Actiflow extensive study helped to complete the risk-assessment, to optimize the design and to determine the energy use of the data centre.
Equinix is the leading global interconnection and datacentre company. Equinix data centres are located throughout most of the world, across five continents. Amsterdam is a network-rich and cloud dense location and is one of the most important central connectivity hubs in the world. Equinix currently operates 8 locations in Amsterdam and 2 in the east of the Netherlands.
In 2017 Equinix opened the AM4 data centre at the Amsterdam Science Park campus. This unique data centre consists out of a lower part (the existing AM3 data centre) and a high-rise part (the AM4 data centre). This combination provided several engineering challenges during the design and construction phase. One of these was the discharge of hot air and fumes from the cooling devices and emergency power units. The relatively small space available makes it less self-evident that the flow of fresh air towards and discharged air and fumes from the data centre are always separated.
Equinix’s subcontractors Unica and RHDHV asked Actiflow to perform an extensive analysis of the flow around the building. Therefore we made a detailed computational model including the surrounding buildings, all cooling devices on the roofs of AM3 and AM4, all exhausts and intakes and details like porous screens, lamellas and falls roofs. We then used this model to calculate the flow of air and the distribution of temperature and water vapour in several scenarios. These scenarios vary in operational mode (normal, maintenance, emergency) and weather conditions (high or low temperature, wind speed and direction).
The study provided both the subcontractors and Equinix with extensive insights in what could happen under a variety of circumstances. It was used to complete the risk-assessment, to optimize the design and to determine the energy use of the data centre. The insight the CFD-study provided made it possible that the related engineering challenges were solved to an optimum.