ASME Weld Ultrasonic Inspection: Testing for Pressure Vessels and Piping

ASME weld ultrasonic inspection is the examination of pressure vessel and piping welds by the methods in ASME Section V Article 4, judged against the acceptance rules of the construction code: Section VIII Division 1 Appendix 12 for vessels and B31.3 para. 344.6.2 for process piping. Indications above 20% of the reference level must be investigated.

The split between "how to examine" and "what to accept" is the single most useful thing to understand about ASME UT. Section V tells the examiner how to calibrate, scan and record. It does not say whether a weld passes. That decision comes from the code the vessel or piping is built to, and from the edition named in the contract.

Contents: What it covers · Which codes · Appendix 12 · B31.3 · UT in lieu of RT · Techniques · Calibration · Worked example · Thickness · Defects · Personnel · Limits · FAQ

A note on code text. ASME codes are paywalled. Clause numbers and limits below come from ASME's public descriptions, from procedures that quote the code, and from several agreeing secondary sources. They are marked as such. Check every value against the edition named in your contract before you use it.

What does ASME weld ultrasonic inspection cover?

It covers the UT examination of welds in ASME-built boilers, pressure vessels and piping, using Section V methods and the acceptance criteria of the referencing construction code. ASME's own description of Section V says it "contains requirements and methods for nondestructive examination which are referenced and required by other BPVC Sections" (ASME BPVC Section V, 2025 edition, publisher's description).

Section V is binding only when another code section or a contract calls it up. The current edition is 2025 (July 2025, per standards resellers).

Key takeaway: a UT report that cites only "ASME V" has said how the weld was examined, not whether it passed.

Which ASME codes call for weld ultrasonic examination?

For vessels, Section VIII Division 1 calls for UT through UW-53 and Appendix 12, and through Code Case 2235 when UT replaces RT. For process piping, B31.3 lets UT stand in for radiography under 341.4.1(b)(1) and judges it under 344.6.2 or Appendix R.

Section VIII Division 1 (pressure vessels)

Division 1 is a radiography-first code. UW-11 decides which joints need full radiography. UW-51 sets how full radiography is done and judged, and UW-52 covers spot radiography. UW-53 covers ultrasonic examination of welded joints and points to Mandatory Appendix 12 for the UT method and acceptance standard.

Appendix 12 applies whenever the code itself, or the designer, calls for UT. Typical cases are nozzle welds (Category D), welds that can't be radiographed in an interpretable way, and final closure seams where the interior is inaccessible. One published case study found that all the Category D nozzle welds on a vessel order "required ultrasonic examination per the requirements of Appendix 12" (Holthaus and Gaskin, Inspectioneering Journal, Jan/Feb 2020). Confirm the exact UW-11 paragraph for closure seams in your edition.

B31.3 (process piping)

In B31.3, radiography is the normal volumetric method. For normal fluid service, 341.4.1(b)(1) calls for random examination of at least 5% of circumferential butt and miter groove welds "by radiography in accordance with para. 344.5 or by ultrasonic examination in accordance with para. 344.6" (as quoted from the code on the AWS forum). UT is therefore a permitted alternative, not the default.

B31.3-2016 added Appendix R. Don Frikken of Becht describes it as fracture-mechanics-based acceptance criteria for UT techniques that size flaws, usable "in lieu of those described in para. 344.6" (Becht blog, June 17, 2016). The same edition moved Category M circumferential butt welds from 20% random to 100% examination by RT or UT.

Key takeaway: in both codes UT is often the alternative to RT, so the contract should say which method is being used and under which paragraph.

How does Section VIII Division 1 Appendix 12 accept or reject a weld?

Appendix 12 has two steps. Any imperfection whose indication exceeds 20% of the reference level is investigated. Cracks, lack of fusion and incomplete penetration then reject at any length, while other imperfections reject only if they exceed the reference level and a thickness-based length.

The numbers below are as reported by secondary sources and procedures that quote the appendix (Med confidence). Check them in your edition.

The word and in that rule is the part most often lost. A slag line at 90% of the reference curve is not rejectable under Appendix 12 however long it is, because it doesn't exceed the reference level. A slag line at 130% of the curve rejects only if it is also longer than the limit. A crack at 25% of the curve rejects outright, because it was above the investigation level and its type is never acceptable.

