Translate | ISO 6520-1:2007 Weld Imperfections, Porosity, Cracks, Fusion and Geometric Classification

To translate an ISO 6520-1:2007 weld imperfection report, you must decide whether a word describes an observed geometric feature, a classification code, a measured dimension or an acceptance decision. “Crack,” “porosity,” “lack of fusion,” “incomplete penetration,” “undercut” and “weld defect” are not interchangeable labels. This guide shows welding inspectors, multilingual fabricators, technical translators and procurement teams how to carry weld imperfection names and observations across languages without turning a classification system into an invented pass-or-fail judgement.

ISO 6520-1:2007 classifies geometric imperfections in fusion-welded metallic materials. Its official ISO catalogue record says that the classification is designed to reduce confusion and notes that metallurgical imperfections fall outside its scope. The 2007 edition remained published following a 2026 review. A separate standard such as ISO 5817:2023 can provide quality levels for certain fusion-welded joints in steel, nickel, titanium and their alloys, subject to its scope; aluminium arc-welded joints have another context in ISO 10042:2018. Naming an imperfection is not the same thing as deciding whether a particular weld is acceptable.

This welding inspection report translation article solves a narrow, high-intent question: how should an inspector’s geometric-imperfection description, identification, location, size and qualification be translated for a reader who does not share the source language? We begin with what the words actually denote, build a practical terminology and inspection-evidence map, examine realistic reporting errors, and close with a review checklist, worked cases, four weeks of exercises and FAQs. The examples are educational, not engineering acceptance criteria. Qualified welding personnel and actual contract documents govern repair, acceptance and safety decisions.

Reading route for technical translators

Classification versus acceptance · Seven-pass document method · Important weld-imperfection terms · Worked examples · Final audit · Four-week learning plan · FAQs · Official references and related learning.

A classification system is not a quality verdict

Picture a fabrication report that contains a diagram, an indication identifier W-07, a term such as “lack of fusion,” a measured indication length and a separate comment about acceptance. The shape described by the classification term, the length recorded by the inspector and the acceptance criterion applied by the project are different pieces of information. A translator who converts “imperfection found” into “defective welding rejected” has crossed a boundary: identification of a feature has been turned into a disposition decision. The reverse is also possible. Changing “crack observed” into “minor surface irregularity” can conceal a serious finding.

The public ISO record presents ISO 6520-1 as a scheme for precise identification and description of fusion-weld geometric imperfections. It distinguishes that purpose from metallurgy. One may find heat-treatment effects, grain-size issues and hardness observations in a broader weld file, but they are not made into ISO 6520-1 geometric categories simply because they appear on the same inspection sheet. A truthful translation keeps the classification vocabulary tied to the geometry that the inspector identified.

Actual welding quality levels are established under the appropriate separate method, material, product and contract requirements. ISO 5817:2023, for example, describes three quality levels B, C and D within its stated scope, with B the most demanding. Those are not numerical imperfection codes in ISO 6520-1. A report might cite classification nomenclature from one source and acceptance from another. It is a serious translation error to merge their references or use a quality letter as though it named the geometric shape.

The translation is also a spatial exercise. A discontinuity may be at the weld root, the toe, the surface, the fusion boundary or a location identified by drawing coordinates. Translation must preserve whether a feature is internal or surface-breaking only when the inspection evidence actually establishes that distinction. A radiographic indication, an ultrasonic response and a visual observation are different evidence channels. The word “indication” may be deliberately more cautious than a confirmed imperfection classification. Do not upgrade preliminary evidence into a definitive defect name.

For cross-border procurement, the most reliable method is to reconstruct the report as a set of linked fields: welding process, joint, material, location, inspection technique, observed feature, classification, dimension, applicable quality requirement, recorded disposition and approval signature. Each field gets translated for what it says, not for the overall message a seller or purchaser hopes to communicate. This can be done in clear, conversational English without losing the technical limitations.

A seven-pass translation method for weld documentation

Pass 1: Identify the exact source document. Establish whether you are reading a weld map, shop observation, NDT indication record, final inspection certificate, repair instruction or acceptance sheet. They may mention the same weld but have different authority. Keep a source-to-target ledger keyed by report and weld identifier. It should be possible to trace every decisive noun, numerical value and conclusion back to one entry. If the source does not contain an acceptance status or a measured depth, leave it unasserted rather than filling the gap with reasonable-sounding professional language.

