1-833-888-0091contact@hvacengineering.ca

Load bank testing for cooling validation

Explains how temporary heat loads are used to prove data center cooling capacity, controls and failure response before IT equipment is installed.

CRAH Rack Cold aisle 18-27 °C Hot aisle Return plenum Liquid rack CDU
Data center cooling and HVAC design

A data center's cooling system is designed to remove a specific amount of heat under specific conditions, with specific responses when equipment fails. The only way to know whether it does so before servers arrive is to give it heat. Load banks, electrical devices that convert power into heat, provide that heat in a controlled and measurable way.

Load bank testing is standard practice in data center commissioning. Done well, it reveals capacity shortfalls, airflow problems, control instability and failure response issues while they can still be fixed without risk to production equipment.

Why use load banks

  • Capacity verification: prove that cooling equipment delivers rated capacity at design conditions.
  • Airflow validation: confirm that air reaches every rack position and that containment works.
  • Control tuning: tune fans, valves and sequences under realistic load rather than at idle.
  • Failure testing: fail cooling units, pumps and chillers under load and measure the temperature response.
  • Electrical testing: the same load banks also test UPS, generators and distribution, often in the same campaign.

Types of load banks and heat simulators

TypeHow it worksBest for
Rack-mount load banksFit in rack positions, draw front-to-back air like serversRealistic airflow and containment testing
Floor-standing heater fansPortable units placed in aisles or rowsRoom-level capacity tests
Large resistive load banksHigh-capacity units, often outdoors or in electrical roomsGenerator and UPS testing; not ideal for cooling distribution
Liquid heat load simulatorsHeat a liquid loop to mimic cold plate loadsDirect-to-chip and CDU testing

For cooling validation, rack-mount load banks give the most representative results because they reproduce server airflow paths. Floor-standing units are cheaper and faster to deploy but can create unrealistic air patterns. Liquid simulators are essential for testing CDUs and liquid cooling loops before IT hardware arrives.

Matching airflow, not just heat

Servers move a certain volume of air per kilowatt. A load bank that produces the right heat but moves much more or much less air will distort the test. Good practice is to select or set load banks so that both heat output and airflow, and therefore temperature rise, approximate the expected IT equipment. Ask the IT team for expected airflow per kW or temperature rise across servers.

Planning the layout

  1. Distribute load to match the design. Place load banks in rows and positions that represent the planned rack layout and density, not clustered near cooling units or power sources.
  2. Include high-density zones. If certain rows are designed for higher loads, load them accordingly.
  3. Install containment and blanking as they will be in operation, so airflow paths are realistic.
  4. Plan power distribution. Load banks are connected to the actual rack PDUs and distribution where possible, which tests electrical paths at the same time.
  5. Stage the load. Plan steps such as 25, 50, 75 and 100 percent of design load.

Test procedure

Preparation

  • Confirm Level 3 commissioning is complete: startup, controls checkout, balancing.
  • Install calibrated temperature sensors or data loggers at rack inlets (top, middle and bottom), cooling unit supply and return, and in containment.
  • Confirm BAS and DCIM trending at suitable intervals.
  • Agree acceptance criteria in advance, usually rack inlet temperatures within the ASHRAE TC 9.9 recommended range or the OPR limits, and maximum temperature during failure events.

Steady-state capacity tests

  1. Apply load in steps, allowing temperatures to stabilize at each step.
  2. Record rack inlet temperatures, cooling unit conditions, water temperatures and flows, and power.
  3. At full load, verify each cooling unit's capacity by measuring airflow and temperature difference, or water flow and temperature difference.
  4. Check for hot spots and recirculation; adjust tiles, blanking or containment if needed.

Failure and transient tests

  1. Fail a single cooling unit and record the temperature response and standby unit start.
  2. Fail pumps, chillers and CDUs in turn.
  3. Simulate loss of power to cooling equipment and measure the rise in inlet temperatures during restart.
  4. Test economizer transitions if outdoor conditions allow.
  5. During integrated systems testing, simulate utility power loss and observe the full sequence with load applied.

Record peak temperatures and recovery times. Fast-sampling data loggers, at intervals of seconds rather than minutes, are needed to capture transient peaks. See continuous cooling and thermal ride-through.

What results tell you

ObservationLikely cause
Some rack inlets hot at full load while others are coolAirflow distribution, missing blanking, containment leaks
Cooling units cannot reach rated capacityLow water flow, wrong water temperature, coil or fan issue
Temperatures oscillateControl loop tuning, competing controls between units
Large temperature spike during unit failureSlow standby start, insufficient redundancy, poor unit placement
Return temperatures lower than expectedBypass air

Results can also validate CFD models, which can then be used confidently for future planning.

Safety

Load banks create large amounts of heat and draw significant electrical current. Safety measures include:

  • Qualified electrical personnel for connections, following CSA Z462 workplace electrical safety practices
  • Temperature limits and automatic shutdown on load banks
  • Fire watch arrangements as required by the site and the authority having jurisdiction
  • Coordination with fire alarm systems, since heat detectors may respond to high temperatures in hot aisles
  • Clear communication and an abort procedure if temperatures exceed safe limits

Canadian considerations

Season affects testing. Economizer modes can only be tested fully in cold weather, and peak mechanical cooling only in hot weather. In Canada, a single test window rarely covers both, so plan for seasonal testing or use test methods that simulate outdoor conditions. Cold weather testing also demands attention to freeze protection on systems that may be partly filled or idle during construction.

Frequently asked questions

Can we test with real IT equipment instead?

Real IT equipment rarely provides full design load at the time of commissioning, and owners are unwilling to risk production hardware in failure tests. Load banks provide controlled, repeatable full load.

How long does load bank testing take?

That depends on facility size and the number of scenarios. Steady-state stabilization can take hours at each step, and each failure test requires setup and recovery. Allow enough time in the schedule.

Key takeaways

  • Load banks provide controlled heat to validate cooling before IT equipment is installed.
  • Rack-mount load banks reproduce server airflow best; liquid simulators test CDUs and cold plate loops.
  • Distribute load to match the design layout and density.
  • Test steady-state capacity and failure transients, with fast data logging.
  • Manage electrical and fire safety, and plan for seasonal testing in Canada.

Talk to an engineer

Have a project or a problem this guide touches on? Tell us about the building and we’ll be in touch within one hour.