In material traceability, the error is usually at the installation, not the certificate

There is no study that reduces the "wrong material rate" to a single number. The figures found measure different things:

• A 1995 technical paper estimates unintended material substitution in 1-3% of alloy components. No sample; it is a warning.
• In a refinery review of more than 420,000 procurement records from 17 sites (Pinnacle Reliability, 2022), about 1.8% of records appeared to need PMI but the flag was missing. This is not a wrong-metal rate, it is a missing control flag; no physical result is given.
• In the Chevron Richmond incident, 6 of 12 samples from the failed circuit had silicon below 0.10%. This is not proof of an MTC error; the specification had no minimum silicon requirement.

In documented incidents the pattern is common: at BP Texas City, similar-looking elbows were mixed up during maintenance and a carbon steel elbow went into a high-temperature line instead of 1.25 Cr; the maintenance procedure did not require PMI after installation; a few months later the elbow failed and a fire followed. This was not certificate fraud; it was an installation and verification failure.

Conclusion: the chain between certificate and installed part can break, and the break is usually in the field. Measuring takes four separate denominators: heat mark/MTC mismatch on the incoming part, PMI on the installed part, alloy verification of weld metal, and parts with an incomplete chain from certificate to installation. "MTC missing" and "wrong material" are not the same thing.

Sources: Inspectioneering (1995); Pinnacle Reliability refinery case page (2022); CSB Chevron Richmond report; BP Texas City RHU bulletin; Becht (2024); Watts Bar heat code plan (nuclear, different context).