Steel Silo Preheating Welding Practices: What Actually Happens on Site

Preheating before welding silo steel isn't optional — it's the difference between a shell that lasts 25 years and one that cracks within 6 months. The short answer: preheat to slow the cooling rate, drive off moisture, and prevent hydrogen-induced cracking in the heat-affected zone. For most silo shell steels (S235 through S355 and beyond), you're looking at preheat temperatures between 75°C and 200°C depending on carbon equivalent, thickness, and ambient conditions. Here's what the standards say, what the field demands, and where most crews get it wrong.

Why Preheating Isn't Just a Line Item on the WPS

The Metallurgy Nobody Explains at the Kickoff Meeting

When you weld two plates together, the weld metal and the surrounding base metal go through a thermal cycle that peaks above 1,500°C and then cools — fast. In thin plate, that cooling rate might be 30°C per second. In a 20mm silo shell course, it's slower, but still fast enough to produce martensite in steels with carbon equivalents above about 0.40%.

Martensite is hard. It's brittle. And when it forms in the heat-affected zone (HAZ) under restraint — which a silo shell absolutely provides, given the circumferential and longitudinal stresses — it cracks. Not always immediately. Sometimes it waits three weeks. Sometimes it waits until the first filling cycle when the wall goes from zero to 150 kPa of hoop stress.

I remember a project in Vietnam where the contractor skipped preheat on the bottom shell course. "It's only 12mm," they said. The carbon equivalent was 0.47%. We found transverse HAZ cracks in 8 of 24 circumferential welds during UT inspection. Every single one had to be cut out and rewound. That cost them 11 days and about $18,000 in rework. The preheat would have cost maybe $400 in propane.

Hydrogen: The Invisible Crack Driver

Preheating does two things. First, it slows cooling — giving the microstructure time to transform to ferrite and pearlite instead of martensite. Second, it drives hydrogen out of the weld zone before it can diffuse into the HAZ.

Where does the hydrogen come from? Moisture on the plate surface. Humidity in the air. Cellulose electrodes (E6010, E7010) that generate hydrogen-rich shielding gas. Even the mill scale on hot-rolled plate can trap enough moisture to push diffusible hydrogen levels above 15 mL per 100g of weld metal. Once that hydrogen sits in a martensitic HAZ under stress, you've got hydrogen-induced cracking (HIC). Also called cold cracking. Also called delayed cracking. Same problem, three names.

The rule of thumb: for every 50°C of preheat above ambient, you roughly double the time hydrogen has to escape before the weld cools below the critical temperature (around 100–150°C for most structural steels). That's not a precise number — it depends on the steel, the hydrogen content, and the restraint — but it's the right order of magnitude.

How to Actually Calculate Preheat Temperature

Carbon Equivalent: The Starting Point

You can't pick a preheat temperature out of thin air. You start with the carbon equivalent (CE) of the steel. The IIW (International Institute of Welding) formula is the one most welding engineers use:

CE(IIW) = C + Mn/6 + (Cr + Mo + V)/5 + (Ni + Cu)/15

For a typical S355J2 silo shell plate, you might see CE values around 0.38–0.43%. For S420ML or higher-strength steels used in large-diameter silos, CE can hit 0.48–0.52%. The higher the CE, the more preheat you need.

EN 1011-2 gives preheat tables based on CE, thickness, and hydrogen class. For a steel with CE = 0.45%, welded with a low-hydrogen process (below 5 mL/100g), at 25mm thickness, the recommended preheat is about 100°C. Bump that to CE = 0.50% and you're at 150°C. Use a high-hydrogen process and add another 50°C on top.

AWS D1.1 takes a different approach — it bases preheat on the steel grade and thickness directly. For ASTM A572 Grade 50 (roughly equivalent to S355), thicknesses above 38mm require a minimum preheat of 66°C. But that's a minimum. In practice, with high restraint and high humidity, you'll want more.

Thickness and Restraint: The Multipliers Nobody Checks

Here's where most welding procedure specifications fall short. They calculate preheat based on CE and plate thickness, then stop. They don't account for restraint.

