Key takeaways
- Steel contracts 4.3 mm over a 6,000 mm panel across a 60 K swing from 35 °C to −25 °C, so fix one edge and slot every other hole by 5 mm or the skin buckles at the joint.
- Water vapour pressure at 30 °C and 70% RH is 2.97 kPa against 0.063 kPa inside a −25 °C room, a 47:1 driving force, so one unsealed penetration grows a frost plug within a season.
- A 50 × 50 mm steel post through 100 mm of polyurethane conducts as much heat as 5.4 m² of panel; break the bridge at the door frame or heat it.
- Hold the threshold heater at 30 W/m and the door frame at 40 W/m, and prove continuity before the panel is closed, because a failed heater shows up as an ice ridge, not an electrical fault.
- Keep the steel skins, trims and frames to ±0.5 mm on feature positions and the assembled opening square to 1.5 mm on the diagonal; the foam panel supplier works to a looser class.
Cold store panel design starts with the steel, because the steel is what fails first. The insulated core does its job quietly for twenty years; what lets a freezer down is a 4 mm gap at a panel joint, a door frame that conducts heat through the insulation, or a fixing that pierces the vapour barrier. All three are drawn, quoted and cut long before anyone notices the frost.
The governing number is thermal movement. Steel moves 12 × 10⁻⁶ per kelvin, so a 6,000 mm panel taken from a 35 °C site temperature to a −25 °C room contracts 12 × 10⁻⁶ × 6,000 × 60 = 4.3 mm. Fix that panel at both ends and something has to give: the skin oil-cans, the joint gasket opens, or the fixing tears the hole.
This guide is about the steel side of a cold store panel design — skins, edge trims, door frames, thresholds and flashings — and the failure modes that come from getting the movement, the vapour seal or the thermal break wrong. It is written for the manufacturing engineer who has to make the panel supplier's drawing and the steel drawing agree.
What fails first in cold store panel design?
Ice on the floor, condensation on the frame and a buckled skin at the joint. Each has one physical cause, and none of them is the insulation thickness.
Ice inside a freezer forms where warm, moist air meets a surface below 0 °C. Air arrives through a leak or diffuses through the panel as vapour. The driving force is large. At 30 °C and 70% relative humidity the vapour pressure is 2.97 kPa. Inside a −25 °C room the saturation pressure is 0.063 kPa, so the ratio across the wall is 47:1. Vapour moves from the warm side to the cold side through any path that is not sealed, and it freezes where it arrives.
That is why the vapour barrier is a continuous system rather than a membrane: the panel joint mastic, the door frame gasket, the penetration seals and the floor membrane all have to overlap. A 3 mm gap in one of them is enough. The frost plug that grows behind a cable penetration can lift a panel off its joint in one season.
Buckling is simpler. A panel restrained at both ends in a 60 K swing has 4.3 mm of contraction to absorb. If the fixing holes are round and the joint is tight, the contraction becomes compressive stress and the skin deflects out of plane. The deflection is small at first — 2–3 mm across a metre — and then the joint gasket loses contact, and the leak starts.
Condensation on the outside of the frame is the third. The outer skin of a well-insulated panel stays close to room temperature. A steel door frame that passes through the insulation does not. Its outer end runs at the cold-room temperature, the room air condenses on it, and in a freezer it frosts. Ice on the warm side of a door frame is the classic signature of a missing thermal break.
Panel thickness, movement and the steel limits
The middle column is the specification for a −25 °C freezer; the right column is what the sheet metal route can hold.
