PMI in practice: when the certificate and the alloy disagree
Positive material identification takes seconds per item. The mix-ups it catches take years to surface — usually in service.
Positive material identification is the least glamorous check in our scope and the one that most often changes an outcome. A handheld XRF or OES instrument reads the alloy composition of the item in front of the inspector, and that reading is compared against what the certificate and the specification say the item should be.
The failures it catches are rarely fraud. They are mix-ups: a 316L fitting in a 304 batch because two bins sat next to each other, a carbon steel elbow in a low-alloy line because the coating hid the marking, a bolt lot part-substituted when the original order came up short. Nobody intended it, and no paperwork records it.
The consequences land later. Alloy substitution in a corrosive or elevated-temperature service does not fail on installation; it fails after commissioning, and by then the item is welded into a system and the supplier relationship is years old.
What matters procedurally: decide the sampling rate before mobilisation and write it into the ITP. One hundred per cent PMI is normal for pressure-retaining components in critical service; a sampling plan is defensible for bulk structural material, provided the plan is agreed in writing rather than improvised at the works. Record the instrument, its calibration status, the reading, and the item's traceable marking — a PMI result with no item identity attached proves nothing.