Steel Silo Leak Detection and Repair: A Field Engineer's Guide

A leaking steel silo loses product, risks structural damage, and creates a safety hazard you can't ignore. The two things you need to know right now: most leaks start at weld seams and bolt holes, not mid-plate, and catching them within the first 6 months of appearance prevents 80% of the expensive downstream damage. This guide covers what actually works on site — not what looks good in a manual.

Why Steel Silos Leak in the First Place

People assume a steel silo is a sealed vessel. It's not. It's a structure with hundreds of connection points, each one a potential leak path. After 15 years of walking around silos on four continents, I can tell you the failure modes are remarkably consistent.

The Weld Seam Problem

Field welds are the number one leak source. Shop welds get proper preparation, controlled temperature, and inspection. Field welds get done in 40°C heat with a breeze blowing slag into the joint. I've seen longitudinal seam welds on 4,000-tonne cement silos with porosity so bad you could thread a wire through them. The weld looks fine from the outside. Then it rains, water finds the pinhole, and three months later you've got a rust streak running 2 metres down the shell.

EN 1090-2 governs execution class for steel structures, but the standard only works if someone's actually checking. On a project in Vietnam, we found that 12% of field welds on a cluster of five slag silos failed dye penetrant testing on the first pass. Twelve percent. The contractor had been signing off their own welds without a third-party inspector. We cut out and re-welded 340 linear metres of seam. Cost the project an extra $28,000 and three weeks. But the alternative was a silo that would've leaked from day one.

Corrosion Under Insulation (CUI)

CUI is the silent killer. You wrap a silo in 100 mm of mineral wool to control temperature, moisture gets trapped between the shell and the cladding, and the steel corrodes from the outside in. By the time you see the stain on the cladding, the shell has often lost 1.5–2.0 mm of thickness. On a 6 mm shell plate, that's a third of your structural margin gone.

The worst case I've seen was a fly ash silo in Malaysia. The plant had insulated the lower ring of the silo to prevent condensation. Five years later, we pulled the cladding during a routine inspection. The shell had pitted to 3.2 mm in spots — original thickness was 6 mm. We ended up replacing the entire bottom ring, 12 plates at 1,200 mm wide each. The silo was offline for 11 days. Total repair cost: $67,000 including scaffolding, plate, and welding. The insulation that was supposed to protect the silo nearly destroyed it.

Thermal Cycling and Fatigue

Silos that handle hot material — think clinker at 80°C or slag at 120°C — expand and contract every filling cycle. The shell moves. The roof moves. The connections between them don't always move together. Over time, the bolts at the roof-to-shell junction loosen by 15–20% of their original torque. Gaskets compress permanently. Flanges warp by 1–2 mm. That's all it takes for a leak.

A cement plant in Shandong called me at 2am. Water was dripping into their silo during a rainstorm, contaminating 200 tonnes of finished cement. The leak was at the roof hatch — a 400 mm diameter opening with a rubber gasket that had been in service for 7 years. The gasket had taken a permanent set and was no longer compressing. We replaced it with a silicone gasket rated for 200°C continuous service. Problem solved in 40 minutes. The lost cement cost them $18,000. The gasket cost $12.

Detection Methods That Actually Work

You can't fix a leak you can't find. And finding leaks in a 30-metre-tall silo with 8 mm shell plates isn't straightforward. Here's what I've seen work reliably, ranked by practicality.

Visual and Acoustic Inspection

Start with your eyes. Walk the silo shell with a high-lumen torch, looking for rust streaks, efflorescence (that white powdery deposit where moisture has leached through), and staining at weld seams. Rust streaks below a weld are a dead giveaway — water follows the seam until it finds a pinhole, then runs down the shell.

Acoustic detection works for pressurised silos. If you're running pneumatic conveying at 0.2–0.3 bar, a leak hisses. You can hear it with a stethoscope pressed against the shell, or with an ultrasonic detector that picks up frequencies between 35–45 kHz. On a coal silo in Indonesia, we found a 3 mm hole in the upper cone using an ultrasonic gun. The hole was on the north face, 18 metres up, invisible from the ground. The hiss was audible from 2 metres away once we knew where to listen.