A silo shell course being welded in place has enormous restraint. The bottom course is anchored to the ring beam or foundation. The circumferential weld joining two shell courses is constrained by the entire weight of the structure above it. Longitudinal welds in the shell are constrained by the hoop stiffness of the already-welded courses.

EN 1011-2 addresses this indirectly through the "rigidity" factor, but in practice, I add 25–50°C to the calculated preheat for silo shell welds compared to what I'd use for the same steel in a less restrained structure. It's not in the standard. It's field experience.

Thickness matters more than people think. A 30mm shell course doesn't just need more preheat because there's more material — it needs more because the weld is cooling faster. Thicker plate acts as a heat sink. The thermal gradient is steeper. The HAZ is narrower but the cooling rate is higher. For thicknesses above 40mm, I've seen preheat requirements jump by 50–75°C compared to the same steel at 20mm.

Ambient Temperature and Wind: The Site Reality

The standards assume an ambient temperature of around 20°C. On a site in northern China in January, you might be at -15°C. That's a 35°C swing that directly affects your preheat requirement.

Wind is worse. A 5 m/s breeze across a preheated weld zone can drop the surface temperature by 30–40°C in under two minutes. I've seen crews in Inner Mongolia preheat to 150°C, confirm it with a tempil stick, then start welding — and by the time they're 200mm into the joint, the plate has cooled to 80°C because the wind picked up.

The fix: wind breaks. Simple canvas or plywood barriers on the upwind side. They don't need to be fancy. On one project in Kazakhstan, we used 2m-high plywood sheets on sawhorses. Cost about $200 in materials. Eliminated every wind-related preheat complaint for the rest of the winter.

Preheat Application: Methods That Work and Methods That Don't

Propane Torches: The Standard Approach

Multi-flame propane heating torches are the workhorse of silo welding preheat. A typical setup uses a 2-meter-long manifold with 8–12 flame outlets, connected to a standard propane cylinder or a manifold of cylinders for continuous work.

The key is even heating. You don't want a hot spot directly under the torch and cold plate 100mm away. Move the torch in a slow, sweeping pattern across the preheat zone. Keep the flame tips about 50–75mm from the plate surface. Too close and you get localized overheating — surface temperatures of 400°C+ while the back of the plate is still at 80°C.

The preheat zone width matters. EN 1011-2 specifies a minimum of 75mm on each side of the weld groove, or three times the plate thickness — whichever is greater. For a 20mm shell course, that's 75mm minimum. For a 40mm course, it's 120mm. I go wider than the minimum. 100mm on each side for anything under 25mm, 150mm above that. The extra width gives you a thermal buffer so the weld zone doesn't cool too fast while you're still laying beads.

Electric Heating Pads: When Precision Matters

Ceramic-mat electric heaters give you controlled, uniform preheat. They're slower — you're looking at 30–60 minutes to bring a 20mm plate up to 150°C versus 10–15 minutes with a torch — but the temperature is even across the entire zone.

I use electric pads for two situations: high-strength steels (S420 and above) where the preheat window is narrow, and repair welds where you can't afford to get it wrong. On a 5,000-tonne clinker silo in Indonesia, we had to repair a 300mm-long crack in a 32mm shell course. The steel was S420ML, CE = 0.49%. We used electric pads to hold 175°C ± 10°C for the entire repair. Took four hours. The repair passed UT on the first attempt.

Cost-wise, a set of ceramic mat heaters for a 2-meter weld joint runs about $800–$1,200. That's a one-time investment if you maintain them. For a project with hundreds of shell welds, the time savings from not having to re-check and re-heat with torches can be significant.

Induction Heating: The Niche Solution

Induction coils give you the fastest, most controllable preheat. They're common in pipe fabrication shops but rare on silo construction sites. The equipment is expensive — a portable induction system capable of preheating silo shell plate runs $15,000–$25,000 — and it requires a power source that many sites don't have.

