Key takeaways
- A 2,000 × 800 × 600 mm enclosure with its back to a wall has about 4.96 m² of effective surface, so 1,500 W of internal loss raises it 55 K above ambient at 5.5 W/m²·K.
- Holding that 1,500 W to a 10 K rise needs 477 m³/h of airflow, which is two 300 × 300 mm louvres at 60% free area moving air at 2.4 m/s, plus a fan that still delivers at 250 Pa.
- Every 10 K above a capacitor's rated temperature halves its life, so a drive running at 55 °C internal instead of 40 °C keeps about 35% of its rated capacitor life.
- Treat a sunlit outdoor enclosure as 10 K hotter than the air around it; at 40 °C air that means designing to a 50 °C internal ambient, not 40 °C.
- Put the loss in watts, the ambient and the acceptable internal temperature on the drawing; a fabricator can cut the openings but cannot invent the thermal duty.
A drive panel with 1,500 W of internal loss, built into a 2,000 × 800 × 600 mm enclosure, standing against a wall in a plant at 45 °C, has an internal temperature of about 100 °C. Nothing on the drawing is wrong. The enclosure is the size the customer asked for, the powder is 75 µm, the door closes. The panel is simply trying to shed 1,500 W through 4.96 m² of steel, and steel at that area gives back 27 W for every kelvin it is above ambient.
Switchgear enclosure heat is decided by three numbers: the watts inside, the square metres of surface that can reject them, and the air changes per hour you are willing to pay for. Everything else — the louvre, the filter, the fan, the derating — follows from those three. A panel that is 15 K hotter than it should be does not fail on the day it is commissioned. It fails two years later, when the capacitors have aged at twice the rate and the breaker has been carrying 90% of its rating since day one.
This guide works the thermal problem for a fabricated enclosure: what the surface can shed, what ventilation adds, how the openings are sized, what the standard limits, and the failure modes that follow a hot panel. The cut-form-weld-coat route that produces the enclosure is covered in how a switchgear enclosure is manufactured; this page is about the thermal duty that route has to serve.
How much of a switchgear enclosure heat load can the steel shed on its own?
A painted steel enclosure in still air sheds about 5.5 W per square metre of effective surface per kelvin above ambient. That single coefficient covers natural convection on the vertical faces and radiation from a painted surface, and it is the number to use before anyone proposes a fan.
The effective area is not the total surface area. For the 2,000 × 800 × 600 mm cabinet, the total is 2 × (0.8 × 0.6) + 2 × (0.8 × 2.0) + 2 × (0.6 × 2.0), which is 0.96 + 3.2 + 2.4 = 6.56 m². Stand that cabinet against a wall and the back face, 1.6 m², is no longer rejecting heat into free air; it is rejecting it into a wall that is at the same temperature. The effective area drops to 4.96 m². Put it on a closed plinth and the base goes too.
At 5.5 W/m²·K and 4.96 m², the cabinet rejects 27.3 W per kelvin. A 500 W panel therefore runs 18 K above ambient, which is acceptable in a 35 °C switch room. A 1,500 W panel runs 55 K above ambient, which at 45 °C is 100 °C inside: past the rating of every device in it, and past the point where PVC insulation softens.
That is the whole decision. Below about 800 W in a cabinet this size, a passive enclosure with sensible device placement works. Above it, the enclosure needs an air path or a heat exchanger, and the drawing has to say which.
The thermal limits for the drawing
The table is the set of values we work to when an enclosure needs a thermal design rather than a wall thickness.
| Parameter | Typical value | Ganesh limit or note |
|---|---|---|
| Surface dissipation coefficient | 5.5 W/m²·K painted steel, still air | unpainted stainless is closer to 4.5 W/m²·K |
| Effective area, 2,000 × 800 × 600 mm | 4.96 m² against a wall | 6.56 m² free-standing, 6.08 m² on an open plinth |
| Internal loss input | watts, from the device list | the number the fabricator cannot invent |
| Reference ambient | 40 °C max, 35 °C 24 h average | IEC 61439-1 basis; 50 °C for a plant roof |
| Temperature rise of accessible metal | 30 K limit | above that the door handle is a burn hazard |
| Airflow for a 10 K rise | 477 m³/h at 1,500 W | 318 m³/h if 15 K is acceptable |
| Louvre face velocity | under 3 m/s | two 300 × 300 mm louvres at 60% free area give 2.4 m/s |
| Louvre free area | 55–60% of the nominal opening | pressed louvre, 250 T press, blade angle 45° |
| Filter class | G3 or G4, dirty-drop rated | change interval written on the door |
| Fan pressure at design flow | 250 Pa with a dirty filter | a 172 mm axial fan will not do it |
| Stratification, 2,000 mm tall | 10–15 K top to bottom | put heat sources low |
| Solar gain, outdoor | design to ambient + 10 K | a dark sunlit face at 40 °C air is a 50 °C panel |
| Capacitor life per 10 K | halves | 55 °C internal gives 35% of rated life |
| Finish inside the panel | 60–80 µm powder | seven-tank pretreatment, 18 ft × 14 ft booth |
Sizing the air path
Ventilation removes heat by replacing air that has already been warmed. The flow needed is the loss divided by the volumetric heat capacity of air times the permitted rise. At 40 °C, air has a density of about 1.13 kg/m³ and a specific heat of 1,005 J/kg·K, so 1,133 J/m³·K. For 1,500 W and a 10 K rise:
Q = 1,500 / (1,133 × 10) = 0.132 m³/s, which is 477 m³/h.