That is why characterization matters so much. The examiner has to decide whether an investigated indication is planar (crack, lack of fusion, incomplete penetration) or volumetric (slag, porosity), using echo dynamics, more than one angle, and where the reflector plots against the weld geometry.

The UW-51(b) radiography limits for elongated indications use the same thickness bands (1/4 in., t/3, 3/4 in.), which makes RT and UT results easier to compare on one vessel. For RT and UT acceptance across codes side by side, see our guide to ultrasonic testing acceptance criteria standards.

Key takeaway: under Appendix 12, amplitude decides whether you look; type, amplitude and length together decide whether the weld is repaired.

How does B31.3 judge ultrasonic indications in process piping?

B31.3 para. 344.6.2 rejects a linear-type indication only when its amplitude exceeds the reference level and its length exceeds a limit set by the weld thickness Tw. The limits, as quoted from the code in forum discussions (Med confidence), are:

The bands match Appendix 12, but secondary sources report one important difference. B31.3 344.6.2 doesn't reject lack of fusion regardless of length the way Appendix 12 does, while B31.3 radiography (Table 341.3.2) does reject lack of fusion. This is Low-to-Med confidence and needs confirming in the edition you work to. If it holds, a weld can pass UT and fail RT for the same flaw, so the choice of method should be agreed before work starts.

For radiography, Table 341.3.2 for girth welds in normal fluid service limits slag and elongated indications to an individual length of Tw/3 and a cumulative length of Tw in any 12Tw, with a width limit that sources quote as 2.5 mm or 2.4 mm. That table is not a UT table, and UT results shouldn't be judged against it.

Appendix R works differently. It is meant for techniques that measure flaw height and length, such as TOFD or encoded PAUT, and compares the sized flaw to fracture-mechanics limits instead of to a reference amplitude.

Key takeaway: in B31.3, check which acceptance route the owner has chosen (344.6.2 amplitude and length, or Appendix R sizing) before the procedure is written.

When can ultrasonic examination replace radiography?

For vessels and boilers, ASME Code Case 2235 allows UT in place of RT when its conditions are met: a written, demonstrated procedure, recorded data, a defined examination volume and thickness-based acceptance criteria. For piping, B31.3 341.4.1(b)(1) and Appendix R give the equivalent route.

Code Case 2235 covers Section I, Section VIII Divisions 1 and 2 and Section XII. Revision 2235-9 (October 2005) is the one most often cited. Copies of that revision state that welds 1/2 in. (13 mm) and thicker may be examined by UT in lieu of RT. Older revisions gave different minimum thicknesses, so read the revision your contract names.

The case is built around recorded, reviewable data. Its requirements, as described by procedures and vendor notes, include:

• a written procedure that has been demonstrated to find flaws in a qualification block
• computer-recorded data that shows probe location, so the examination can be reviewed after the fact
• an examination volume that covers the weld and the heat-affected zone on both sides
• acceptance criteria that depend on wall thickness and on whether the flaw is surface or embedded
• personnel qualified for the technique, with data analysts trained on the system used

Evident, an instrument maker, notes that much of this has since been "codified directly into the main bodies of the major code books", citing the 2010 and later editions of Section V Article 4 (Evident: phased array in lieu of radiography). That is a vendor source, so check the current edition before relying on the code case alone.

Key takeaway: UT in lieu of RT is not manual UT with a new name. It needs encoded, recorded data and a demonstrated procedure.

Which UT techniques are used on ASME welds?

Four techniques cover almost all ASME weld work: manual pulse-echo, phased array (PAUT), time of flight diffraction (TOFD) and automated UT (AUT). It is wrong to say there are "two types" of UT, conventional and PAUT. PAUT is a pulse-echo technique with a multi-element probe, and TOFD and AUT are separate approaches with their own rules.

Two cautions apply. ASTM E2700 warns that phased array can oversize flaws (section 1.2.1), and an encoded PAUT scan still travels along the weld; it doesn't "scan without moving". TOFD depth is calculated from diffraction geometry, and interpreting it still needs a trained analyst, so it is amplitude-independent but not operator-independent.

The ASTM practices that sit behind these techniques are ASTM E2700-26 for contact PAUT of welds and ASTM E2373/E2373M for TOFD. Section V Article 4 has its own mandatory appendices for TOFD and phased array, and an inspenet summary of those appendices covers scan overlap and data dropout limits for encoded PAUT. For a fuller comparison, see the difference between UT and PAUT and our guide to automated ultrasonic testing.