Where I've seen it work: large-diameter silos (30m+) with thick lower shell courses (40–60mm) where the preheat requirement is 200°C+ and propane torches can't keep up with the heat sink effect. On a 40,000-tonne coal silo in South Africa, the bottom course was 55mm ASTM A516 Grade 70. Preheat requirement was 225°C. We used induction coils. The alternative — six propane torches running simultaneously — would have consumed about 40 kg of propane per hour and still struggled to maintain temperature across the full joint width.

Measuring Preheat: Trust Nothing You Haven't Verified

Temperature-Indicating Crayons (Tempil Sticks)

Tempil sticks are the most common field tool. You pick a stick rated for your target temperature (say, 105°C), draw a mark on the preheated plate, and wait for it to melt. When it liquefies, you've hit the target temperature.

They're cheap — about $3–$5 per stick — and they work. But they have limitations. They measure surface temperature at a single point. They don't tell you if the back of the plate is at the same temperature. And they're single-use: once the mark melts, you need a new stick for the next check.

I carry tempil sticks on every site visit. But I don't rely on them alone. For critical welds — shell-to-bottom-plate joints, ring beam connections, any weld in steel above S355 — I want a contact thermocouple or an infrared thermometer as a cross-check.

Infrared Thermometers: Fast but Tricky

IR thermometers give you instant readings across a wide area. Point, shoot, read. The problem is emissivity. Mill-scale steel has an emissivity of about 0.80–0.85. Bright, clean steel is closer to 0.30–0.40. If your IR thermometer is calibrated for one surface condition and you measure another, you can be off by 30–50°C.

The fix: calibrate against a contact measurement. Take a reading with a surface thermocouple, then adjust the IR thermometer's emissivity setting until it matches. Do this at the start of each shift, because the surface condition changes as the plate heats up and the mill scale oxidizes.

Thermocouples: The Gold Standard

A Type K thermocouple spot-welded to the plate gives you continuous, accurate temperature monitoring. For critical welds, I specify thermocouples at three locations: the center of the weld zone, and at the edge of the preheat zone on each side. This tells you not just whether you've hit the target temperature, but whether the gradient across the preheat zone is acceptable.

The gradient matters more than most people realize. If the center of the joint is at 150°C but the edge of the preheat zone is at 60°C, you've got a 90°C gradient across 100mm. That's steep enough to create differential thermal expansion stresses that can distort thin shell plate. For shell courses under 16mm, I try to keep the gradient below 50°C across the preheat zone.

Welding Procedure Specifications: Writing What You'll Actually Enforce

The Preheat Clause That Gets Ignored

Every WPS for silo shell welding should specify: minimum preheat temperature, maximum interpass temperature, preheat zone width, measurement method, and ambient conditions under which welding must stop. Most WPS documents I see specify the first two and ignore the rest.

The interpass temperature is just as important as the preheat. If you let the weld zone cool below the minimum preheat between passes, you've lost the benefit. If you let it exceed the maximum interpass (typically 250–300°C for structural steels), you risk grain growth in the HAZ that reduces toughness. I've seen WPS documents with no interpass limit at all. That's a gap.

Here's a clause I put in every silo WPS: "Welding shall not commence when the ambient temperature is below 5°C unless preheat is increased by 25°C above the specified minimum. Welding shall not commence when wind speed exceeds 8 m/s at the weld location unless wind protection is in place. Preheat shall be verified at the start of each weld pass and at intervals not exceeding 500mm of weld length."

That last part — verifying every 500mm — is what catches the problem before it becomes a defect. On a 10-meter circumferential weld, that's 20 temperature checks. Takes an extra 3 minutes. Worth it.

Qualification Testing: Does Your WPS Match Your Steel?

Welding procedure qualification tests (WPQR) should be performed with the actual steel grade and thickness range you'll be using on the silo. Too often, a contractor qualifies a WPS with 16mm S235 plate and then applies it to 30mm S355 plate without re-qualification. The preheat that worked for the qualification test may be completely inadequate for the thicker, higher-CE steel.