Pass 2: Record the classification edition. Retain ISO 6520-1:2007 where that is the edition actually used. Do not assume a revised ISO 5817 citation replaces the classification source. Version control matters because readers must be able to reconstruct the original language and codes. Keep a source-to-target ledger keyed by report and weld identifier. It should be possible to trace every decisive noun, numerical value and conclusion back to one entry. If the source does not contain an acceptance status or a measured depth, leave it unasserted rather than filling the gap with reasonable-sounding professional language.

Pass 3: Locate the feature. Tie every imperfection phrase to weld ID, joint ID, position and dimensional reference. A term without its location can be attached to the wrong component when the translated drawing is assembled. Keep a source-to-target ledger keyed by report and weld identifier. It should be possible to trace every decisive noun, numerical value and conclusion back to one entry. If the source does not contain an acceptance status or a measured depth, leave it unasserted rather than filling the gap with reasonable-sounding professional language.

Pass 4: Keep detection separate from identification. Determine whether an inspection technique produced an indication, whether it was subsequently characterised, and whether the final reviewer assigned a geometric class. Translate all three states with their own evidentiary status. Keep a source-to-target ledger keyed by report and weld identifier. It should be possible to trace every decisive noun, numerical value and conclusion back to one entry. If the source does not contain an acceptance status or a measured depth, leave it unasserted rather than filling the gap with reasonable-sounding professional language.

Pass 5: Read measurements literally. Check mm versus µm, indication length versus depth, surface extent versus total weld length, and whether a value is maximum, average or nominal. A correctly spelled word beside a wrong measurement can be dangerous. Keep a source-to-target ledger keyed by report and weld identifier. It should be possible to trace every decisive noun, numerical value and conclusion back to one entry. If the source does not contain an acceptance status or a measured depth, leave it unasserted rather than filling the gap with reasonable-sounding professional language.

Pass 6: Find acceptance elsewhere. If an acceptance criterion is present, identify its source and material scope. Do not translate a classification code into a rejection threshold. Never infer approval from an absent defect note. Keep a source-to-target ledger keyed by report and weld identifier. It should be possible to trace every decisive noun, numerical value and conclusion back to one entry. If the source does not contain an acceptance status or a measured depth, leave it unasserted rather than filling the gap with reasonable-sounding professional language.

Pass 7: Reverse-check a sample. Ask a bilingual reviewer and qualified welding reviewer to reconstruct the original finding, location, size, method, classification, disposition and unresolved questions. Resolve differences before final publication. Keep a source-to-target ledger keyed by report and weld identifier. It should be possible to trace every decisive noun, numerical value and conclusion back to one entry. If the source does not contain an acceptance status or a measured depth, leave it unasserted rather than filling the gap with reasonable-sounding professional language.

Field-by-field welding vocabulary and translation traps

1. Fusion welding

The ISO classification addresses fusion welding of metallic materials. The welding process and joint type provide context for interpreting the imperfection description. To preserve meaning, build a line that identifies the geometry, the weld and the evidence source. Technical vocabulary is not a decorative synonym list: it is the interface between the inspector’s physical observation and the reader’s decision. If any element is missing, a polished sentence can still be wrong.

Misleading shortcut: Extending the label to all joining processes without checking scope. Faithful approach: Preserve the actual process or procedure reference reported by the fabrication record. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A shop certificate combines fusion welding and a different joining process under one generic heading. Write one precise table caption and one natural explanation to a project manager. Evidence check: Verify the WPS or project welding process where cited. Try a reverse reading without the image. Can the next reviewer identify the same feature and where it lies? If not, improve the caption rather than adding a conclusion.

2. Geometric imperfection

A geometric imperfection describes an observed departure in weld shape or continuity. Naming it does not itself determine that the weld is unacceptable. This entry is a test of source authority. The same everyday word can appear in a technician’s comment, a formal NDT report and a signed acceptance sheet with different levels of finality. Translating them all in one assertive voice loses that difference.