Key takeaway: choose the technique by what the acceptance route needs: amplitude for Appendix 12 and 344.6.2, sized height and length for Appendix R and some code-case criteria.

How is the equipment calibrated for an ASME weld examination?

Article 4 calibration starts with a basic calibration block made to T-434.2.1, from which the examiner builds a distance amplitude correction (DAC) curve. Indications are then judged as a percentage of that curve, with 20% DAC as the usual investigation level and 50% DAC as a common recording level.

The basic calibration block has three side-drilled holes (SDHs) at 1/4 T, 1/2 T and 3/4 T, and the holes are at least 1.5 in. deep according to block makers. Block thickness and hole diameter step up with the weld thickness range, so a block for a 1 in. weld is not the block for a 3 in. weld. These values come from vendor listings (Med confidence); check Figure T-434.2.1 in your edition.

The DAC curve joins the peak echoes from those holes at one fixed gain. The highest response is normally set to 80% of full screen height (FSH); that is common practice rather than a fixed rule. Points must lie beyond the near field. Evident's DAC tutorial describes the curve as one that "graphically compensates for material attenuation, nearfield effects, and beam spreading" (Evident: DAC and TVG).

Three more settings come from Article 4 as quoted in procedures (Med confidence):

• Scanning gain at least twice the reference level, which is +6 dB, so small reflectors aren't missed during the search
• Transfer correction for the difference in surface condition and attenuation between the block and the part
• Calibration checks at set intervals; secondary sources report that data must be re-examined if a DAC point drops by more than 20% or 2 dB (confirm the exact wording in your edition)

A 6 dB drop halves the echo amplitude, not its energy. Halving the energy is 3 dB.

Our DAC curve ultrasonic testing guide builds a full curve on a 1-1/2 in. block, with near-field checks and transfer correction.

Key takeaway: the 20% and 50% DAC lines are evaluation and recording levels; neither is an acceptance limit on its own.

Worked example: evaluating an indication on a 1 in. vessel weld

This example evaluates one indication on a 1.00 in. thick single-V longitudinal seam to Section VIII Division 1 Appendix 12. The indication is investigated, plots on the fusion face, and the result depends on whether it is characterized as slag or as lack of fusion. All numbers are illustrative; the procedure and the code edition govern a real job.

Step 1: setup

• Weld: single-V butt, 30° bevels, 1/8 in. root opening, t = 1.00 in.
• Probe: 60° shear wave. Leg length = t ÷ cos 60° = 1.00 ÷ 0.500 = 2.00 in. of sound path.
• Block: T-434.2.1 basic block, 1.00 in. thick, SDHs at 0.25, 0.50 and 0.75 in. depth.
• First-leg sound path to each hole = depth ÷ cos 60°: 0.50, 1.00 and 1.50 in. Second-leg sound path = (2T − depth) ÷ cos 60°: 2.50, 3.00 and 3.50 in.

Step 2: the DAC curve

The 0.50 in. point is the highest response, is confirmed beyond the near field, and is set to 80% FSH. The other points are recorded at that gain. Each dB value is 20 log₁₀(80 ÷ peak).

A measured transfer correction (say +2.0 dB) is added to the reference gain to give the evaluation gain. Scanning is done at least 6 dB above that.

Step 3: locate the indication

The peak echo is at a sound path of 1.30 in. That is less than the 2.00 in. leg, so it is a first-leg signal.

• Depth = 1.30 × cos 60° = 1.30 × 0.500 = 0.65 in. (inside the 1.00 in. wall)
• Surface distance from the probe index = 1.30 × sin 60° = 1.30 × 0.866 = 1.13 in.
• The probe index was 1.39 in. from the weld centerline, so the reflector plots 1.39 − 1.13 = 0.26 in. from the centerline, on the probe side.
• The fusion face at 0.65 in. depth lies 0.0625 + (1.00 − 0.65) × tan 30° = 0.0625 + 0.35 × 0.577 = 0.26 in. from the centerline.

The reflector plots on the bevel. That makes lack of sidewall fusion a real possibility, and it must be characterized, not assumed to be slag.