Pressure Decay Testing

For sealed silos — fly ash, cement, any powder that's aerated — pressure decay testing is the most reliable method. You pressurise the silo to 0.1 bar, isolate it, and measure the pressure drop over 30 minutes. A healthy silo loses less than 5% of pressure in that window. If you're losing 10% or more, you've got a leak somewhere.

The catch: pressure decay testing tells you there's a leak, not where it is. You still need to combine it with visual inspection or soap solution testing on the shell. We use a 5% soap-and-water solution sprayed on weld seams and bolt holes. Bubbles form at the leak point. It's low-tech, but it works. On a 2,500-tonne alumina silo in Australia, we found four leaks this way — all at bolt holes in the roof flange where the gasket had shrunk by 3 mm.

Tracer Gas Methods

For silos where you can't pressurise easily — large diameter, thin shell, no fan capacity — tracer gas detection is the answer. You introduce a helium-nitrogen mix into the silo at 0.05 bar and scan the exterior with a helium detector. Helium molecules are small enough to find leaks that water won't penetrate. Detection sensitivity goes down to 1×10⁻⁶ mbar·L/s.

It's expensive. A helium detection rig rents for $800–$1,200 per day, and you need a trained operator. But for a 10,000-tonne cement silo where a single leak could contaminate the entire contents, it's worth it. We used this method on a project in Saudi Arabia where the silo was storing food-grade calcium carbonate. The client couldn't risk any moisture ingress. We found two leaks at the shell-to-bottom junction — both were incomplete penetration welds from the original construction.

Repair Approaches by Leak Type

Not all leaks are equal. The repair method depends on what's leaking, where it is, and how much structural integrity you've still got.

Weld Defect Repair

For pinholes, porosity, and cracks in weld seams, the repair is straightforward: grind out the defect, weld it back up, and re-inspect. The key is grinding deep enough. I've seen contractors grind 2 mm into a 6 mm weld, slap a bead on top, and call it fixed. The defect was 4 mm deep. It came back in three months.

The rule: grind to sound metal, then go another 2 mm. For a 6 mm fillet weld, that means removing at least 6 mm of weld metal. Then re-weld using the same procedure as the original — same electrode, same preheat, same pass sequence. EN 1090-2 requires that repair welds meet the same acceptance criteria as the original. No exceptions.

On a slag silo in South Africa, we repaired 14 weld defects in the lower ring. Each repair took 45 minutes of grinding and 30 minutes of welding. Total time: 17 hours over two days. Cost: $4,200 including NDT re-inspection. The silo had been leaking for 8 months before they called us. The moisture had corroded the shell around each leak point, so we had to plate-patch three of the repairs instead of just re-welding.

Corrosion Patch and Overlay

When corrosion has eaten into the shell, welding over the top doesn't work. You need to cut out the damaged area and weld in a patch plate. The minimum patch size is 150 mm × 150 mm — anything smaller and the heat-affected zone from welding overlaps the patch edges and creates new stress concentrations. Patch plate thickness should match the original shell, and the material grade must be equivalent or better.

For widespread corrosion — say, the entire bottom ring has thinned from 8 mm to 5 mm — you can overlay a new plate over the old one. We did this on a 3,000-tonne coal silo in India. The lower 2 metres of the shell had thinned to 4.5 mm from 8 mm original. We overlaid 6 mm plates, stitch-welded at 150 mm intervals with 50 mm welds. The silo stayed in service during the repair — we worked in 1.2-metre vertical bands, going around the circumference. Took 6 days. Cost: $34,000. A full ring replacement would've been $90,000 and 3 weeks offline.

Sealant and Coating Systems

Not every leak needs welding. For non-structural leaks — roof flanges, hatch seals, bolt holes — a high-quality sealant can be the right fix. Polyurethane sealants work for static joints. Silicone sealants handle temperatures up to 250°C. For chemical resistance — think alumina or gypsum silos — epoxy-based sealants are better.

