This project used computational fluid dynamics (CFD) to evaluate how wind and heat flows across the roof of a large data centre in London affected the performance and risk profile of its external plant. The assessment represented all major rooftop systems, including dry air coolers (DACs), air-cooled chillers, hybrid coolers, and the exhaust from a diesel rotary uninterruptible power supply (DRUPS). The study characterised how wind direction and speed influenced recirculation, plume dispersion, and the potential for warm air to be drawn into hybrid cooler intakes. It found that most equipment operated with limited cross-interaction under typical conditions, while localised wind-driven recirculation around taller structures or exhausts could modestly elevate intake conditions. Where such effects were indicated, operational responses (such as temporary fan speed adjustments consistent with manufacturer guidance) were deemed sufficient to maintain performance margins without additional physical mitigation.
From a risk perspective, the analysis provided confidence that intake criteria could be maintained with routine operational controls. The likelihood of sustained adverse intake conditions was assessed as low, and the consequences were manageable through pre-defined responses and monitoring. The results also strengthened data-driven design and operations by identifying roof areas where wind effects were more influential, informing control set-points, alarm strategies, and maintenance priorities. Overall, the project offered a high-level assurance that the roof plant arrangement and control philosophy supported resilient cooling performance across a range of wind conditions, while providing clear guidance for operational risk management and ongoing performance monitoring.