Misleading shortcut: Automatically translating “imperfection” as “defective and rejected.” Faithful approach: Use “geometric imperfection” or the specific feature name, reserving “rejected” for a documented decision. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: An inspection table lists a feature with no final evaluation column. Write one precise table caption and one natural explanation to a project manager. Evidence check: Look for a separately signed acceptance or disposition entry. Mark the author and date beside the term in your bilingual ledger. Separate what was observed from what a named reviewer decided.

3. Metallurgical imperfection

Metallurgical conditions concern material structure and behaviour, and the public ISO scope explicitly excludes them from this classification. Do not let the apparent simplicity of a short label mislead you. A one-word geometrical term can depend on orientation, cross-section and inspection method. Use the adjacent sketch, legend and numeric fields to resolve the intended feature before selecting target-language wording.

Misleading shortcut: Labelling a hardness or microstructure result as a listed geometric ISO 6520-1 code. Faithful approach: Translate metallurgical observations according to their actual method and report. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A weld file contains a micrograph and an ultrasonic scan. Write one precise table caption and one natural explanation to a project manager. Evidence check: Keep microstructure evidence separate from geometric findings. Create one accurate engineer-facing label and a separate plain-English explanation. The two should agree about location, physical form and certainty.

4. Classification identifier

A classification designation is a stable way of referring to the type of imperfection. Its digits or symbols must be copied exactly as provided. Numbers make a statement look authoritative even after a noun has been translated incorrectly. Link each measurement to its own property: length, depth, height, angle or position. This avoids the very common mistake of preserving the correct digits beside a different physical dimension.

Misleading shortcut: Inventing a code from a remembered generic list or changing a digit in layout. Faithful approach: Retain the source code and its stated description together; query inconsistencies. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A report references a numeric classification but the scan is partly illegible. Write one precise table caption and one natural explanation to a project manager. Evidence check: Check the licensed method and original digital record. Audit the final file after export. A lost degree symbol or an ambiguous slash in a unit can invalidate an otherwise careful translation.

5. Imperfection versus defect

An imperfection is a descriptive term; “defect” often carries a judgement that an applicable requirement was not met. A disciplined translator is prepared to leave a question unanswered. Welding decisions can have consequences, but translation does not confer authority to diagnose a flaw from a photo or to declare a fabricated structure acceptable. Preserve the record and route unresolved technical choices to the responsible qualified personnel.

Misleading shortcut: Calling every recorded imperfection a defect before acceptance has been assessed. Faithful approach: Prefer the descriptive term until a sourced rejection or defect classification is given. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A quality summary uses “defect” casually while the certified inspection report says “imperfection.” Write one precise table caption and one natural explanation to a project manager. Evidence check: Identify whose status statement has authority. Use a small issue log naming the weld ID, phrase, evidence and question. The reviewer should answer by citing a source, not by providing reassurance.

6. Indication

NDT instruments and visual reviews can identify indications that require interpretation. An indication is not necessarily a proven crack or cavity. To preserve meaning, build a line that identifies the geometry, the weld and the evidence source. Technical vocabulary is not a decorative synonym list: it is the interface between the inspector’s physical observation and the reader’s decision. If any element is missing, a polished sentence can still be wrong.

Misleading shortcut: Translating all indications as confirmed physical defects. Faithful approach: Use “indication detected” unless the inspector actually identified its type. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: An ultrasonic response appears at a given location but is not classified. Write one precise table caption and one natural explanation to a project manager. Evidence check: Read the final NDT interpretation and disposition. Try a reverse reading without the image. Can the next reviewer identify the same feature and where it lies? If not, improve the caption rather than adding a conclusion.

7. Crack

A crack is a discontinuity described by the inspector; it must not be softened into ordinary surface texture in translation. This entry is a test of source authority. The same everyday word can appear in a technician’s comment, a formal NDT report and a signed acceptance sheet with different levels of finality. Translating them all in one assertive voice loses that difference.

Misleading shortcut: Replacing “crack” with “mark” or “scratch” to make a report sound less alarming. Faithful approach: Keep the explicit crack designation, site and observed extent. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A visual record states surface crack at an identified weld toe. Write one precise table caption and one natural explanation to a project manager. Evidence check: Confirm observation location and any subsequent examination. Mark the author and date beside the term in your bilingual ledger. Separate what was observed from what a named reviewer decided.