Step 4: amplitude against the curve

The DAC height at 1.30 in. is interpolated between the 1.00 in. (62%) and 1.50 in. (48%) points: 62 + (48 − 62) × (0.30 ÷ 0.50) = 62 − 8.4 = 53.6% FSH. At evaluation gain the indication peaks at 41% FSH.

• Amplitude = 41 ÷ 53.6 = 76.5% DAC, or 20 log₁₀(0.765) = −2.33 dB
• The 20% DAC line here is 0.2 × 53.6 = 10.7% FSH, so the indication is well above it and must be investigated
• Measured length, using the method in the procedure: 0.45 in.
• Appendix 12 length limit for t = 1.00 in. (between 3/4 and 2-1/4 in.): t/3 = 1.00 ÷ 3 = 0.333 in.

Step 5: the decision

Case A and Case D are the same echo. The only difference is the examiner's call on its type, which is why the plot on the fusion face, a second angle and echo dynamics matter more than the screen height. If this were a B31.3 pipe weld of the same thickness (Tw = 25.4 mm), the 344.6.2 limit would also be Tw/3 = 8.5 mm (0.33 in.), but Case D would be judged on amplitude and length as reported in the B31.3 section above.

Key takeaway: run the geometry before the amplitude. Where the reflector plots often decides its type, and its type decides the outcome.

Is there a minimum thickness for ASME weld ultrasonic testing?

Section V Article 4 doesn't set a blanket minimum such as "6 mm and up". Thickness limits come from the referencing code, the code case and the written procedure. The claim that Section V "generally supports" UT from about 1/4 in. (6 mm) has no source in the Article.

Where a limit exists, it belongs to a specific document:

Thin welds are harder for practical reasons: the probe dead zone, close-spaced root and cap echoes, and pipe curvature under flat wedges. Probe choice, curved wedges and procedure demonstration handle them, not a rule in Section V.

Key takeaway: if a procedure claims a thickness limit, it should cite the paragraph it comes from.

Which weld defects does UT find, and which does it miss?

UT is strongest on planar flaws that face the beam: lack of sidewall fusion, incomplete penetration and cracks. It is weaker on scattered porosity and on planar flaws tilted away from the beam. That is the reverse of radiography, which shows volumetric flaws well and can miss tight planar ones.

For two defect types in depth, see our guides to slag inclusion in welding and ultrasonic testing and lack of fusion in welds. Most of these flaws trace back to the welding procedure and how it is followed, which is covered in our guide to what a welding procedure specification is. For how UT compares with RT, MT, PT and the other methods, see our comparison of non-destructive testing methods.

Key takeaway: if porosity is the expected defect, RT may be the better method; if lack of fusion is, UT usually is.

Who may perform and evaluate the examination?

ASME UT is done by personnel certified by their employer under a written practice, usually based on ASNT SNT-TC-1A or ANSI/ASNT CP-189, as the referencing code requires. ASNT doesn't certify the technician under SNT-TC-1A; the employer does. The documents are listed on ASNT's standards page.

Techniques that record data, such as encoded PAUT, TOFD and AUT, add a further requirement: the person evaluating the data must be trained on that technique and software, and code-case work usually requires a demonstration. CP-189 states its requirements as mandatory, while most of SNT-TC-1A is written as recommendations, so check which one the code edition and contract call up.

What a good procedure and report contain

A published case study gives a sense of the stakes. On a new vessel order, inspectors found rejectable flaws in Category D nozzle welds, and 57 nozzle repairs were needed before the vessels went into service. The authors blamed inadequate written procedures, poor implementation, weak reporting and undertrained examiners (Holthaus and Gaskin, Inspectioneering Journal, 2020).

A procedure that holds up to review names the code and edition, the acceptance paragraph, the probe angles and frequencies for each thickness, the block, the DAC or TCG method, scanning gain, transfer correction, calibration-check intervals, scan coverage from both sides where accessible, and the recording and investigation levels. The report should give each indication's location, depth, length, amplitude in % DAC or dB, its characterization and the acceptance decision with the paragraph cited.

Key takeaway: the examiner's certificate matters, but a written procedure that matches the code is what makes the result defensible.

What ASME weld UT cannot do

UT gives an amplitude, a position and a measured length. It doesn't show a picture of the flaw, it can't prove a weld is flaw-free, and its result is only as good as the procedure, the coverage and the examiner's characterization.