| Parameter | Typical value for a −25 °C freezer | Ganesh limit or note |
|---|---|---|
| Panel thickness | 150 mm PU at 0.023 W/mK | Skins and trims cut on the 12 kW laser, 40 ft × 16 ft bed |
| Panel U-value | 0.15 W/m²K | Calculated to ISO 6946 by the panel supplier, not by us |
| Thermal movement, 6,000 mm steel | 4.3 mm over 60 K | Slot 5 mm at every non-fixed fixing |
| Thermal movement, 6,000 mm aluminium trim | 8.3 mm over 60 K | Slot 10 mm; aluminium trim needs its own allowance |
| Fixing pitch on trim | 300 mm | 4.8 mm rivet or M5 screw into a closed section |
| Door frame thermal break | 10 mm rigid PVC or equivalent | Frame fabricated square to 1.5 mm on the diagonal |
| Threshold heater | 30 W/m under the door line | Continuity proved before the panel closes |
| Door frame heater | 40 W/m on the frame rebate | Separate circuit from the threshold |
| Door leaf | 1,200 × 2,100 mm, 100 mm thick, about 70 kg | Three hinges; hinge plate in 3.0 mm, not 1.6 mm |
| Door gasket | Magnetic, 30% compression | Checked with a feeler gauge at four points |
| Skin tolerance | ±0.5 mm on feature positions | Tighter than the panel supplier's class |
| Bend radius, 1.2 mm stainless | 1.5 × t minimum | Visible faces 2.0 mm radius to avoid marking |
Two rows carry most of the cost. The door frame thermal break is a machined or pultruded component that has to be inserted into the frame profile, and it is the first thing a value-engineered quote removes. The heater is the second: 30 W/m of self-regulating cable around a 1,200 mm threshold is a small line item that prevents a large service call.
Five failure modes, in the order they appear
Panel buckling at a restrained joint
A 6,000 mm skin fixed at both ends in a 60 K swing has nowhere to put 4.3 mm. The skin buckles at the panel mid-height, the joint opens 2–3 mm, and the gasket that was designed for 30% compression now sits at zero.
Fix one edge only. Every other hole is a 5 mm slot, with a washer large enough to cover the slot, and the screw is tightened to just past snug rather than torqued. On a vertical panel the fixed edge is the top, so the panel hangs and contracts downward without fighting the floor.
Vapour seal breached at a penetration
Every pipe, cable and bracket that crosses the envelope is a hole in the barrier. The vapour pressure ratio of 47:1 does the rest. The failure is not the hole; it is the sealing method. Silicone alone does not bond to a dusty or damp surface, and it does not survive the movement at a pipe that runs from +35 °C to −25 °C.
Use a butyl collar bonded to the barrier, clamped by a stainless flange, with the pipe sleeved so it never touches the skin. On a 25 mm pipe the collar is 75 mm diameter, giving a 25 mm bond width all round. That is cheap and it is the detail most often drawn as "seal with mastic".
Thermal bridge at the door frame
A steel frame passing through 150 mm of insulation is a short circuit around the panel. Compute the penalty. A 50 × 50 mm steel post through 150 mm has a conductance of 50 × 0.0025 / 0.15 = 0.83 W/K. The panel around it, at 0.15 W/m²K, carries 0.15 W/K per square metre. So one small post equals 5.5 m² of good panel — and that is before the fixing plates.
Break the bridge with a 10 mm PVC or polyamide insert at the frame rebate, or heat the frame. Heating is the fallback for existing rooms; a break is the design answer. On a sliding door, the track bracket is the worst offender because it is a direct steel path from the warm floor to the cold frame.
Gasket freeze and door seal loss
A freezer door gasket compresses against a frame that is held above freezing by the frame heater. If the heater fails or the frame is unheated, moisture on the gasket freezes, the gasket takes a set, and the door leaks air from the first cold snap.
Specify the gasket at 30% compression, check it at four points with a feeler gauge and a strip of paper, and keep the heater circuit independent of the door interlock so it stays on when the door is closed and the room is at temperature.
Floor heave under the threshold
Frost heave lifts a floor slab when moisture below the insulation freezes. The visible symptom is a door that will not close, or a threshold that has risen 20 mm above the finished floor. The cause is a failed or missing heater under the floor, or a floor membrane that was punctured during fit-out.
The heater runs at 30 W/m under the door line and typically 10–20 W/m² across the door approach. It has to be proved continuous and megger-tested before the floor screed goes down, and recorded, because after that it is inaccessible.
The stainless and galvanised details
Cold rooms are washed down. Cleaning agents are alkaline or chlorinated, and both attack galvanised steel at a cut edge first. A galvanised trim with a laser-cut edge shows white rust at the edge within a few months in a wet room.
Three rules keep the steelwork alive. Laser-cut edges get a deburr and a coating that wraps the edge, or the trim is stainless. Fasteners are A2 or A4 stainless, not zinc-plated, and where stainless meets galvanised steel the joint gets an isolating washer to slow the galvanic couple. And any horizontal surface inside the room is a shelf for ice — a 20 mm upstand beats a 200 mm flat ledge every time.
The powder coating failure guide explains the edge-coverage mechanism in detail; it applies to a cold room as much as to an outdoor canopy, because both are wet environments with cut edges.
