Pneumatic versus Hydrostatic Proof Test on a Fuel-Gas Line
A 220 m long fuel-gas line to a furnace, 150 mm nominal bore (internal diameter about 152 mm), is complete and must be proved before commissioning. Design pressure is 12 barg. The contractor proposes a pneumatic test at 13.2 barg because there is no water supply on site, the line will not need drying, and the test can be done in a day. The route runs alongside a workshop holding 40 people, the nearest point of the pipe is 25 m from the workshop, and a 400 mm live steam main runs 2 m away. The contractor's method statement has no exclusion zone, no calibrated chart recorder and no defined hold period. The client's project manager says: "We tested at 12 bar for 30 minutes with a gauge and it held, so just sign the certificate; the plant loses money every day." You are the engineer asked to sign.
1. Quantify the stored energy in the proposed pneumatic test and explain what that number means for the decision.
Read the worked answer
Internal volume: V = (π/4) × 0.152² × 220 = 0.01815 × 220 ≈ 4.0 m³.
Stored energy for isothermal expansion of a compressed gas to atmosphere: E = P₁·V·ln(P₁/P_atm)
At a test pressure of 13.2 barg, P₁ ≈ 1.42 MPa absolute: E = 1.42 × 10⁶ × 4.0 × ln(1.42/0.101) = 5.68 × 10⁶ × 2.64 ≈ 15 MJ
15 MJ is roughly 3.6 kg of TNT equivalent - order-of-magnitude, but the correct order of magnitude for the decision. The significance is not the number itself, it is the failure mode: a compressed-gas release is an uncontrolled, nearly instantaneous expansion. A 220 m line that ruptures can whip, launch fittings and insulation, and project a jet across 25 m. Water at the same pressure stores almost no energy - a small leak is a puddle, and a joint failure is a local spill rather than an explosion.
So the question is not whether a pneumatic test can be made to pass, it is whether the site can safely absorb a failure. With an occupied workshop at 25 m, a live steam main 2 m away and no exclusion zone, the answer is no. The proposal must be changed, not merely documented.
2. Specify the test you would approve, including what the contractor must provide, and say what you would accept if water truly cannot be used.
Read the worked answer
Primary specification: hydrostatic test at 1.5 × design = 18 barg, using clean water.
What the method statement must include:
- Test volume about 4 m³, so roughly 4 t of water: verify pipe supports, anchors and expansion loops for the added weight and for the test temperature, and check that the line is not restrained in a way that locks in stress.
- Vents at all high points, drains at low points, and temporary strainers removed or protected.
- Two calibrated pressure gauges (one at the low point, one at the high point where the reading is lowest) plus a chart recorder or data logger; record water and ambient temperature throughout.
- A defined hold period per the applicable code, with the pressure stabilised before the hold starts, and a documented acceptance criterion with no visible leakage.
- A drying procedure afterwards (swabbing plus dry air or nitrogen sweep) and a dew-point check before gas is admitted, since the line feeds a furnace.
- Barricading and a competent person in charge, with a written test procedure approved before the day.
If water genuinely cannot be used, pneumatic testing is not simply forbidden, but it must be a managed operation: reduce stored energy by testing in short sections, define and enforce an exclusion radius with the workshop evacuated, pressurise remotely in stages, use calibrated recording instruments and a defined hold, and obtain documented management approval against a formal risk assessment. That is a much larger exercise than the contractor's one-day plan, which is the point.
3. The project manager wants the 12 bar/30 min gauge test signed off. Give your response, including the technical reason the pressure reading alone proves very little.
Read the worked answer
I will not sign it, and the reason is technical before it is procedural:
- Pressure is too low. A 12 barg design line requires 13.2 barg for a pneumatic proof (1.1×) or 18 barg hydrostatically (1.5×). A test at 12 barg has not proved the line to any recognised margin.
- Temperature correction was not applied. In a sealed gas-filled line, pressure varies with absolute temperature: ΔP ≈ P × ΔT/T. For P = 1.42 MPa absolute and a 20 K fall (say 45 °C at test to 25 °C later), ΔP = 1.42 × 20/318 ≈ 0.09 MPa ≈ 0.9 bar. So a line that was correctly pressurised in the afternoon can read almost a bar low the next morning, and a line that appears to "hold" may simply have been tested during a warming period. Without logged temperature, the reading cannot be interpreted.
- No traceability. No calibration certificate on the gauge, no chart recorder, no defined hold period, no record of stabilisation time; and a pneumatic holding test is far less sensitive to small leaks than a hydrostatic test.
My response: issue a non-conformance, refuse the certificate, and state in writing what is required to close it - a properly specified and witnessed test per the approved procedure. I would also escalate to the client's engineering authority, because a project manager's schedule pressure is not a basis for certifying a gas line that will run beside an occupied workshop. If the client wishes to commission without a valid test, that decision must be taken above me and recorded as such; it will not carry my signature.