Allow a 15 K rise and the figure drops to 318 m³/h. Allow 20 K and it is 239 m³/h, but a 20 K rise on a 45 °C site puts the panel at 65 °C internal, which is a derating conversation with every device supplier. Ten to fifteen kelvin is the practical band.
The openings, not the fan, set the limit
The air has to get in and out through holes in the steel, and the holes are usually the smallest part of the system. At 477 m³/h, or 0.132 m³/s, a single 300 × 300 mm opening with 60% free area gives 0.054 m² and a face velocity of 2.4 m/s. That is acceptable: above about 3 m/s a louvre starts to whistle and to pull in rain and dust rather than letting them drop out.
Two louvres, one low on the door and one high on the opposite side, give a cross-flow path across the devices. Put both on the same face and the air short-circuits across the top of the panel without touching the drive at the bottom. The louvre free area guide covers the blade geometry that gets 60% free area from a pressed louvre without weakening the door, and the canopy louvre design guide applies the same numbers to a weather-facing panel.
The fan has to be chosen at the operating point, not at free air. A 172 mm axial fan that is quoted at 500 m³/h free air delivers about 200 m³/h at 250 Pa, and a dirty G4 filter plus two louvres is a 250 Pa system. That fan will not hold the panel to 10 K. A centrifugal blower of the same physical size, or two axial fans in parallel, will. This is the single most common thermal design error in a fabricated panel: the fan was sized from a catalogue's first column.
When air is not enough
Above about 3 kW in a cabinet this size, or where the ambient is over 45 °C, or where the air is oily, conductive or explosive, an open air path stops being the answer. A closed-loop air-to-air heat exchanger keeps the IP rating and moves heat without letting plant air in; a cooling unit does the same with a refrigerant circuit. Both are bought-in assemblies, and the enclosure's job is to provide a flat, gasketed cut-out to the supplier's template, with the return-air path designed so the cold air reaches the devices rather than the door.
The failure modes that follow a hot panel
Capacitors that age at double speed. An electrolytic capacitor's life halves for every 10 K above its rated temperature. A drive rated for 40 °C running at 55 °C keeps about 35% of its rated life, which turns a ten-year capacitor into a three-and-a-half-year one. The panel does not fail; it starts tripping on DC bus ripple, and the maintenance team replaces drives instead of fixing the airflow.
Breakers that carry less than the label. A thermal-magnetic breaker is calibrated at 30 °C or 40 °C and derates above it. A 63 A breaker in a 50 °C enclosure carries about 57 A, and in a 60 °C enclosure about 52 A. A panel designed at 63 A per feeder and built into a hot corner trips on a normal load, and the trip is blamed on the breaker.
Hot joints on the busbar. A bolted joint at 80 °C oxidises faster than one at 50 °C, and the oxide raises the resistance, which raises the temperature. The runaway is slow and invisible until the joint discolours. Thermal imaging at commissioning and at the first annual shutdown is the only practical check.
Condensation on the cold metal. A ventilated panel in a humid plant pulls moist air in during the day and condenses it on the door at night. A 10–20 W heater per cubic metre, or a filter with a desiccant breather, is cheaper than replacing a corroded mounting plate.
Dust tracking. Covered in the warning above: an unfiltered fan, or a filter that was never changed, deposits conductive dust across clearances that were designed for clean air.
A fan that stopped. A fan with no alarm is a fan that stopped three months ago. Fit a differential pressure switch or a thermostat with a volt-free contact, and wire it to the panel's own alarm.
What the standard actually limits
IEC 61439-1 sets the reference conditions for a low-voltage assembly: an ambient air temperature of 40 °C maximum, an average over 24 hours of 35 °C, and a lower limit of −5 °C, unless the manufacturer declares otherwise. It also limits the temperature rise of accessible parts, with 30 K for metal parts that can be touched, so that a door handle in a 40 °C room is not a burn hazard. Those limits are for the assembled switchgear, declared by the panel builder, not for the empty box.
The empty enclosure is covered by IEC 62208, which includes the thermal behaviour of the shell. IEC 60890 gives the calculation method for verifying the temperature rise of an assembly by calculation rather than by test, and it is the method behind the 5.5 W/m²·K figure used above, with adjustments for the installation.
