Radiographic testing gets more attention in the field, but ultrasonic testing is the NDE method of choice for most structural CJP groove welds thicker than 5/16 in (8 mm). UT finds planar discontinuities — incomplete fusion, incomplete joint penetration, cracks — that RT can miss when the defect plane is not parallel to the X-ray beam. Understanding how AWS D1.1:2025 defines acceptance for UT indications is essential for any CWI signing off on structural weld reports.
How AWS D1.1:2025 UT acceptance criteria work
AWS D1.1:2025 Clause 8 covers inspection, and the UT acceptance criteria for CJP groove welds in structural connections rely on an indication rating system rather than a fixed flaw-size limit. The rating accounts for two factors simultaneously: how strong the reflected signal is relative to a calibrated reference, and how long the indication appears to be along the weld axis.
The key parameter is the d-value (also called the indication rating):
- The UT technician calibrates the instrument to a reference reflector in a standard calibration block. This establishes a baseline signal level — the reference.
- For each reportable indication, the technician measures the peak signal amplitude relative to that reference (in decibels).
- The amplitude reading is adjusted for beam path length using an attenuation correction to arrive at the d-value.
- The d-value and the evaluated indication length together determine whether the indication is acceptable.
A large positive d-value (high amplitude, high severity) is rejectable regardless of length. A moderate d-value is rejectable only if the indication extends beyond a length limit. A sufficiently negative d-value (small reflector well below reference) is acceptable regardless of length.
This system reflects engineering reality: a long, weak indication is less structurally significant than a short, high-amplitude one. It also means that a UT report showing only "accept/reject" is not informative enough — the d-value and indication length must both be recorded for the acceptance call to be verified.
Static vs. cyclically loaded structures
AWS D1.1:2025 establishes two acceptance tiers:
Statically loaded structures (most buildings, industrial structures, non-seismic frames) use the less restrictive criteria. Higher d-values and longer indication lengths are permitted before rejection.
Cyclically loaded structures (bridges designed under fatigue provisions, crane runway girders, structures with repetitive live load) use stricter criteria. The tighter limits exist because fatigue crack propagation can grow a below-threshold discontinuity into a critical crack over the structure's service life. An indication accepted under static criteria can be rejectable under cyclic criteria.
The engineer of record or the project specification designates which category applies. When in doubt, the CWI should confirm the designation in writing before issuing the UT report. Applying static criteria to a cyclically loaded structure is a quality escape that is very difficult to retroactively close.
Weld types and scanning coverage
AWS D1.1:2025 UT scanning requirements for CJP groove welds specify:
Base metal scanning. Before scanning the weld, scan the base metal adjacent to the weld to identify any laminar reflectors that could mask weld discontinuities. Document any base metal reflectors separately.
Weld scanning angles. Shear wave scanning is performed with the transducer on the base metal surface adjacent to the weld. Multiple scan angles (typically 45°, 60°, and/or 70°) are required to provide full volume coverage of the weld cross section. Which angles are required depends on weld geometry and joint configuration.
Scanning coverage. The technician scans the weld from both sides of the weld axis where access allows. For T-joint and corner welds where one side is inaccessible, the procedure must address how coverage is achieved from the accessible side.
Scan documentation. The UT report must record: identification of the weld, weld joint type and base metal thickness, instrument and transducer data, calibration data, scan coverage achieved, each recordable indication with its location (depth, distance from reference edge), d-value, evaluated length, and accept/reject determination.
Always-rejectable conditions
Regardless of d-value classification, the following conditions are always rejectable under AWS D1.1:2025:
- Cracks — any indication interpreted as a crack is automatically rejected. There is no d-value threshold below which a crack is acceptable.
- Incomplete fusion at the root of a CJP groove weld — planar lack-of-fusion at the root is treated as a crack-like indication for acceptance purposes.
- Incomplete joint penetration in a weld specified as CJP — by definition, any area of the specified joint that was not fused through the full thickness is rejectable.
When a UT indication has the characteristics of a planar reflector (sharp leading edge, high amplitude relative to its apparent length, specular reflection), the technician should interpret it as a planar discontinuity and reject it regardless of d-value, rather than downgrading it to a volumetric classification to pass the amplitude threshold.
Connecting UT to the WPS and PQR
The WPS specifies NDE requirements in the inspection section of the form. When UT is specified, the WPS should identify: the applicable code and edition (AWS D1.1:2025), the weld type (CJP groove), the load type classification (static or cyclic), and any additional project NDE specification requirements.
The PQR does not drive UT acceptance criteria — the governing code does. But if the PQR test included UT as part of procedure qualification testing, the UT results should be documented in the PQR record. This is particularly relevant for weld procedures qualified under Annex A (alternative prequalified joint details) or for thick section procedures where the qualification plate was UT inspected.
For fabricators building audit-ready welding procedure libraries, the UT documentation package (reports, calibration records, technician certifications) is a required component of the weld audit packet. Third-party auditors verifying AISC or AWS D1.1 compliance will pull UT reports and cross-check them against the acceptance criteria for the structure type. See NDE documentation and audit packet requirements for how to organize this package.
Alternative UT methods: PAUT and TOFD
AWS D1.1:2025 Annex B provides qualification requirements for phased array UT (PAUT) and time-of-flight diffraction (TOFD) as alternatives to conventional single-element shear wave inspection. Both methods offer advantages for specific joint configurations:
- PAUT allows electronic steering and focusing, providing better coverage of complex joint geometries and the ability to generate encoded scan images (B-scan, S-scan) that can be archived and reviewed remotely.
- TOFD is highly sensitive to planar reflectors and provides accurate depth sizing, but requires interpretation by a qualified technician and has near-surface dead zones that require supplemental conventional scanning.
Both Annex B methods must be qualified through a written procedure demonstration before use on production welds. The qualification requirements and scan documentation standards are more extensive than conventional UT, but the improved detection and sizing capability is often worth it for high-value or fracture-critical welds.
Practical notes for CWIs
When reviewing a UT report:
- Confirm the technician certification level (AWS QC1 Level II or equivalent ASNT or SNT-TC-1A Level II minimum for interpretation).
- Verify that the calibration block matches the base metal type and that calibration was performed and documented before and after the inspection.
- Check that the scan coverage table accounts for both sides of the weld axis and all required angles.
- For each rejection, verify that the d-value, indication length, and location are documented precisely enough to locate the indication for repair.
- Confirm that the acceptance criteria column cites the correct table for the structure load type.
UT is a skill-dependent method. A report that shows 100% acceptance on a weld known to have challenging joint geometry (tight backing clearance, high restraint, 100% E71T-8 or similar) warrants a closer look at scan documentation. Compare to common WPS deficiencies found in third-party audits for patterns that often correlate with UT problems downstream.
Rule library based on AWS D1.1:2025; verify against your governing edition.