7.2 Ultrasonic Thickness Measurement (Article 23, SE-797)

Ultrasonic Thickness Measurement (Article 23, SE-797)

The September 2026-May 2027 API 510 effectivity sheet includes ASME Section V Article 23 only through SE-797 for ultrasonic testing. It does not include Article 4. The exam task is manual pulse-echo contact thickness measurement: obtaining a defensible wall reading, recognizing bad data, and preserving information for corrosion-rate and remaining-life decisions. Article 4 weld-flaw calibration blocks, DAC/TCG, PAUT, and TOFD may matter in field practice but are outside the listed Section V scope.

1. What the instrument measures

A gauge sends a longitudinal-wave pulse into the component and measures round-trip travel time to the back wall. Displayed thickness is acoustic velocity x round-trip time / 2. The gauge interprets time using its entered velocity; it does not directly see metal. Material, temperature, surface, coating, couplant, and reflector geometry can all change the indication.

2. Procedure, equipment, and calibration

A written technique identifies instrument, probe, material, expected range, surface, couplant, calibration standard, scan pattern, temperature limits, and recording rules. Equipment range and resolution must cover expected wall.

Calibrate with a known, acoustically similar standard. A two-point calibration bracketing expected thickness is preferable when supported because it checks zero offset and velocity. Verify calibration at the start and end and at procedure intervals. If verification is outside tolerance, readings since the last acceptable check are suspect and must be evaluated or repeated; adjusting the gauge does not validate prior data.

Couplant removes the air gap. Use enough for stable transmission without a thick variable layer that distorts a delay-line measurement. Keep probe pressure and orientation consistent, particularly on small-radius nozzles and rough surfaces.

3. Repeatable sequence

1. Confirm component, CML, material, nominal wall, and expected mechanism.
2. Verify instrument, battery, transducer face, cable, and calibration status.
3. Clean loose scale without grinding away meaningful base metal.
4. Calibrate or verify over the expected range.
5. Seat the probe squarely and obtain a stable repeatable indication.
6. Scan enough to find localized minima instead of recording the first value.
7. Repeat an unexpected value after cleaning and calibration checks; preserve a low until disproved.
8. Record value, units, location, equipment/probe, temperature, coating, calibration, operator, and date.
9. Reverify calibration at completion and resolve drift.

4. Temperature, coatings, curvature, and local damage

An ambient calibration on hot carbon steel usually overstates thickness because acoustic velocity falls as temperature rises. A field rule of thumb is about one percent indicated-thickness correction per 100 F above calibration, but use the approved correction for actual material and temperature. A rule of thumb is not a qualified procedure.

Paint and polymers have different velocities from steel. A first-echo reading can include coating time and read too thick. Coating-discrimination, multiple-echo, or echo-to-echo modes remove it only when instrument, probe, range, and procedure are suitable.

One point is inadequate for pitting, grooving, erosion, impingement, or under-deposit attack. Treat the CML as an area: scan, confirm the minimum, and record its reproducible location. Curved small nozzles require a suitable small footprint and stable seating.

5. Turning data into evidence

Compare new, baseline, and previous values, calculate long- and short-term rates, and investigate trend changes. A negative rate may reflect changed location, coating, calibration, temperature, or prior quality; it does not prove growth. An abrupt low may be real damage. Review data with the mechanism, required thickness, remaining allowance, and interval.

Exam traps

• Article 23 SE-797 is listed; Article 4 weld examination is not.
• Section V explains performance and records; acceptance comes from the referencing code.
• Stable coupling does not eliminate calibration, velocity, temperature, or location error.
• Never delete a low simply because it disrupts a trend.

6. Worked data-quality decision

Suppose a CML previously read 0.500 in. at ambient temperature and now reads 0.535 in. on a hot coated surface. Do not calculate a negative corrosion rate and extend the interval. First confirm the exact location, clean or account for coating by the qualified mode, verify calibration and material velocity, record temperature, and repeat the scan. If the corrected minimum is 0.470 in., that verified value enters the trend. The scenario tests the essential distinction between instrument display and accepted inspection data.