• Orientation: a planar flaw tilted away from the beam can return less than 20% DAC and still be large. More than one angle is the protection.
• Coarse grain: austenitic stainless steel and nickel alloy welds scatter and skew the beam. They need refracted longitudinal wave or dual-element probes and a procedure demonstrated on representative welds.
• Geometry: nozzles, fillet welds, tapers and weld caps limit where the probe can go, and the cap may need grinding flush for full coverage.
• Thin wall and small bore: dead zones and curvature reduce reliability on thin sections and small pipe.
• Amplitude is not size: an echo below the reference curve can come from a large but poorly oriented flaw. Sizing needs tip diffraction (TOFD), PAUT sizing techniques or the methods Appendix R allows.
• Manual UT leaves no record of coverage. Where the owner needs reviewable data, specify encoded PAUT, TOFD or AUT.
• The code edition decides. Every limit in this article can change between editions. The contract's edition governs.

How EMA Industries applies this

EMA Industries provides NDT Level III services and weld ultrasonic examination from Southern California, mobilizing nationwide, with AWS D1.1 structural weld UT, MT and PT as core work. On an ASME job, a Level III review checks the procedure against the code edition named in the contract: the acceptance paragraph (Appendix 12, 344.6.2 or a code case), the calibration block, the DAC and recording levels, scan coverage and the characterization rules. For structural steel built to AWS instead of ASME, the rules differ; see our AWS D1.1 weld inspection criteria. To review a procedure or scope weld UT, see our ultrasonic examination services or contact EMA Industries.

Related reading: Phased array ultrasonic testing · Non-destructive testing for welding inspection · WPS and PQR

Frequently asked questions

What is the ASME standard for ultrasonic weld testing?

ASME BPVC Section V, Article 4 sets the UT methods for welds: equipment, calibration blocks, DAC and scanning. Acceptance comes from the construction code, such as Section VIII Division 1 Appendix 12 for vessels or B31.3 para. 344.6.2 for process piping.

What are the main types of ultrasonic weld testing?

Manual pulse-echo (angle and straight beam), phased array (PAUT), time of flight diffraction (TOFD) and automated UT (AUT). PAUT is a pulse-echo variant with a multi-element probe, so splitting UT into just "conventional and PAUT" leaves out TOFD and AUT.

What does 20% DAC mean in ASME UT?

It is the line 20% of the height of the reference curve, about 14 dB below it. Under Appendix 12 and Article 4-based procedures, any indication above it must be investigated and characterized before the acceptance criteria are applied. It is not an acceptance limit.

Can ultrasonic testing replace radiography on ASME vessels?

It can under Code Case 2235 for Section I and Section VIII vessels and boilers, provided the procedure is demonstrated, the data is recorded and the case's acceptance criteria are used. Copies of revision 2235-9 give 1/2 in. (13 mm) as the minimum thickness. For B31.3 piping, 341.4.1(b)(1) already allows UT for random examination.

How is the length of a UT indication measured?

By moving the probe along the weld and marking where the echo falls to the level the procedure specifies, often a 6 dB drop or the 20% DAC line. Appendix 12 compares that length to 1/4 in., t/3 or 3/4 in. depending on thickness, but only for indications above the reference level.

Sources

• ASME BPVC Section V, Nondestructive Examination, 2025 edition, publisher's description via AFNOR store listing (July 2025, current)
• ASME BPVC Section V product page, ASME (2025 edition, stabilized maintenance)
• ASME B31.3 – Substantive Changes in the 2016 Edition, Don Frikken, Becht, June 17, 2016 (Appendix R, Category M)
• Pressure vessel UT procedures and practices case study, Part 2, Holthaus and Gaskin, Inspectioneering Journal, Jan/Feb 2020
• Phased array ultrasound as a replacement for radiography, Evident (vendor application note)
• DAC and TVG tutorial, Evident
• ASTM E2700-26, Contact Ultrasonic Testing of Welds Using Phased Arrays, ASTM International
• ASTM E2373/E2373M-19(2024), Ultrasonic Time of Flight Diffraction (TOFD) Technique, ASTM International
• ASNT standards (SNT-TC-1A, CP-189), American Society for Nondestructive Testing
• Inspection with PAUT under ASME Section V Article 4, inspenet (secondary summary of the PAUT appendices)