8. Longitudinal crack

Orientation relative to the weld axis matters in describing this kind of crack. A translation must preserve the stated direction. Do not let the apparent simplicity of a short label mislead you. A one-word geometrical term can depend on orientation, cross-section and inspection method. Use the adjacent sketch, legend and numeric fields to resolve the intended feature before selecting target-language wording.

Misleading shortcut: Using transverse when the source means along the weld. Faithful approach: Write “longitudinal crack” with the weld reference direction intact. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A sketch arrow is parallel to the seam while the text is abbreviated. Write one precise table caption and one natural explanation to a project manager. Evidence check: Compare the original diagram and legend. Create one accurate engineer-facing label and a separate plain-English explanation. The two should agree about location, physical form and certainty.

9. Transverse crack

A transverse description refers to orientation across the weld direction; the term should follow the source geometry. Numbers make a statement look authoritative even after a noun has been translated incorrectly. Link each measurement to its own property: length, depth, height, angle or position. This avoids the very common mistake of preserving the correct digits beside a different physical dimension.

Misleading shortcut: Treating all cracks as directional synonyms. Faithful approach: Translate “transverse crack” and retain coordinate information. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: The crack is drawn crossing the weld bead rather than tracking along it. Write one precise table caption and one natural explanation to a project manager. Evidence check: Check the axis in the inspection sketch. Audit the final file after export. A lost degree symbol or an ambiguous slash in a unit can invalidate an otherwise careful translation.

10. Crater crack

A crater-associated crack is related to a particular part of the deposited weld geometry as identified by the inspector. A disciplined translator is prepared to leave a question unanswered. Welding decisions can have consequences, but translation does not confer authority to diagnose a flaw from a photo or to declare a fabricated structure acceptable. Preserve the record and route unresolved technical choices to the responsible qualified personnel.

Misleading shortcut: Replacing crater crack with generic crater depression or incomplete fill. Faithful approach: Preserve the source’s specific crack terminology. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A weld-end record distinguishes a crack from a shallow termination depression. Write one precise table caption and one natural explanation to a project manager. Evidence check: Review the photographic evidence and inspector note. Use a small issue log naming the weld ID, phrase, evidence and question. The reviewer should answer by citing a source, not by providing reassurance.

11. Cavity

A cavity is a void or hollow feature; the term is not a generic synonym for every NDT indication. To preserve meaning, build a line that identifies the geometry, the weld and the evidence source. Technical vocabulary is not a decorative synonym list: it is the interface between the inspector’s physical observation and the reader’s decision. If any element is missing, a polished sentence can still be wrong.

Misleading shortcut: Calling a solid inclusion a cavity or assuming any cavity is gas porosity. Faithful approach: Use the exact named void or cavity class when confirmed. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: An image shows a dark feature but has no final classification. Write one precise table caption and one natural explanation to a project manager. Evidence check: Ask whether the observation has been identified as a void. Try a reverse reading without the image. Can the next reviewer identify the same feature and where it lies? If not, improve the caption rather than adding a conclusion.

12. Gas porosity

Porosity relates to pores or voids in the weld; details such as distribution and location may matter in inspection reporting. This entry is a test of source authority. The same everyday word can appear in a technician’s comment, a formal NDT report and a signed acceptance sheet with different levels of finality. Translating them all in one assertive voice loses that difference.

Misleading shortcut: Translating pores as solid slag or non-metallic inclusions. Faithful approach: Use “porosity” for the confirmed pore-related finding. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A report uses a short symbol for distributed pore indications. Write one precise table caption and one natural explanation to a project manager. Evidence check: Check the inspector’s legend and evidence basis. Mark the author and date beside the term in your bilingual ledger. Separate what was observed from what a named reviewer decided.

13. Clustered porosity

A group or cluster of pores is a description of distribution, not just the presence of one pore. Do not let the apparent simplicity of a short label mislead you. A one-word geometrical term can depend on orientation, cross-section and inspection method. Use the adjacent sketch, legend and numeric fields to resolve the intended feature before selecting target-language wording.

