Size an industrial thermal management system around the heat that must be removed at the equipment’s most demanding credible operating condition, then check that the selected cooling method can reject that heat at the site’s actual temperatures. A cooling unit’s nominal capacity alone is not enough: its output can fall as ambient air or coolant temperature rises, while the equipment’s heat load may increase.

For technical evaluators, the decision comes down to four questions: how much heat enters the cooling boundary, where that heat must go, what temperatures the equipment can tolerate, and how much capacity remains under adverse conditions. The answers determine whether ventilation, a heat exchanger, active refrigeration, or liquid cooling is appropriate.

Define the cooling boundary before calculating the load

Start by identifying what the thermal system is expected to protect. The boundary might be an electrical enclosure, a motor, an inverter power module, or a complete machine with several cooling circuits. Each boundary has different heat sources and temperature limits.

A cabinet cooler, for example, must handle the heat released inside the cabinet and any net heat entering through its walls. It does not need to remove all the heat generated by a motor mounted outside it. Conversely, a liquid loop connected directly to inverter cold plates must remove the heat transferred into the coolant, even if most of the surrounding enclosure is cooled separately.

Record the location of each significant loss source and its path to the cooling system. Avoid adding the same heat twice: losses conducted from a component into a cold plate are already part of the liquid-loop load. Only the portion released into cabinet air belongs in the cabinet-air calculation.

Temperature limits should be assigned to specific points, such as cabinet air, coolant supply, winding temperature, or semiconductor junction temperature. These are not interchangeable. A cabinet air temperature below its limit does not prove that a poorly cooled component inside it is operating safely.

Build a heat-load estimate from operating losses

For electrical equipment, heat generation is related to losses, not simply rated power. When input power and efficiency are known at the relevant operating point, a first estimate is:

Heat loss = input power × (1 − efficiency)

If rated output power is the available figure, first estimate input power as output power divided by efficiency. Check the manufacturer’s definitions: an efficiency value may cover an entire drive, a motor, or only one part of the system. Adding losses calculated from an overall efficiency figure to separate component losses can overstate the load.

Suppose an inverter takes 100 kW of electrical input at an operating point where its efficiency is 97%. Its estimated loss is 3 kW. If the enclosure also contains a control transformer dissipating 200 W and other devices dissipating 300 W, their combined internal loss is approximately 3.5 kW. That figure is a starting point for the enclosure only if all of those losses enter the enclosure air. If part of the inverter loss leaves through an external heatsink or liquid circuit, allocate it to the appropriate cooling path instead.

Use loss data at the expected load and duty cycle wherever possible. Rated efficiency at full load may not describe part-load operation, switching losses, braking events, or repeated starts. For variable-speed drives, the worst thermal condition may occur during a sustained operating point rather than at the highest instantaneous power.

Include auxiliary heat sources within the boundary, such as power supplies, reactors, contactors, braking resistors, lighting, and fan motors. Also consider heat entering from adjacent machinery, hot process surfaces, sunlight, or warm incoming ventilation air when relevant. Solar and wall heat gains depend on the installation and temperature difference; they should not be treated as a fixed allowance for every cabinet.

Separate continuous load from short peaks

A cooler intended for steady operation must handle the sustained heat load under its specified conditions. Short peaks require a separate check of thermal storage and component limits. A heatsink, coolant volume, or machine structure can absorb some energy temporarily, but the allowable duration depends on its thermal mass, starting temperature, and heat rejection during the event.

For a simplified first check, excess heat energy is the difference between heat generated and heat rejected, multiplied by the duration of the event. That energy must be accommodated without any protected temperature exceeding its limit. Do not rely on average heat load alone when peaks can overheat a semiconductor, winding, or coolant circuit before the average becomes meaningful.

Set the temperatures that determine usable capacity

Cooling performance depends on temperature differences. Define the highest credible surrounding air temperature for air-cooled equipment, or the highest available facility-water temperature for a water-cooled system. Include local conditions: air near an oven, compressor discharge, or sun-exposed enclosure can be hotter than the temperature measured elsewhere in the building.

Then establish the maximum acceptable temperature at the point being controlled. Leave room for temperature rises between that point and sensitive components. A drive may permit a certain inlet-air temperature while its internal power devices operate substantially hotter; use the equipment manufacturer’s limits and any required clearances to establish the allowable operating window.

