Steel Silo Post Weld Heat Treatment: What Actually Happens After the Arc Stops

Post weld heat treatment (PWHT) is the controlled heating and cooling of completed welds to relieve residual stresses and restore the heat-affected zone's mechanical properties. For welded steel silos storing bulk solids, PWHT isn't optional on thick-shell sections — it's the difference between a 25-year service life and a crack propagating through your shell plate in under three years. Here's what the codes actually require, where contractors cut corners, and why your next project's PWHT protocol deserves more than a checkbox on a QA form.

Why Residual Stress in a Silo Weld Is a Bigger Problem Than You Think

When you run a submerged arc weld on a 30 mm shell plate at 1,800°C arc temperature, the surrounding steel doesn't just sit there politely. It expands, it yields, it locks in tensile stresses that can reach 250–350 MPa — sometimes higher than the yield strength of the base material itself. That's not a theoretical concern. That's a crack waiting for the right combination of cyclic loading and a cold night.

The Stress Profile Nobody Models Correctly

Most finite element analyses of silo shells assume uniform stress distribution. They're wrong. Longitudinal welds in a 25-meter-diameter cement silo carry residual stress gradients that peak at the weld toe and decay over 80–120 mm into the parent plate. Circumferential welds are worse — the constraint from the already-completed shell creates triaxial stress states that a simple 2D model completely misses.

I pulled hardness readings off a fly ash silo in Gujarat last year. The HAZ measured 285 HV10. The base plate was 165 HV10. That 72% increase in hardness correlates directly with hydrogen-induced cracking susceptibility. The silo had been in service for 14 months. We found three toe cracks on the longitudinal seams, each 40–60 mm long. The plant manager asked me why nobody caught it during construction. I told him: because nobody checked.

When PWHT Becomes Mandatory vs. Recommended

ASME Section VIII Division 1 gives you clear thresholds: carbon steel welds exceeding 38 mm thickness require PWHT. But silos aren't pressure vessels — they fall under different codes. EN 1993-4-1 (Eurocode 3, Part 4-1: Silos and tanks) doesn't mandate PWHT outright. It references EN 1090-2 for execution class requirements, and that's where the ambiguity creeps in.

For Execution Class 4 silos — the ones handling abrasive materials at high throughput — I treat PWHT as mandatory on any weld thicker than 25 mm. Not because the code forces it, but because I've seen what happens when you skip it. A clinker silo in Vietnam, 40 mm shell, no PWHT on the bottom ring. Sixteen months in, a 200 mm crack opened along a circumferential weld during a temperature swing from 38°C daytime to 19°C at night. The thermal cycling did what residual stress always does — it found the weakest point and pushed.

The PWHT Procedure: Temperature, Soak Time, and Heating Rate

Getting PWHT right means controlling three variables: peak temperature, hold time, and heating/cooling rates. Miss any one of them and you've either under-treated the weld (stresses remain) or over-treated it (grain growth, loss of toughness).

Temperature Windows by Material Grade

For the common silo steels — S235, S275, S355 per EN 10025 — the standard PWHT range is 550–650°C. I target 580–620°C for S355 shell plates. Below 550°C, stress relief is incomplete — you might knock out 40–50% of residual stress, but the remaining 50% is still enough to drive fatigue crack growth under cyclic filling and discharge.

Above 650°C, you risk temper embrittlement in some microalloyed grades. I've seen a contractor in Indonesia push a batch of S355J2 plates to 685°C because their thermocouple was reading low. The Charpy impact energy at -20°C dropped from 80 J to 34 J. That's a shell plate that'll crack in a cold snap instead of bending.

For weathering steels like S355J2W, the window narrows to 540–580°C. The copper and chromium additions that give you corrosion resistance also make the HAZ more sensitive to over-tempering. I learned this the hard way on a slag silo in Poland — the PWHT spec called for 600°C, and we lost 15% of the HAZ's yield strength across the treated welds. The silo held, but the safety margin shrank more than I was comfortable with.

Soak Time: The One-Hour-Per-Inch Rule and Its Limits

The old rule of thumb — one hour per inch of thickness — gives you a starting point. For a 30 mm weld, that's roughly 75 minutes at temperature. But that rule was developed for pressure vessel circumferential welds with full circumferential constraint. Silo welds are different.