Misleading shortcut: Rendering a cluster as a single isolated hole. Faithful approach: Retain the grouping and any measured extent. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: An inspector marks multiple adjacent pores within an area. Write one precise table caption and one natural explanation to a project manager. Evidence check: Verify the source grouping language and area definition. Create one accurate engineer-facing label and a separate plain-English explanation. The two should agree about location, physical form and certainty.

14. Linear porosity

An elongated arrangement of pores should not be confused with a single continuous elongated void without confirming the observation. Numbers make a statement look authoritative even after a noun has been translated incorrectly. Link each measurement to its own property: length, depth, height, angle or position. This avoids the very common mistake of preserving the correct digits beside a different physical dimension.

Misleading shortcut: A translated description collapses alignment into cavity shape. Faithful approach: State “linearly distributed porosity” when the original supports it. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A weld map shows a sequence of separate aligned indications. Write one precise table caption and one natural explanation to a project manager. Evidence check: Compare classification text with the diagram. Audit the final file after export. A lost degree symbol or an ambiguous slash in a unit can invalidate an otherwise careful translation.

15. Solid inclusion

An inclusion is material trapped within the weld; the term differs from a void or a lack-of-fusion interface. A disciplined translator is prepared to leave a question unanswered. Welding decisions can have consequences, but translation does not confer authority to diagnose a flaw from a photo or to declare a fabricated structure acceptable. Preserve the record and route unresolved technical choices to the responsible qualified personnel.

Misleading shortcut: Using porosity whenever a scan shows an internal indication. Faithful approach: Keep inclusion material and classification when given. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: An inspection conclusion specifies a solid non-metallic inclusion. Write one precise table caption and one natural explanation to a project manager. Evidence check: Confirm that the report identifies the material rather than speculating. Use a small issue log naming the weld ID, phrase, evidence and question. The reviewer should answer by citing a source, not by providing reassurance.

16. Slag inclusion

Slag inclusion refers to retained slag material and should be distinguished from other solid inclusions or surface slag residue. To preserve meaning, build a line that identifies the geometry, the weld and the evidence source. Technical vocabulary is not a decorative synonym list: it is the interface between the inspector’s physical observation and the reader’s decision. If any element is missing, a polished sentence can still be wrong.

Misleading shortcut: Assuming a visible surface deposit proves an internal inclusion. Faithful approach: Use the inspector’s confirmed “slag inclusion” wording. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A weld process creates slag, but the NDT record only says “indication.” Write one precise table caption and one natural explanation to a project manager. Evidence check: Check whether the final identification actually names slag. Try a reverse reading without the image. Can the next reviewer identify the same feature and where it lies? If not, improve the caption rather than adding a conclusion.

17. Tungsten inclusion

A tungsten inclusion is a specific classification and should not be translated as generic debris or slag. This entry is a test of source authority. The same everyday word can appear in a technician’s comment, a formal NDT report and a signed acceptance sheet with different levels of finality. Translating them all in one assertive voice loses that difference.

Misleading shortcut: Replacing the metal name with a familiar welding consumable. Faithful approach: Retain “tungsten inclusion” when specifically documented. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A method note contains the term adjacent to another inclusion class. Write one precise table caption and one natural explanation to a project manager. Evidence check: Check the recorded process and classification. Mark the author and date beside the term in your bilingual ledger. Separate what was observed from what a named reviewer decided.

18. Lack of fusion

Lack of fusion concerns missing joining at a required fusion interface; it is not identical to insufficient penetration. Do not let the apparent simplicity of a short label mislead you. A one-word geometrical term can depend on orientation, cross-section and inspection method. Use the adjacent sketch, legend and numeric fields to resolve the intended feature before selecting target-language wording.

Misleading shortcut: Treating every incomplete fusion observation as incomplete root penetration. Faithful approach: Name the fusion boundary or surface where the source identifies it. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A report distinguishes sidewall lack of fusion from a root-related indication. Write one precise table caption and one natural explanation to a project manager. Evidence check: Check the drawing and inspector conclusion. Create one accurate engineer-facing label and a separate plain-English explanation. The two should agree about location, physical form and certainty.