The practical consequence for a fabricator is a scope boundary. We cut, form, weld, coat and mask the enclosure to the thermal drawing. We do not test the assembly to IEC 61439-1, we do not declare its temperature rise, and we do not size the devices. What we can do is refuse to cut a louvre that is too small for the flow the customer has calculated, which happens more often than it should.
What Ganesh runs for ventilated enclosures
Louvre panels are pierced on the 250 T press with a dedicated tool, which gives a consistent 45° blade and a free area of 55–60% without the distortion a laser-cut slot pattern leaves in a 1.5 mm door. The enclosure itself is nested and cut on the 12 kW fibre laser with the 40 ft × 16 ft bed, formed on the 200 T, 5-axis brake with the 4,450 mm bed, and welded in fixtures built in our own tool room so the door opening stays square while the heat goes in.
Filter openings, fan cut-outs and heat exchanger apertures are cut to the supplier's template and checked against it before coating, because a 2 mm error on a fan cut-out is a gasket that does not seal. Coating runs through seven-tank pretreatment into the 18 ft × 14 ft powder booth at 60–80 µm, with DFT and adhesion recorded per batch. The data centre rack range covers the frames and enclosures that share this route, and the racks and enclosures hub collects the thermal, earthing and ingress guides around it.
What to put on the drawing
- Internal loss in watts, from the device list, with the largest single device called out.
- Site ambient, including whether the panel is in sun, and the acceptable internal temperature.
- Whether the enclosure is passive, ventilated, or fitted with a heat exchanger, and the supplier's cut-out template.
- Louvre size, position and free area, with the inlet low and the outlet high on opposite faces.
- Filter class and the change interval to be marked on the door.
- Fan or blower model, flow at the system pressure, and the alarm contact.
- Device layout: heat sources low, temperature-sensitive devices low, clearance above the busbar.
- Finish, DFT and whether internal faces are coated, plus the IP rating the openings must preserve.
When a fabricated enclosure is not the right choice
A welded steel enclosure with louvres is the right answer for most indoor panels up to about 3 kW. It is the wrong answer where the site is a hazardous area, because the openings and the fan both need certification, and that is an Ex-rated assembly, not a sheet metal box.
It is also wrong where the ambient is above 50 °C or the load is above 5 kW, because the enclosure then becomes a housing for a bought-in cooling unit and the thermal design belongs to that supplier. And it is wrong where the panel has to hold IP66 or better with no air path at all: at that point the only options are a heat exchanger through the wall or a derated device list, and both are decisions the electrical designer makes before the steel is cut.
Send the device list, the losses and the site conditions with the drawing, or put the duty on the RFQ form. The quotation will state the openings, the louvre free area, the filter class and the finish it assumes.
Frequently asked questions
How much heat can a sealed switchgear enclosure dissipate?
About 5.5 W per square metre of effective surface per kelvin above ambient. A 2,000 × 800 × 600 mm cabinet with its back against a wall has 4.96 m², so it sheds roughly 27 W per kelvin, or 1,100 W at a 40 K rise. Above that the internal temperature climbs faster than the load.
How do I calculate the airflow a panel needs?
Divide the internal loss in watts by 1,133 J/m³·K times the permitted rise in kelvin. For 1,500 W and a 10 K rise that is 0.132 m³/s, or 477 m³/h. Then size the openings for under 3 m/s at the louvre face and pick a fan that delivers that flow at the filter's dirty pressure drop.
Is a filtered fan enough for a dusty plant?
Only with a G3 or G4 filter and a maintenance interval written on the panel. An unfiltered fan turns the enclosure into a dust trap, and the failure shows up as tracking across a terminal rather than as an over-temperature. Where the air is oily or conductive, use a closed-loop heat exchanger instead.
Why does the top of my panel run hotter than the bottom?
Stratification. In a natural-convection enclosure 2,000 mm tall, the air at the top runs 10–15 K above the air at the bottom, so a device mounted high sees a higher ambient than the calculation assumes. Put the heat-producing devices low and the temperature-sensitive ones low as well, or force the air upward with a fan.
What ambient should I design a switchgear enclosure for?
40 °C maximum and 35 °C average over 24 hours is the reference in IEC 61439-1, and it is what most component ratings assume. A panel in an Indian plant roof or next to a furnace needs 50 °C or 55 °C, and the devices have to be derated to match.
Sources and further reading
- IEC 61439-1:2020 — Low-voltage switchgear and controlgear assemblies, general rules including ambient and temperature-rise limits · reviewed 3 Nov 2026
- IEC 60890:2022 — Method of temperature-rise verification of low-voltage assemblies by calculation · reviewed 3 Nov 2026
- ISO 12944-2:2017 — Corrosivity of atmospheres, classification of environments · reviewed 3 Nov 2026
Designing a panel that runs hot?
Send the device list with the losses, the ambient, the IP rating and the site conditions. We will confirm the openings, louvre free area and finish before quotation.
Request a quote