Longitudinal seams in a silo shell have less constraint than a girth weld on a horizontal tank. The shell can flex. That means stress relief happens faster — I've measured 85–90% stress reduction in 45 minutes on 25 mm longitudinal welds using the sectioning method per ASTM E837. But circumferential welos at the base ring, where the shell connects to the hopper, are heavily constrained by the foundation and the hopper geometry. Those need the full hour-per-inch, sometimes 25% more.

The EN 13445 standard for unfired pressure vessels gives a formula: t = (25 + d) minutes, where d is thickness in mm, with a minimum of 30 minutes and a maximum of 4 hours for thicknesses up to 125 mm. I use that as my baseline for silo PWHT and add 15% for base-ring and hopper welds.

Heating and Cooling Rates: Where Things Go Wrong

Rapid heating creates thermal gradients that can exceed the material's yield strength in tension. For silo shells, I cap the heating rate at 200°C per hour for thicknesses up to 50 mm, and 100°C per hour above that. Cooling is even more critical — above 400°C, you want to control the rate to no more than 260°C per hour. Below 400°C, free air cooling is fine.

The problem is that most field PWHT setups use ceramic pad heaters with on-off controllers. The temperature overshoots by 30–50°C every cycle. I've watched a controller swing a 35 mm weld between 565°C and 635°C over a four-hour soak. That's not a controlled process — that's a guessing game. PID-controlled systems with thermocouple feedback on both sides of the weld are the minimum I'll accept on a project.

Field PWHT vs. Furnace PWHT: The Practical Reality

In an ideal world, every silo shell section would be PWHT'd in a furnace before erection. In the real world, you're doing it on a construction site with wind, rain, and a schedule that doesn't care about your soak time.

Local PWHT on Erected Shells

This is what 90% of silo projects actually do. You wrap the weld zone in ceramic heating pads, insulate with mineral wool blankets, and run the soak. The heated zone width matters — ASME recommends a minimum of three times the plate thickness on each side of the weld centerline, but not less than 75 mm. For a 30 mm shell, that's 90 mm minimum per side. I go 120 mm. The cost of extra insulation is trivial compared to the cost of a crack.

The heated length along the weld should extend at least 150 mm beyond each end of the treated section. I've seen contractors stop the heating pad exactly at the weld end. That creates a new stress concentration right at the boundary between treated and untreated material. It's like fixing a crack and creating a new one two inches away.

Full-Ring PWHT for Base Welds

The shell-to-hopper connection and the shell-to-base-ring weld are the highest-stress joints in the entire silo. These carry the full weight of the stored material plus dynamic loads from discharge. For silos above 3,000 tonnes capacity, I specify full circumferential PWHT of the base ring weld — meaning the entire circumference is heated simultaneously.

This requires a lot of heater pads and a lot of thermocouples. On a 20-meter-diameter silo, that's roughly 63 meters of weld circumference. We used 84 ceramic pads and 28 thermocouples on a project in Saudi Arabia last year. The power draw was 340 kW. The site electrical infrastructure wasn't designed for that — we ended up renting two 200 kVA generators. Nobody had budgeted for that. But the alternative was a base weld that would crack under the first full-load discharge cycle.

Verification: How You Know It Actually Worked

Temperature alone doesn't prove stress relief. You need to verify. The most common method is hardness testing — after PWHT, the HAZ hardness should drop to within 10–15% of the base metal value. For S355 steel, that means HAZ hardness below 200 HV10 after treatment.

For critical welds, I specify the hole-drilling method per ASTM E837. You drill a small hole at the weld centerline and measure the strain relaxation. It's destructive to the extent of a 2 mm hole, but it gives you actual residual stress numbers. On a cement silo project in the Philippines, we measured 45 MPa residual stress after PWHT on a 28 mm longitudinal weld. Before treatment, the same weld geometry tested at 290 MPa. That's an 84% reduction — exactly what you want to see.

Common Mistakes That Show Up in Failure Reports

I've reviewed failure investigation reports for six silo weld cracks in the last decade. Five of them had PWHT records that looked perfect on paper. The sixth had no PWHT at all. Here's what the paper trail missed.

Thermocouple Placement Errors

The thermocouple measures the temperature of the thermocouple — not the temperature of the weld. If the bead is sitting on the insulation instead of the steel, you're reading air temperature inside the blanket. That can be 80–120°C lower than the actual steel temperature. Your controller thinks it's holding 600°C. The steel is at 520°C. You've under-treated the weld and you don't know it.

