Induction heating on a windy day: under-cured field-applied FBE

The control panel proves the powder is capable of curing; it does not prove the pipe reached temperature. The failure is a heat-balance and measurement problem:

• Single-point measurement at 12 o'clock. The top of the pipe is not the coldest point on a joint in a crosswind. The windward side loses heat by forced convection and is consistently the coldest quadrant. A thermocouple at 12 o'clock can read 240 C while the windward surface sits well below the 232 C window.
• Crosswind destroys the coil's symmetry. A 35 km/h wind at 2 C strips heat from the exposed joint faster than the coil can replace it. The induction coil heats by eddy current, but the windward surface is convecting away energy throughout the soak.
• Soak time is meaningless if the surface never reaches temperature. Thirty seconds at 232 C is not the same as thirty seconds averaged over an unequal thermal profile.
• Gel time of 22 s is short. With a fast-gelling powder, any surface below temperature will fail to sinter and wet out fully - producing the dull appearance and low MEK resistance, and later pinholes and poor cathodic disbondment.
• No verification of the coldest point: no second/third thermocouple, no thermal imaging, no data logging of the heat cycle.

The dull finish plus low solvent resistance is the classic fingerprint of under-cure: incomplete conversion, low Tg, residual exotherm on DSC.

Equipment and environment

• Build a windbreak/enclosure around the joint and insulate the exposed pipe either side of the coil, so the heat balance is not weather-dependent.
• Increase induction capacity or coil coverage, or use a coil configuration with better axial and circumferential uniformity; confirm the joint is centred in the coil.

Measurement

• Minimum four thermocouples at 12, 3, 6, and 9 o'clock, with the coldest reading controlling the acceptance decision; magnetically attached and shielded from the coil field.
• Data-log the heat cycle for every joint, with a strip chart retained in the joint record.
• Use thermal imaging as a cross-check on the windward quadrant before powder application, not as the sole record.
• Verify the thermocouple calibration and the contact condition - a loose bead reads low and causes over-heating elsewhere.

Procedure

• Rewrite the window around the coldest measured point: all four positions must be at or above 232 C before the soak starts, and remain there through the soak.
• Consider a slower-gel FBE qualified for field application, so small temperature variations do not produce a dry, unsintered surface.
• Re-qualify the procedure with a full joint under simulated wind, then verify cure by DSC (Tg and residual heat of reaction) and MEK rub on coupons from the qualification joint.

Do not accept on appearance alone, and do not accept a promise to 're-test later' without a defined test plan.

1. Immediate hold on the 12 joints; no lowering-in or backfill.
2. Objective testing on each joint: DSC for Tg and residual exotherm against the qualified values, MEK rub, and adhesion/peel. Cut a small sample from the windward quadrant - the location most likely to have failed - not from an accessible, convenient spot.
3. Interpretation. A joint meeting Tg and residual-exotherm criteria plus MEK and adhesion passes, regardless of appearance. A dull finish with low MEK almost always fails DSC; if so, remove the coating completely (grit blast to bare metal, Sa 2 1/2) and reapply. Partial patching of an under-cured thermoset is not acceptable - cure state is a bulk property, not a surface one.
4. Trend review. If any joint in the run fails, treat the whole run as suspect and test a statistical sample before accepting any of it.
5. Records. Document the disposition of each joint by number, the test method, the result, and the corrective action, and issue a nonconformance report that changes the procedure before the crew resumes.

The cost of removing 12 joints is small; the cost of 12 under-cured girth welds on a buried line is discovered five to ten years later through ILI.