The surface prep is everything. I've seen sealant applied over rust, over old sealant, over oil. It fails every time. The surface needs to be clean, dry, and profiled. For steel, that means wire-brushing to bright metal and wiping with solvent. On a fly ash silo in Thailand, we re-sealed the roof-to-shell junction with a polyurethane sealant after cleaning the flange with a needle scaler. The previous sealant had been applied over rust and lasted 8 months. The new sealant, properly applied, was still holding after 4 years when I last checked.

When Repair Isn't Enough

Sometimes the damage is too widespread for patching. You need to know when to stop repairing and start replacing.

Structural Assessment Thresholds

Eurocode 3 (EN 1993-1-1) gives you the framework for assessing damaged steel structures. For silo shells, the critical threshold is when remaining thickness drops below 50% of the original design thickness. At that point, the shell can no longer carry the full design load — including wind, seismic, and material pressure — and you need to either reinforce or replace.

We assessed a 15-year-old cement silo in the Philippines where the lower ring had thinned to 3.0 mm from 6.0 mm original. The silo was 22 metres tall, 12 metres diameter, holding 5,000 tonnes. The shell was below 50% thickness over 40% of the lower ring. We recommended full ring replacement. The plant wanted to patch it. We walked away from the job. Six months later, the shell buckled during a filling cycle. No one was hurt, but the silo was a total loss. The replacement cost $380,000. The patch job would've been $45,000.

Partial vs. Full Panel Replacement

If the damage is localised — say, one or two plates in a specific ring — you can replace individual panels. The cut-out needs to extend at least 50 mm beyond the damaged area in all directions. The new plate is beveled for full-penetration welding, and the welds are 100% inspected by ultrasonic or radiographic testing per EN ISO 5817, quality level B.

Full panel replacement on a 10-metre-diameter silo means cutting out a plate that's 3 metres wide and 2 metres tall. That's 6 square metres of steel, roughly 280 kg at 6 mm thickness. You need a crane, a welding crew, and a structural engineer on site to sign off the repair. Budget $8,000–$15,000 per panel including materials, labour, and inspection. A full ring replacement — 12 to 16 plates — runs $80,000–$150,000 depending on diameter and access.

Prevention: Designing Leaks Out

The best leak is the one that never happens. Design choices made before construction determine whether your silo leaks in year 2 or year 20.

Material Selection

For most industrial applications — cement, clinker, slag, coal — carbon steel with a proper coating system is sufficient. The coating matters more than the base metal. A three-coat system — zinc-rich epoxy primer (75 µm), epoxy intermediate coat (150 µm), polyurethane topcoat (50 µm) — gives you 275 µm total dry film thickness. That's good for 15–20 years in a normal industrial environment.

For corrosive materials — gypsum, alumina, chemical powders — consider stainless steel for the lower rings where moisture accumulates. Duplex 2205 handles chloride environments that would eat through carbon steel in 3 years. It costs 4–5 times more per tonne, but you only need it for the bottom 3–4 metres of shell. On a gypsum silo in Qatar, we used carbon steel for the upper rings and duplex 2205 for the lower two rings. The silo's been in service for 9 years with zero shell corrosion. A neighbouring plant built an all-carbon-steel silo at the same time. They replaced the bottom ring after 6 years.

Drainage and Sealing Details

Water is the enemy. Every leak path starts with water getting somewhere it shouldn't. Design your silo with positive drainage: sloped roofs (minimum 1:50 gradient), drip edges on all penetrations, and weep holes at the base of insulation cladding. The roof-to-shell junction is the most common leak point — use a continuous gasket, not spot sealant, and specify a compression stop that maintains 30% gasket compression after 10 years of service.