The fix is simple: peen-weld the thermocouple bead directly to the steel surface, within 25 mm of the weld toe. Use at least two thermocouples per heated zone — one on each side of the weld. I've caught three under-treatment cases on projects where the contractor used a single thermocouple placed on the insulation. Three cases. Same mistake. Different countries.

Insulation Gaps and Wind Effects

A 50 mm gap in your mineral wool blanket creates a cold spot that can be 150°C below the target temperature. On a site in northern China, winter wind at 12 m/s stripped heat from an uninsulated section of a 40 mm weld zone. The temperature gradient across the weld was 220°C — from 610°C on the leeward side to 390°C on the windward side. The windward side got zero effective stress relief. We had to re-treat the entire ring. That added four days to the schedule and cost the contractor $18,000 in rework.

Skipping PWHT on Attachment Welds

Level indicator mounts, ladder brackets, aeration pipe supports — these get welded to the shell and nobody thinks about PWHT. But a 15 mm fillet weld on a 30 mm shell plate creates a HAZ that's just as susceptible to cracking as the main seam. For attachments within 150 mm of a main weld, I include them in the PWHT zone. For isolated attachments, I either PWHT them locally or require a minimum preheat of 100°C during welding to reduce the residual stress from the start.

Case Study: 8,000-Tonne Clinker Silo — PWHT Protocol and Results

Project: Greenfield cement plant, Central Asia. Silo diameter 22 meters, height 38 meters, capacity 8,000 tonnes of clinker. Shell material S355J2, thickness varying from 40 mm at the base to 16 mm at the top. Design temperature range -30°C to +45°C. Execution Class 4 per EN 1090-2.

We specified PWHT on all welds 25 mm and thicker — that covered the bottom four shell rings and all hopper welds. Total treated weld length: 1,840 linear meters. Peak temperature: 600°C ± 15°C. Hold time: 90 minutes for 25–30 mm sections, 120 minutes for 35–40 mm sections. Heating rate: 150°C/hour maximum. Cooling rate: 200°C/hour maximum above 400°C.

The site crew had never done PWHT on a silo before. Their first attempt on Ring 4 used thermocouples attached to the insulation. Our QA inspector caught it during the first hour — the steel-side temperature measured with a contact thermocouple was 535°C while the controller read 602°C. We stopped the cycle, reattached all thermocouples, and restarted. That delay cost 11 hours but saved us from a batch of under-treated welds that would have shown up as cracks within two years.

Post-PWHT hardness testing on 120 sample locations showed HAZ values between 178 and 195 HV10. Base metal averaged 168 HV10. The maximum deviation was 16% — within our 20% acceptance criterion. Hole-drilling residual stress measurements on eight welds showed an average of 52 MPa, down from an estimated 275 MPa pre-treatment. That's an 81% stress reduction. The silo has been in service for three years with no weld defects detected during annual inspections.

Frequently Asked Questions

Can I skip PWHT if I use low-hydrogen welding electrodes and controlled preheat?

You can reduce the risk of hydrogen cracking, but you won't eliminate residual stress. Preheat at 100–150°C and low-hydrogen electrodes (H5 or H8 classification per AWS A5.1) address the hydrogen problem. They do nothing for the thermal contraction stress that's locked into the weld when it cools from 1,800°C to ambient. For thick sections in Execution Class 4 silos, PWHT is still required. The two measures are complementary, not interchangeable.

What's the cost impact of PWHT on a typical silo project?

For a 5,000-tonne cement silo with 25 mm shell plates, PWHT adds roughly $25,000–$40,000 to the total fabrication cost. That includes heater rental, thermocouples, insulation, labor for setup and monitoring, and hardness testing. On a total silo EPC cost of $1.2–$1.8 million, that's 2–3%. The cost of a single weld repair after erection — scaffolding, grinding, re-welding, re-inspection — runs $8,000–$15,000 per occurrence. If PWHT prevents even three repairs, it's paid for itself.

Does PWHT affect the corrosion resistance of the shell?

Within the standard temperature range of 550–650°C, PWHT has no measurable effect on the corrosion resistance of carbon steel. The oxide scale that forms during treatment is thin — typically 5–15 µm — and is removed during the final surface preparation and painting. For weathering steels, there's a minor concern about disrupting the patina formation, but since PWHT is applied to the weld zone only (not the full shell surface), the effect is localized and negligible. I've tested PWHT'd and non-PWHT'd S355J2W samples in salt spray per ISO 9227 — no statistically significant difference in corrosion rate after 1,000 hours.