This simulator models the airflow path around detailed server racks: a cold-aisle supply that feeds rack fronts, a hot-aisle return that carries away exhaust, and a containment barrier between them. Change IT heat load, air volume, fan speed, containment effectiveness and recirculation, then watch server-inlet and return-air temperatures respond.
• Facility laboratory tab: a real-time 3D workbench of rack-mounted servers with dual power supplies, the cold-aisle containment/return path, the raised-floor supply plenum and the high-level hot-air return, with Home view, Focus selected part, Show full enclosure/exploded view toggle, Auto rotate, Expand and a clickable component list with callouts; live stats (server inlet, hot return air, coil leaving air, heat removed, chilled water supply, water return); an operating-sequence readout; and experiment controls — Pause/advance simulation (10 ms or 1 s steps), playback speed, and sliders/toggles for IT load, air volume flow at full fan, fan speed multiplier, containment effectiveness and uncontained recirculation fraction. • Curves & measurements tab: a temperature history chart (server inlet, return air, water supply) and a heat-flow chart (IT heat, coil heat removal, chiller electricity), the underlying model equations (Qcoil = ε·Cmin·(Treturn−Twater) limited by rated capacity; Qwater = ṁwater·cp,water·ΔTwater; room and water thermal-mass balances; Pchiller = Qchiller/COP) and snapshot measurements. • Experiments tab: guided scenarios (a contained baseline at 90% containment effectiveness, an opened-containment case with 70% recirculation, a reduced fan-speed case at half speed, and a higher rack-density case at 140 kW), a model verification bench (Run model checks) against independent fresh models, and a timestamped event log with a trial-report export. • Learn & assess tab: lessons on rack-front/supply-aisle orientation, how containment prevents mixing, the relationship between airflow and heat capacity, and why inlet (not return) is the measurement that matters, a knowledge-check quiz with reset, and a scope/references note.
Server racks are arranged so their intake fronts face a cold supply aisle and their exhaust faces a separate hot aisle or high-level return. In this lumped model, room air is treated as a single thermal-mass node: IT heat load raises the return-air temperature, and the CRAH/coil path removes heat from that return air before it recirculates back to the supply. Containment — physical panels, end doors and curtains — is what keeps the cold supply and hot exhaust from mixing.
The containment-effectiveness and recirculation-fraction controls determine what fraction of warm return air bypasses the coil and leaks back into the cold aisle before reaching server fronts. The simulator calculates server-inlet temperature as a mix of coil-leaving (supply) air and this recirculated warm air — so with poor containment, inlet temperature can run meaningfully hotter than the coil's leaving-air temperature would suggest, even though the coil itself is working correctly.
Watch the gap between server-inlet temperature and coil-leaving-air temperature: a small gap means containment is doing its job and little warm air is recirculating; a large gap under the same coil settings signals containment leakage. The heat-flow chart shows IT heat input against coil heat removal — at steady state these track together, and coil removal is capped by the coil/chiller capacity setting, so if IT load exceeds available capacity the room's thermal mass will absorb the surplus and return-air temperature will keep climbing rather than reach a stable value.
The Run model checks bench validates the underlying heat-balance model with independent fresh instances, leaving your live trial untouched. This is a lumped sensible-heat model (400 kJ/K room thermal mass, 850 kJ/K water-loop mass, cp,air = 1.005 kJ/kg·K, cp,water = 4.186 kJ/kg·K) — it does not solve CFD, humidity, condensation, refrigeration-cycle or hydraulic-network physics, and this particular airflow experiment fixes water supply at 12°C, coil capacity at 140 kW and water flow at 4 kg/s (adjust those in the dedicated CRAH or water-loop labs).
Both aim to separate cool supply air from warm exhaust air. Cold-aisle containment encloses the rack-front supply aisle so only conditioned air reaches server intakes; hot-aisle containment instead encloses the exhaust aisle and ducts it directly back to the cooling units. This simulator models the cold-aisle supply/hot-return arrangement with an adjustable containment-effectiveness control.
Because inlet temperature is a mix of coil-leaving supply air and any warm return air that bypasses containment and recirculates. If containment effectiveness is low, a larger recirculation fraction mixes into the supply before it reaches server fronts, raising inlet temperature independently of how well the coil itself is performing.
At a fixed heat load, moving less air mass per second means each unit of air must absorb more heat to carry the same total load away, which raises the temperature rise from supply to return air.
It is a lumped sensible-heat model with fixed room and water thermal masses — it does not solve computational fluid dynamics, humidity or condensation effects, the refrigeration cycle, or a detailed hydraulic network. Water supply temperature, coil capacity and water flow are held fixed in this lab so you can isolate the airflow and containment variables; adjust those water-side variables in the dedicated CRAH cooling lab.