, as well as the heat exchanges in the hot and cold aisles, are captured with great precision. The first calculations verified the absence of any flow problem and made it possible to assess the energy consumption of the recyclers and to size their supply.
Once the digital twin is calibrated, the CFD serves as a test bench to size the recyclers' supply: it links the supply flow rate and the return temperature of each unit to the thermal load actually seen by the aisles. This verifies that the installed cooling power covers the servers' dissipation without over-ventilating, and that the redundancy keeps its promises when one or more units shut down.
Load balance per aisle: the dissipated power of the server lines set against the flow rate and return temperature of the recyclers.
Recycler-shutdown cases: load transfer onto the remaining units and verification of the intake temperatures.
Mixing homogeneity: spotting the under-supplied aisles and the residual recirculation zones.
Safety margin: validating the behaviour in degraded mode before moving on to the extreme scenarios.
The principle of providing backup recyclers (N+1 or more) so that the shutdown of one unit, in failure or maintenance, does not compromise the cooling. CFD verifies that the remaining units keep the intake temperatures below the thresholds.
Key takeaway. Sizing the recyclers as tightly as possible and testing the redundancy by calculation avoids both costly over-ventilation and under-cooling in degraded mode.
Numerical modelling helps designers to predict the performance and behaviour of the systems, to test different configurations and to decide in full knowledge. It is a powerful tool for detecting hot spots during an extreme scenario (heatwave, air-conditioning-system failure), taking into account the location of the air inlets and outlets, the temperature and velocity of the ambient air and the equipment configuration.
Finally, a transient study assessed the temperatures near the servers during a power outage, over the time between the shutdown and the restart of the ventilation systems by the generator, in order to verify that the electronic hardware is not damaged.
An unsteady simulation that follows the evolution of the temperatures second by second during an event, such as a power outage. It quantifies the delay before reaching the critical thresholds and compares this delay with the recovery time of the backup systems.
Key takeaway. The detailed digital twin makes visible the recirculation and hot spots invisible to the eye, and the transient study proves the room's behaviour before the real incident.
Digital twin, recyclers and transient study: answers to the questions operators and designers ask before an internal CFD study.
Mastering the thermo-aeraulic phenomena induced by the hot and cold aisles and assessing the efficiency of the air mixing that cools the servers, by visualising various recycler-shutdown cases to optimise the HVAC systems, as in our internal study of the DC28 data center.
It reproduces the air volumes, the server lines, the lighting, the material conductivity and all the HVAC systems, with precisely modelled diffusion grilles. This level of detail allows highly accurate calculations and the rapid identification of thermal problems.
Through the simulation of extreme scenarios (heatwave, air-conditioning-system failure), taking into account the location of the air inlets and outlets, the temperature and velocity of the ambient air and the equipment configuration, in order to identify the hot spots, their sources and their consequences.
It assesses the temperatures near the servers during a power outage, over the time between the shutdown and the restart of the ventilation systems by the generator, to verify that the hardware is not damaged.
Because their outlet velocities and temperatures govern the exchanges in the aisles. Capturing them finely makes it possible to assess their energy consumption and to size their supply correctly.
Explore our expertise, projects and technical papers to go beyond the FAQ.
Internal CFD study of the DC17 data center: hot-spot identification, cold-air distribution and containment efficiency.
Digital twinData center study – Data Hall and UPS rooms
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
ExternalData center – D14 – External
ExternalData center – DC25 & DC26 – External
InternalData Center – DC10 – Internal
InternalData center – DC25 – Internal
ExternalData Center – DC25 & DC26 – External
CFD expertise, delivered projects and technical dossiers: the whole Data Center field in one place.