For passive or forced-air cooling, heat can only flow continuously from the hotter surface or internal air toward cooler ambient air. As the allowable temperature difference narrows, the airflow or heat-transfer area required for a given load increases. If the required cabinet temperature is at or below the surrounding air temperature, ordinary ambient-air ventilation cannot maintain it continuously. Active cooling or a cooler external medium is needed.

Review published capacity curves at the intended conditions, not just the model’s headline rating. An enclosure air conditioner, for example, may be rated at a different cabinet and ambient temperature from those at the installation. A liquid-to-air radiator is likewise affected by coolant inlet temperature, ambient temperature, airflow, and fouling.

Compare cooling methods against the installation

The practical choice depends on the required temperature, the quality of the environment, and the available heat sink. Each method has a different boundary and set of failure modes.

• Filtered ventilation: A relatively simple option when ambient air is sufficiently cool and the enclosure can tolerate air exchange. Check filter pressure drop, dust loading, humidity, and whether contaminants can reach sensitive components.
• Air-to-air heat exchanger: Keeps internal and external air streams separate. It suits sealed enclosures when ambient air is cool enough and there is sufficient temperature difference, but it cannot normally cool the cabinet below ambient.
• Enclosure air conditioner: Can maintain cabinet air below ambient and isolate it from outside contaminants. Evaluate capacity at site conditions, electrical consumption, condensate management, and service access.
• Liquid cooling: Can transfer concentrated heat from cold plates or equipment jackets to a remote heat exchanger. It is useful where air cooling is constrained, but adds requirements for flow, pressure, coolant quality, leak management, and facility heat rejection.

A liquid loop does not eliminate heat; it moves it. Confirm that the radiator, cooling tower, chiller, or facility-water system can reject the combined load at the required temperature. A skid sized for a given coolant supply temperature may fail to achieve it if the building water becomes warmer during peak conditions.

For a preliminary liquid-flow estimate, use:

Heat load = mass flow rate × coolant specific heat capacity × coolant temperature rise

Use consistent units and coolant properties at the relevant concentration and temperature. A smaller permitted temperature rise requires more flow for the same heat load. The resulting pressure drop must remain within the pump’s capability, including the cold plates, hoses, valves, filters, and heat exchanger. Also check each component’s minimum flow and maximum inlet-temperature requirements; an adequate total flow does not guarantee adequate flow through every parallel branch.

Apply a margin that addresses uncertainty

Once the load and conditions are defined, compare the required capacity with the cooler’s capacity at the design point. Add an allowance for uncertainties that the calculation does not capture well, such as incomplete loss data, filter fouling, heat-exchanger degradation, measurement tolerance, and foreseeable changes in equipment loading.

There is no single margin that fits every installation. A well-characterized sealed cabinet with measured losses may justify a different allowance from an outdoor machine exposed to variable solar loading and dusty air. Document the assumptions behind the margin so that future equipment additions do not silently consume it.

Oversizing also deserves scrutiny. A much larger refrigeration unit can introduce cycling and control difficulties if it cannot modulate at low load. Excessive ventilation can bring in more contaminants or moisture, while an oversized liquid pump may waste energy or exceed component pressure limits. Match capacity and control range to both the maximum sustained load and the expected range of normal operation.

Check the design as a complete operating system

Before selecting a unit, assemble a short design-point schedule: sustained and peak heat loads, ambient or facility-water limits, required equipment temperatures, operating duty, available power, and environmental constraints. Ask for performance data at those conditions, including the effect of any airflow restriction or coolant pressure drop.

Review what happens when cooling is reduced or lost. Temperature sensors, alarms, controlled derating, and shutdown functions should protect the equipment before critical limits are reached. Sensor location matters: a measurement near the cooler outlet may miss a hot pocket at a drive inlet, while a single coolant return sensor may not reveal inadequate flow through one branch.

Finally, verify the installed system under representative high-load and high-ambient conditions where practical. Measure temperatures at equipment inlets and other critical points, along with airflow or coolant flow where accessible. A satisfactory cooler nameplate rating is useful evidence; stable component temperatures under the intended duty are the stronger basis for accepting the selection.