19. Incomplete penetration

Penetration terminology concerns weld development through the joint geometry and is different from fusion at every interface. Numbers make a statement look authoritative even after a noun has been translated incorrectly. Link each measurement to its own property: length, depth, height, angle or position. This avoids the very common mistake of preserving the correct digits beside a different physical dimension.

Misleading shortcut: Translating incomplete penetration as inadequate overall weld size without evidence. Faithful approach: Preserve the original penetration phrase and joint location. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A butt-weld record refers to a root condition. Write one precise table caption and one natural explanation to a project manager. Evidence check: Read the joint preparation and section drawing. Audit the final file after export. A lost degree symbol or an ambiguous slash in a unit can invalidate an otherwise careful translation.

20. Weld root

The root is a specific region of joint geometry; it is not the cap or outer weld face. A disciplined translator is prepared to leave a question unanswered. Welding decisions can have consequences, but translation does not confer authority to diagnose a flaw from a photo or to declare a fabricated structure acceptable. Preserve the record and route unresolved technical choices to the responsible qualified personnel.

Misleading shortcut: Confusing root indications with visible face irregularities. Faithful approach: Keep root, face and toe separate in the bilingual key. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: An inspector’s note says “root” while the annotated image shows the back side. Write one precise table caption and one natural explanation to a project manager. Evidence check: Compare the weld cross-section and report notation. Use a small issue log naming the weld ID, phrase, evidence and question. The reviewer should answer by citing a source, not by providing reassurance.

21. Undercut

Undercut describes a groove-related geometric feature near the weld toe or root in the applicable context. To preserve meaning, build a line that identifies the geometry, the weld and the evidence source. Technical vocabulary is not a decorative synonym list: it is the interface between the inspector’s physical observation and the reader’s decision. If any element is missing, a polished sentence can still be wrong.

Misleading shortcut: Using “cut” to imply a deliberate machining operation. Faithful approach: Use the technical noun “undercut” with measured location and extent. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A seam report records a toe groove with depth. Write one precise table caption and one natural explanation to a project manager. Evidence check: Check the measuring basis and illustration. Try a reverse reading without the image. Can the next reviewer identify the same feature and where it lies? If not, improve the caption rather than adding a conclusion.

22. Overlap

Overlap is associated with excess metal extending without the required fusion relationship; it is not a synonym for reinforcement. This entry is a test of source authority. The same everyday word can appear in a technician’s comment, a formal NDT report and a signed acceptance sheet with different levels of finality. Translating them all in one assertive voice loses that difference.

Misleading shortcut: Calling a smoothly reinforced bead an overlap solely because it projects outward. Faithful approach: Use the inspector’s actual identified feature. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: The record distinguishes an edge overlap from excess cap height. Write one precise table caption and one natural explanation to a project manager. Evidence check: Refer to the observation rather than guess from appearance. Mark the author and date beside the term in your bilingual ledger. Separate what was observed from what a named reviewer decided.

23. Excess reinforcement

Reinforcement refers to weld metal above the reference surface and may have a measured amount, subject to joint and method. Do not let the apparent simplicity of a short label mislead you. A one-word geometrical term can depend on orientation, cross-section and inspection method. Use the adjacent sketch, legend and numeric fields to resolve the intended feature before selecting target-language wording.

Misleading shortcut: Describing all reinforcement as a defect or using the total weld height as the excess. Faithful approach: Keep “excess reinforcement” and its measured value where the source reports it. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A cap measurement is quoted in millimetres. Write one precise table caption and one natural explanation to a project manager. Evidence check: Check what baseline the inspection used. Create one accurate engineer-facing label and a separate plain-English explanation. The two should agree about location, physical form and certainty.

24. Root concavity

Root concavity is a geometry description; it should not be translated as full penetration failure without supporting evidence. Numbers make a statement look authoritative even after a noun has been translated incorrectly. Link each measurement to its own property: length, depth, height, angle or position. This avoids the very common mistake of preserving the correct digits beside a different physical dimension.