EN ISO 15614-1 requires that the qualification test covers the thickness range of the production welds. A test on 25mm plate qualifies production welds from 5mm to 50mm (for butt welds). But the preheat specified in the qualified WPS was based on 25mm. If you're welding 50mm plate with the same preheat, you're under-preheating.

The solution: either qualify at the maximum thickness you'll use, or write the WPS with thickness-dependent preheat values. The second approach is more practical for silo work, where shell thickness varies from 6mm at the top to 40mm+ at the bottom.

Case Study: 8,000-Tonne Cement Silo, Eastern Turkey

Project: 32m diameter cement silo, 8,000-tonne capacity, shell height 28m. Shell steel: S355J2, thickness ranging from 8mm (top course) to 36mm (bottom course). Carbon equivalent: 0.41–0.44% across all courses. Welding processes: GMAW (CO₂ shielding) for longitudinal and circumferential shell welds, SMAW for attachments and manways.

Site conditions: elevation 1,400m, ambient temperature range -8°C to +12°C during construction (October–December). Relative humidity averaging 75%. Wind speeds frequently exceeding 10 m/s in the afternoon.

Preheat specification: 75°C minimum for courses 16mm and below. 100°C for 20mm courses. 125°C for 25mm courses. 150°C for 30mm and 36mm courses. All values measured on the inner shell surface, 75mm from the weld groove edge. Wind barriers required for all welding when wind speed exceeded 6 m/s.

What happened: The first two shell courses (36mm and 32mm) were erected in late October. Ambient temperature was 3–8°C. The contractor used propane torches but didn't install wind barriers. Preheat checks with tempil sticks showed 150°C at the torch side, but IR measurements on the opposite surface showed 90–110°C. The welds passed visual inspection.

Two weeks later, UT inspection of the circumferential welds in the bottom course found 14 indications — 9 of them confirmed as HAZ cracks. All were in welds done on days with wind speeds above 8 m/s. The cracks were 15–40mm long, located 3–8mm from the fusion line. Classic hydrogen-induced cold cracking.

Rework: All 9 cracked welds were gouged out, preheat was reapplied with wind barriers in place, and the joints were rewelded. The rework took 6 days and cost approximately $12,000. The wind barriers cost $350 in materials. The lesson was expensive but clear: preheat without wind protection in a high-wind environment is just going through the motions.

For the remaining courses, we enforced the wind barrier requirement strictly. No UT indications were found in any subsequent welds. The silo was commissioned in February and has been in continuous operation since.

FAQ

Q: Can I skip preheat if I use low-hydrogen electrodes?

A: No. Low-hydrogen electrodes (E7018, E7016) reduce the hydrogen input, which lowers the preheat requirement — but they don't eliminate it. EN 1011-2 still specifies minimum preheat for low-hydrogen processes, typically 25–50°C less than for cellulose electrodes. For a steel with CE = 0.45% at 25mm thickness, you might drop from 100°C to 75°C by switching from E6010 to E7018. But 75°C isn't zero. And if the steel is wet or the humidity is above 80%, even low-hydrogen electrodes can't compensate for the moisture on the plate surface.

Q: How long do I need to hold preheat before starting the weld?

A: Long enough for the entire preheat zone to reach the target temperature through the full plate thickness. For a 20mm plate heated with propane torches, that's typically 5–8 minutes after the surface hits temperature. For 40mm plate, 12–15 minutes. The standard doesn't specify a hold time — it specifies a temperature. But if you start welding while the center of the plate thickness is still 30°C below the surface temperature, you're effectively under-preheating. Use a contact thermocouple on the opposite surface to verify through-thickness temperature.

Q: What's the maximum interpass temperature for silo shell welding?

A: For S235 through S355 steels, 250°C is the typical maximum. For higher-strength steels (S420, S460), some specifications limit interpass to 200°C or even 180°C to prevent excessive grain growth in the HAZ. Exceeding the maximum interpass doesn't cause cracking the way under-preheating does — it reduces toughness and can lead to brittle fracture under load. In a silo, where the shell is subjected to cyclic loading from filling and discharging, HAZ toughness matters. I've seen WPS documents with no interpass limit. That's a gap you should close.