Silo Pressure Testing Procedures for New Installations: A Field Engineer's Guide

Pressure testing a new silo isn't paperwork — it's the moment you find out whether the thing will hold. The procedure involves pressurizing the shell to 1.5× the design pressure, holding for 30 minutes, and checking every weld, bolt, and penetration for leaks or deformation. Skip it, and you're gambling with a structure that holds 5,000 tonnes of material over a busy plant. Here's how we do it right.

1. Why Pressure Testing Isn't Optional

1.1 The Physics You Can't Ignore

A bulk solids silo isn't a static tank. It's a dynamic pressure vessel that cycles through filling, storage, and discharge — each phase loading the shell differently. Internal pressures during pneumatic conveying can spike to 0.3–0.5 bar above ambient. Dust explosion scenarios? Those hit 2.5 bar in milliseconds if your suppression system doesn't fire.

EN 1991-4 (Eurocode 1, Part 4 — Silos and Tanks) defines the pressure regimes you need to design for. But design calculations are only as good as the assumptions baked into them. Pressure testing validates those assumptions against reality. It's the difference between "the model says it works" and "we pressurized it and it held."

1.2 What Happens When You Skip It

I remember a project in Vietnam — a 3,000-tonne cement silo, 16 metres diameter, 32 metres tall. The contractor skipped the pressure test to save four days. Six weeks after commissioning, a weld seam on the lower cone split during a filling cycle. 800 tonnes of cement poured onto the plant floor. The cleanup took 11 days. The root cause? A lack-of-fusion defect in a circumferential weld that a 0.35 bar test would have caught instantly.

That's the thing about pressure testing. It doesn't just verify strength. It reveals fabrication defects that ultrasonic testing missed, bolt torques that drifted during assembly, and gasket seats that looked fine under visual inspection but leak under load.

2. Types of Pressure Tests You'll Actually Run

2.1 Pneumatic Pressure Test (Internal)

This is the standard test for most welded steel silos. You seal all openings, pressurize the interior with compressed air, and hold at the test pressure. For cement and clinker silos, the test pressure is typically 1.5× the maximum operating pressure — usually 0.35 to 0.5 bar for gravity-discharge systems, up to 1.0 bar for pressurized vessels.

The procedure follows ASME Section VIII principles adapted for thin-shell structures. You don't just slam it to full pressure. You stage it: 50% of test pressure, hold for 10 minutes, inspect. 75%, hold, inspect. 100%, hold for 30 minutes minimum. At each stage, you're checking with soapy water solution on every weld seam, bolt circle, and nozzle penetration.

Here's a detail most specs don't mention: ambient temperature matters. If the sun's been baking the shell at 40°C and you pressurize with 15°C compressed air, the differential contraction can mask leaks or create false readings. We schedule tests for early morning when the shell temperature is stable — ideally within 5°C of the air supply temperature.

2.2 Hydrostatic Test (Structural Integrity)

For concrete silos — and we build a lot of those for capacities above 10,000 tonnes — hydrostatic testing is the go-to. You fill the silo with water to the maximum design level and hold for 24–48 hours. The water load simulates the weight of bulk solids at approximately 1,000 kg/m³, which is conservative for most materials (cement bulk density runs 1,100–1,300 kg/m³, clinker around 1,400 kg/m³).

The advantage? Water is incompressible. If the shell fails, you get a leak, not an explosion. The disadvantage? You need a water source, drainage plan, and enough time for the concrete to reach design strength before testing. We don't hydrostatic test until the concrete has cured for at least 28 days and reached 90% of its 28-day compressive strength — typically 25 MPa for C30/37 concrete.

During the hold period, you're monitoring settlement at a minimum of four points on the foundation. Total settlement should stay under 25 mm. Differential settlement — the difference between any two points — shouldn't exceed 1:500. We use survey instruments reading to 0.1 mm, checked every 4 hours.

