The short answer
Most pressure-part radiography failures come from six causes: lack of fusion, lack of penetration, porosity, slag inclusion, undercut and cracking. All six are preventable through fit-up control, purging discipline, correct heat input and qualification-tested welders.
A failed radiograph on a pressure boundary is expensive in a way that is easy to underestimate. The joint has to be cut out, re-prepared, re-welded, re-heat-treated where applicable and re-shot. On a critical path that is not hours, it is days — and the hydro test moves with it.
The six defects that account for most rejections
| Defect | Usual cause | Prevention |
|---|---|---|
| Lack of fusion | Low heat input, wrong angle, dirty surface | Correct parameters, clean prep, trained hand |
| Lack of penetration | Root gap too tight, wrong root pass technique | Fit-up control, verified root gap |
| Porosity | Moisture, loss of purge, contaminated consumable | Oven-baked consumables, purge discipline |
| Slag inclusion | Poor interpass cleaning | Enforced grinding between passes |
| Undercut | Excess current, wrong weave | Parameter control, welder qualification |
| Cracking | Restraint, hydrogen, missed pre-heat | Pre-heat control, low-hydrogen consumables |
Fit-up is where most of the outcome is decided
Root gap, alignment and internal mismatch determine whether a good welder can make a good joint. If the fit-up is wrong, no amount of skill at the torch recovers it. Fit-up should be a signed hold point on pressure parts — inspected before the root pass, not after the cap.
Consumable control is not administrative overhead
Low-hydrogen electrodes absorb moisture from the air. Once damp, they produce hydrogen-induced cracking that may not appear until after the weld has cooled — sometimes days later. Mother ovens, portable quivers and a genuine issue-and-return discipline are the control. A site where electrodes sit in an open box is a site that will have cracking.
Pre-heat and PWHT are measurements, not intentions
Pre-heat temperature has to be measured at the joint, on the correct side, at the correct distance — not assumed from the time the torch has been playing on it. Post-weld heat treatment needs calibrated equipment and a traceable chart per joint or per group. If there is no chart, from an audit point of view the heat treatment did not happen.
The hold-point discipline that works
- Welder qualification verified against the WPS before deployment, not at the gate
- Fit-up inspected and signed before the root pass
- Pre-heat measured and recorded before the arc is struck
- Interpass cleaning checked on multi-pass joints
- PWHT chart raised and filed at the time, not reconstructed
- NDT to the approved QAP percentage, with repair joints tracked separately
Track your first-time acceptance rate
Repair rate is the single most useful quality metric on a boiler job, and it is trivially easy to measure: joints accepted first time divided by joints shot. Track it by welder and by front. It tells you where the problem is while you can still fix it, rather than at the end when it is a commercial argument.
Topics