Skip to the content

Solutions

Data centres

Dry coolers through a data centre's year

A data centre's dry coolers are sized for its hottest hours and run most of the year below full speed. DraftSolver rates plate-fin dry coolers, forced and induced draft, on the coil's own geometry, and runs them through a weather year with the fans held to the supply temperature: the fan energy, the hours without a chiller and the water the pre-cooling pads use.

What DraftSolver does for data centres

Data-centre MEP consultants and design-build firms.

  • The coil's own geometryThe air side of a plate-fin coil is calculated from its own dimensions, on the relations of Wang, Chi and Chang (2000) by default or of Rich (1973), with no fitted characteristic required. The tubes carry water or an antifreeze solution.
  • Fans held to the supply temperatureAn annual run solves the fans' speed at every hour so that the water leaves at its set point: slower while the air is cool, faster as it warms. The year reports the fan energy, and the hours the fans alone held the set point; on a data centre those are the hours without a chiller.
  • Adiabatic pre-coolingWetted pads or sprays cool the inlet air towards its wet bulb once the dry bulb reaches the temperature set for them, with the pads' own pressure drop on the fans. The year reports the water they use, their bleed included.
  • Sizing and sweepsDesign fixes a target and solves for one dimension of the cooler, and a parametric sweep repeats the rating along one or more axes, such as the ambient temperature.
  • Banks of coolersA bank of units is rated as a bank: with its width stated, or under a wind, the warm air that recirculates from its plume into its inlets is charged.
  • Fan and load transientsThe Transient simulation follows a cooler through a fan's start from rest or its stop, a step in the load or in the ambient temperature, or the ambient temperature following a profile.

The systems

Dry coolers

What it was compared with

The dry coolers' claims as the app states them, and the cases on plate-fin coils.

Table 1. What the dry coolers' models predict, forced and induced draft alike, and what that was compared with.
AnalysisPredictsCompared withAgreement
Steady rating, and the sweep, design and annual runs built on itAir mass flow rate, heat transfer rate, fan power and fluid outlet temperature of a plate-fin coil coolerWang, Chi and Chang's (2000) coil database; the draft side, fan and platform on Kröger's Examples 8.1.1 and 8.1.288.6% of the coil database within 15%; no measured dry cooler is reproduced
Discretised transientFluid outlet temperature, heat transfer rate and air mass flow rate through a fan ramp or a load changeIts own steady rating, the state it settles to; no measured start-up or run-downSettles to the steady rating within 0.1% on the air mass flow rate and the heat transfer rate

The validation record, case by case

The plan that fits

The Data Centre edition holds the dry coolers alone, for rating, sweeps, design and annual runs, monthly or yearly. The Transient simulation and CFD of the coolers are on Expert.

  • Data Centre

    Data-centre MEP consultants and design-build firms sizing dry coolers.

    • Rating, sweeps, design and annual runs
    • Systems: Data-centre dry coolers

    Request early access

  • Expert

    Owner's engineers, performance teams and test agencies: acceptance tests, guarantees and disputes.

    • Rating, sweeps, design and annual runs
    • ASME PTC 30.1 curves and acceptance
    • The discretised transient
    • The Monte Carlo band
    • CFD
    • CFD lab
    • 2,000 CFD credits a year
    • Systems: Data-centre dry coolers, Mechanical draft and Natural draft

    Request early access

The plans side by side