2.3 Negative Pressure (Vacuum) Test

This one's often overlooked, and it shouldn't be. During discharge, especially with fluidizing systems, internal vacuum can develop. EN 1991-4 requires you to check for external pressure buckling. The test is simple: pull a vacuum to the design negative pressure (typically -0.1 to -0.2 bar for most applications) and hold for 15 minutes. Watch the shell for any inward deformation.

On a bolted silo, this test will find missing or undertorqued bolts faster than anything else. I've seen panels bow inward 3 mm at a single missing bolt location. That's not a failure — yet. But it tells you the bolt pattern isn't continuous, and under repeated cycling, that panel will work-harden and crack.

3. Pre-Test Preparation — Where Most Projects Go Wrong

3.1 Structural Readiness Checklist

Before you even think about connecting the compressor, the silo has to be mechanically complete. That means:

• All shell welds finished and visually inspected (VT per EN ISO 5817, Level C minimum)
• All bolts torqued to specification — and I mean torqued, not just tightened. For M20 high-strength bolts in a bolted silo, that's typically 280–320 Nm. We check 10% of bolts randomly, and 100% of bolts in the lower third of the shell where stresses are highest.
• All appurtenances installed: manways, nozzles, ladder mounts, roof penetrations. You can't test a silo with temporary blank plates and then say it passed — the real hardware has to be in place.
• Dust collection and aeration systems disconnected or isolated. You're testing the shell, not the process equipment.

The concrete foundation needs to be fully cured and the anchor bolt nuts double-checked. We've seen anchor bolts that were torqued during installation but loosened as the grout settled. A quick re-torque with a calibrated wrench takes 20 minutes and prevents a very expensive surprise.

3.2 Instrumentation and Monitoring Setup

You need at least two independent pressure gauges on the silo — one at the compressor outlet and one at the highest point of the shell. The high-point gauge is the one that matters, because that's where pressure is lowest if there's any stratification or temperature gradient. Gauges should be calibrated within the last 6 months and have a range of at least 1.5× the test pressure with 1% accuracy.

For hydrostatic tests, you need settlement monitoring points — minimum four, installed on the foundation ring before the silo is built. For pneumatic tests, strain gauges on the shell at critical locations: mid-height of the cylindrical section, the cone-to-shell junction, and the top ring. We typically install 8–12 strain gauges and log data every 15–30 seconds throughout the test.

And you need a pressure relief valve set at 110% of test pressure. Non-negotiable. If the compressor malfunctions or someone misreads a gauge, that relief valve is the only thing between you and an overpressure event. We set ours, test them, and seal them with tamper-evident tags.

3.3 Safety Exclusions and Site Control

Nobody inside the silo during a pneumatic test. Nobody within 10 metres of the shell during pressurization. We establish a cordon with physical barriers — not just tape — and post a test supervisor with authority to abort.

The exclusion zone extends further for larger silos. For a 12-metre diameter silo tested to 0.5 bar, the stored energy in the compressed air is roughly 340 kJ. That's enough to launch a loose panel fragment at lethal velocity. We use a 15-metre radius for silos under 10 metres diameter, 25-metre radius for anything larger.

All personnel in the exclusion zone wear hard hats, safety glasses, and high-visibility vests. The test supervisor carries a radio and has direct line-of-sight to the compressor operator. If communication is lost, the test stops. Period.

4. Step-by-Step Test Execution

4.1 The Pressurization Sequence

Here's the sequence we follow for a standard pneumatic test on a welded steel cement silo:

1. Pre-test briefing. All personnel on site understand the procedure, the exclusion zones, the abort criteria, and the communication protocol. Takes 15 minutes. Saves lives.
2. Seal verification. Every opening is blanked or valved off. Manway covers are bolted with full gasket sets. We do a quick leak check at 0.05 bar — just enough to reveal major sealing errors before you commit to the full test.
3. Stage 1 — 50% test pressure. Pressurize to 50% of the target test pressure. Hold for 10 minutes. Inspect all accessible welds, bolts, and penetrations with soapy water. Listen for hissing. Feel for air movement with your hand (carefully — don't put your face near the shell). Record strain gauge readings.
4. Stage 2 — 75% test pressure. Increase to 75%. Hold for 10 minutes. Repeat inspection. Check pressure gauge stability — a drop of more than 2% during the hold period indicates a leak that needs investigation.
5. Stage 3 — 100% test pressure. Pressurize to full test pressure. Hold for 30 minutes minimum. Continuous monitoring of pressure gauges and strain gauges. Full visual inspection of the shell exterior. For large silos, we use binoculars or a drone for the upper sections — no climbing on a pressurized structure.

The pressurization rate matters. You don't want to shock the shell. We limit the rate to 0.05 bar per minute. For a 0.5 bar test, that means the ramp-up from zero to full takes 10 minutes. Slow. Controlled. Boring is good.

4.2 Hold Period and Inspection Protocol

During the 30-minute hold at full pressure, you're looking for three things: pressure stability, visible deformation, and audible leaks.

Pressure stability: the gauge shouldn't drop more than 1% over 30 minutes. Some variation is normal — temperature changes cause pressure fluctuations. If the ambient temperature drops 5°C during the test, you'll see a pressure drop of about 1.7% in a sealed volume. That's physics, not a leak. We correct for temperature using the ideal gas law and log ambient temperature every 5 minutes.

Visible deformation: we mark reference points on the shell before the test — chalk lines at 2-metre intervals around the circumference. During the hold, we check each mark for any outward bulging. For a 12-metre diameter silo, the maximum allowable deflection under test pressure is typically 15 mm at mid-height. Anything beyond that triggers an engineering review.

Audible leaks: a hissing sound means a leak. A means water test will find it. We mark every leak location with chalk, note the pressure at which it appeared, and photograph it for the test report.

4.3 Depressurization and Post-Test Verification

Depressurize at the same controlled rate — 0.05 bar per minute. Rapid depressurization can cause the shell to snap back and create stress reversals that fatigue the welds. Once at atmospheric pressure, wait 15 minutes for the shell to stabilize, then do a final visual inspection.

Check for any permanent deformation at the reference points. Measure bolt torques again on a random 5% sample — the test can loosen bolts that were marginally tight. Review all strain gauge data for any readings that exceeded 80% of yield stress. For S355 structural steel, that's 280 MPa. If you hit that, the shell is overdesigned for stiffness but underdesigned for the actual loads, and you need to revisit the calculations.

5. Common Failures and What They Actually Mean

5.1 Leakage at Bolted Connections

This is the most common finding, and it's usually not a structural problem — it's an assembly problem. Bolted silos rely on gasket compression to seal. If the gasket material is wrong (EPDM instead of silicone for high-temperature applications), if it's pinched during assembly, or if the bolt torque is uneven, you'll get leaks.

The fix is straightforward: re-torque the bolts in a star pattern to the specified value, replace damaged gaskets, and re-test. But here's what matters — the leak location tells you something about the assembly process. If leaks cluster at one elevation, the crew that worked that section probably had a technique issue. If leaks are random, the gasket batch might be defective.

5.2 Panel Deformation Beyond Tolerance

If a panel bows more than the allowable deflection under test pressure, you've got a structural issue. Could be an under-thickness panel (we've caught 4 mm panels installed where 6 mm was specified — the contractor substituted to save money). Could be a missing stiffener. Could be a fabrication error in the rolling — the panel curvature doesn't match the design radius.

We had a case in Malaysia where a 10-metre diameter fly ash silo showed 12 mm of outward deflection at mid-height during the 0.4 bar test. The allowable was 8 mm. Turned out the shell thickness was correct but the steel grade was wrong — yield strength tested at 220 MPa instead of the specified 355 MPa. The entire shell had to be replaced. The mill certificates had been falsified. That's why we do material verification testing, not just paperwork checks.