What Ganesh actually runs
Skins, trims, door frames and thresholds are cut on the 12 kW fibre laser with the 40 ft × 16 ft bed, which takes a 6,000 mm trim in one pass and leaves a square edge for the fold. Long returns run on the 200 T, 5-axis brake with the 4,450 mm bed and Delem control, which forms a full-length upstand without a mid-panel join. Door frames are welded on the manual stations or the CLOOS robotic cells, depending on volume, on fixtures built in our tool room so the diagonal stays within 1.5 mm. Coating is seven-tank pretreatment and 60–80 µm powder where the drawing calls for it, with DFT and cross-hatch checks per batch.
We do not laminate the foam core. The panel, its joint geometry and its U-value come from a panel supplier, and our scope is the steel that meets it — frames, trims, flashings and thresholds. That split has to be settled on the drawing, because a 2 mm difference in joint geometry between the two suppliers becomes a 2 mm gap in the vapour barrier.
What to put on the drawing
- Room temperature and the warm-side design condition, since they set the movement and the dew point.
- Panel thickness, joint geometry and the panel supplier's edge profile.
- The fixed edge and the direction of movement, with slot size and washer type.
- Door frame break material, thickness and position.
- Heater wattage per metre, circuit, and the continuity test record required before closing.
- Gasket type, compression range and the test method.
- Coating system and which faces are food-contact or washdown.
- Tolerance split: ±0.5 mm on steel features, 1.5 mm on the frame diagonal, and ISO 2768-1 class m for anything not called out.
When sheet metal is the wrong scope for a cold store
If the project is a full envelope, the panel system decides the performance and the steel is a component of it. A room below −40 °C usually needs a different panel system and a more aggressive vapour strategy than a steel trim can rescue. And a cold store built inside an existing building may need structural work — a raised floor, a separate roof — that is outside a fabrication shop's scope.
Within those limits, the steelwork is where the failures are prevented. The railway and infrastructure hub groups the related guides on frames, panels and coating, and the long panel bending guide covers the forming side of a 6,000 mm trim. A freezer that stays dry for ten years is not a more expensive freezer; it is the same freezer with four details drawn correctly — a fixed top edge, 5 mm slots, a 10 mm break and 30 W/m under the door. The skins, trims, frames and thresholds are listed under custom sheet metal parts, the forming limits sit on the bending capability page, and the drawings can be sent through request a quote with the room temperature and panel thickness so we can check the movement allowance before pricing.
Frequently asked questions
How much does a cold store panel move when it cools down?
Steel expands and contracts at 12 × 10⁻⁶ per K. Over a 6,000 mm panel and a 60 K swing from 35 °C to −25 °C the change is 12 × 10⁻⁶ × 6,000 × 60 = 4.3 mm. Aluminium at the same size moves 8.3 mm, which is why aluminium trim needs a bigger slot.
Why is there ice at the bottom of my cold room door?
Either the threshold heater has failed or the frame is a thermal bridge with no break. The surface falls below freezing, moisture in the room condenses on it and freezes. Check heater continuity first, then the frame detail.
What thickness of panel does a −25 °C freezer need?
150 mm of polyurethane at 0.023 W/mK gives 0.15 W/m²K, which is the usual specification for −25 °C. A 100 mm panel gives 0.23 W/m²K and is fine for chill rooms at 0 to 5 °C, not for a freezer.
Can powder coating be used inside a cold store?
On the warm side, yes, at 60–80 µm over seven-tank pretreatment. On the cold side use a food-safe washdown finish agreed with the operator, and never rely on coating for the vapour seal — the seal is the gasket and the mastic, not the film.
Do you make the insulated panels themselves?
No. We make the steel skins, edge trims, door frames, thresholds, bump rails and flashings to drawing. The foam core and the panel lamination come from a panel supplier, and the two sets of drawings have to agree on joint geometry.
Sources and further reading
- ISO 6946 — Building components and elements: thermal resistance and transmittance · reviewed 3 Nov 2026
- IS 875 (Part 3) — Wind loads on buildings and structures · reviewed 3 Nov 2026
- ISO 2768-1 — General tolerances for linear and angular dimensions · reviewed 3 Nov 2026
Have cold store steelwork to quote?
Send the door frame, threshold, trim and flashing drawings with the room temperature, panel thickness and quantity. We will check the movement allowance and the thermal break before quotation.
Request a quote