Bolted connections need torque specs and a re-torque schedule. We specify initial torque values based on bolt diameter and grade — for M20 bolts in tension, that's typically 250 Nm for Grade 8.8. Then we require re-torque at 6 months, 12 months, and annually thereafter. On a project in Brazil, the contractor skipped the 6-month re-torque. Within a year, 30% of the roof flange bolts had loosened by more than 20%. We found water ingress at 8 bolt locations during the first heavy rain. Re-torqueing the entire flange took 4 hours and cost $600. The water damage to the cement inside cost $22,000.

Case Study: 4,000-Tonne Clinker Silo — Leak Detection and Repair

A cement plant in East Africa contacted us about a 4,000-tonne clinker silo that was leaking during the rainy season. The silo was 28 metres tall, 14 metres diameter, built 8 years earlier. Water was entering the shell and contaminating the clinker, causing quality rejects at the mill.

Detection: We started with a visual inspection and found rust streaks at three locations on the lower ring, all below field weld seams. Pressure decay testing confirmed a leak rate of 15% over 30 minutes — well above the 5% threshold. Soap solution testing on the seams identified six pinhole leaks, each 1–3 mm in diameter, all at the root pass of single-sided fillet welds.

Root cause: The original welds had been done without back-grooving. The root pass had incomplete penetration of 2–3 mm. Over 8 years of thermal cycling — clinker at 90°C going in, ambient at 25°C at night — the residual stress in the root pass caused micro-cracks that opened into pinholes.

Repair: We ground out the defective root passes to a depth of 8 mm, back-grooved from the inside using a carbide burr, and re-welded with a qualified procedure — E7018 electrodes, 150°C preheat, two-pass fill. All repairs were inspected by dye penetrant testing per EN ISO 3452-1. We also applied a two-coat epoxy system (200 µm DFT) over the repaired areas as additional protection.

Result: The repair took 3 days, cost $11,500 including materials, labour, and inspection. Pressure decay testing after repair showed a 2% drop over 30 minutes — within acceptable limits. The plant reported zero clinker contamination in the following rainy season. We also recommended a re-torque of all roof flange bolts and replacement of the roof hatch gasket, which added $1,200 to the job.

Frequently Asked Questions

Q: How often should I inspect my steel silo for leaks?

A: Walk-around visual inspection every 3 months. Pressure decay testing annually for sealed silos. Detailed NDT inspection of weld seams every 5 years, or immediately if you notice rust streaks, efflorescence, or product contamination. If your silo handles hot material, add a thermal imaging survey every 2 years — it finds insulation failures before they become corrosion problems.

Q: Can I repair a leak while the silo is full?

A: For non-weld repairs — sealant application, gasket replacement, bolt re-torque — yes, you can work on a full silo. For welding or cutting, the silo needs to be empty and ventilated. Welding on a silo with combustible contents is an explosion risk. ISO 15012 requires gas monitoring and a hot work permit. We've done external patch welding on silos with material inside, but only after confirming the headspace dust concentration is below 10% of the minimum explosible concentration. That's rare in practice. Plan repairs during scheduled shutdowns.

Q: What's the typical cost range for silo leak repair?

A: Minor repairs — sealant, gasket replacement, bolt re-torque — run $500–$3,000. Weld defect repair on a few locations: $3,000–$12,000. Single panel replacement: $8,000–$15,000. Full ring replacement: $80,000–$150,000. Structural assessment and engineering: $5,000–$15,000. These are on-site costs including labour, materials, and inspection. They don't include lost production during downtime, which is usually the bigger number.

Get Your Silo Leak Assessment from Manxing

We've inspected, assessed, and repaired steel silos across 30+ countries — cement plants in Southeast Asia, slag facilities in the Middle East, coal handling terminals in Africa. Our engineers don't just find leaks; we tell you why they happened and how to stop the next one.

If you've got a silo that's leaking, or you're not sure whether it is, contact Manxing for a site assessment. We'll send a senior engineer, not a salesperson. You'll get a written report with photos, measurements, repair options, and a cost estimate you can actually budget from.

Email: [email protected] | Website: manxingstorage.com