5.3 Foundation Settlement During Hydrostatic Test

Excessive settlement during a hydrostatic test is a foundation problem, not a silo problem. But it's your problem if you're the EPC contractor. Total settlement over 25 mm or differential settlement exceeding 1:500 means the foundation design didn't match the soil conditions.

The fix depends on the cause. If it's consolidation of the soil beneath the foundation, you might be able to grout under the foundation ring and re-test. If it's a bearing capacity failure, you're looking at foundation reconstruction — and that's a project delay measured in months, not days. This is why we insist on a geotechnical investigation with boreholes to at least 1.5× the foundation depth before we finalize the foundation design. It costs $15,000–$25,000. Foundation failure costs $500,000 and up.

6. Documentation and Sign-Off

The test report is a legal document. It's what you hand to the client, the insurer, and — if things go wrong — the investigator. It needs to be thorough.

At minimum, the report includes: silo identification and design parameters, test procedure reference (which standard you followed), calibration certificates for all instruments, ambient conditions throughout the test, pressure readings at each stage with timestamps, strain gauge data, photographs of any defects found, corrective actions taken, and final pass/fail determination signed by the test supervisor and the client's representative.

We also include the raw data — not just summaries. If there's a dispute six months later, you want to be able to reconstruct exactly what happened. We archive test data for the life of the silo plus 10 years. That's our standard. Some clients require longer.

Case Study: 5,000-Tonne Clinker Silo — Vietnam

Project: New clinker storage silo, 12 m diameter × 38 m height, bolted steel construction, design capacity 5,000 tonnes.

Test type: Pneumatic pressure test to 0.35 bar (1.5× maximum operating pressure of 0.23 bar during pneumatic filling).

Procedure: Three-stage pressurization (0.175 bar, 0.26 bar, 0.35 bar) with 10-minute holds at stages 1 and 2, 30-minute hold at full pressure. 12 strain gauges installed at mid-height and cone junction. 4 settlement monitoring points on foundation ring.

Findings: At 0.26 bar (stage 2), a leak was detected at a bolted connection on the north elevation, 8 metres above the cone. Soap test showed bubbling at 6 bolt locations. Investigation revealed the gasket had been pinched during assembly — a section was folded under itself, creating a channel. The gasket was replaced and bolts re-torqued to 300 Nm.

Result: After repair, the silo passed the full 0.35 bar test with zero leaks and maximum deflection of 6 mm (allowable: 15 mm). Strain gauge data showed maximum stress of 12 MPa at the cone junction — well below the 80% yield threshold of 280 MPa. Total test duration: 4.5 hours including the repair. The silo was commissioned two weeks later and has been in continuous service since 2021.

Frequently Asked Questions

How much does pressure testing add to the project cost?

Typically 1.5–2.5% of the total silo installation cost. For a $400,000 bolted steel silo, that's $6,000–$10,000. The cost covers compressor rental, instrumentation, technician time, and documentation. It's the cheapest insurance you'll ever buy. A single weld failure during operation can cost $50,000–$200,000 in downtime and repair — not counting the safety risk.

How long does a full pressure test take?

Plan for a full day. The actual pressurization and hold takes 2–4 hours for a pneumatic test, 24–48 hours for a hydrostatic test. But you need time for setup (instrumentation, sealing, safety cordon), the test itself, depressurization, post-test inspection, and documentation. We schedule pressure testing as a dedicated activity — not squeezed between other tasks. Rushing a pressure test is how mistakes happen.

Is pressure testing required by code?

Yes, in most jurisdictions. EN 1991-4 requires verification of shell integrity under design loads. EN 14491 (dust explosion protection) requires pressure testing of explosion relief devices and suppression systems. In China, GB 50128 (Code for Construction of Vertical Cylindrical Welded Steel Storage Tanks) mandates pressure testing for all new installations. Even where it's not explicitly required by code, it's required by common sense. We've never had a client regret doing a pressure test. We've had several who regretted skipping one.