Misleading shortcut: Replacing a concave profile with incomplete penetration as though they were equivalent. Faithful approach: Translate the shape and depth exactly as identified. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A section view shows a concave root but no declared fusion defect. Write one precise table caption and one natural explanation to a project manager. Evidence check: Check the precise feature classification. Audit the final file after export. A lost degree symbol or an ambiguous slash in a unit can invalidate an otherwise careful translation.

25. Misalignment

Misalignment concerns relative positioning of joined components, not a flaw in the weld metal composition. A disciplined translator is prepared to leave a question unanswered. Welding decisions can have consequences, but translation does not confer authority to diagnose a flaw from a photo or to declare a fabricated structure acceptable. Preserve the record and route unresolved technical choices to the responsible qualified personnel.

Misleading shortcut: Translating an offset as an angular distortion or a length defect. Faithful approach: Keep linear offset and angular distortion separately named. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: Two plates are vertically offset at the joint. Write one precise table caption and one natural explanation to a project manager. Evidence check: Check the defined reference planes and measurement direction. Use a small issue log naming the weld ID, phrase, evidence and question. The reviewer should answer by citing a source, not by providing reassurance.

26. Angular distortion

Angular change describes relative rotation or shape after joining and is not numerically interchangeable with a linear offset. To preserve meaning, build a line that identifies the geometry, the weld and the evidence source. Technical vocabulary is not a decorative synonym list: it is the interface between the inspector’s physical observation and the reader’s decision. If any element is missing, a polished sentence can still be wrong.

Misleading shortcut: Changing an angle in degrees into a millimetre value. Faithful approach: Preserve the measured angle and coordinate reference. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A fabrication report records an angular departure from the drawing. Write one precise table caption and one natural explanation to a project manager. Evidence check: Check the reference dimension and unit. Try a reverse reading without the image. Can the next reviewer identify the same feature and where it lies? If not, improve the caption rather than adding a conclusion.

27. Spatter

Spatter describes unwanted deposited particles; it is distinct from an internal slag inclusion or weld porosity. This entry is a test of source authority. The same everyday word can appear in a technician’s comment, a formal NDT report and a signed acceptance sheet with different levels of finality. Translating them all in one assertive voice loses that difference.

Misleading shortcut: Calling all visible particles “internal inclusions.” Faithful approach: Use “weld spatter” if the inspector recorded external deposits. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A visual inspection lists particles adjacent to the seam. Write one precise table caption and one natural explanation to a project manager. Evidence check: Review whether the location is surface or internal. Mark the author and date beside the term in your bilingual ledger. Separate what was observed from what a named reviewer decided.

28. Measurement of length

Indication length, continuous flaw length and total affected weld length may differ; the translation must reproduce the source’s measurement basis. Do not let the apparent simplicity of a short label mislead you. A one-word geometrical term can depend on orientation, cross-section and inspection method. Use the adjacent sketch, legend and numeric fields to resolve the intended feature before selecting target-language wording.

Misleading shortcut: Quoting an overall weld length as the imperfection length. Faithful approach: State which feature dimension is measured in mm. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A 2 m seam contains a recorded 12 mm indication. Write one precise table caption and one natural explanation to a project manager. Evidence check: Check the drawing scale and dimension label. Create one accurate engineer-facing label and a separate plain-English explanation. The two should agree about location, physical form and certainty.

29. Depth and height

Depth, height, throat dimension and cap reinforcement represent different geometrical quantities. They should not share a generic “size” label if the report distinguishes them. Numbers make a statement look authoritative even after a noun has been translated incorrectly. Link each measurement to its own property: length, depth, height, angle or position. This avoids the very common mistake of preserving the correct digits beside a different physical dimension.

Misleading shortcut: Inverting measured depth and protrusion height. Faithful approach: Keep axis, reference plane, symbol and unit explicit. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: An inspection form has two millimetre columns with similar headings. Write one precise table caption and one natural explanation to a project manager. Evidence check: Use the original symbol key to confirm both. Audit the final file after export. A lost degree symbol or an ambiguous slash in a unit can invalidate an otherwise careful translation.

30. Weld map and indication ID

A weld map joins findings to a real joint; an indication ID may recur on different sheets unless linked to a report. A disciplined translator is prepared to leave a question unanswered. Welding decisions can have consequences, but translation does not confer authority to diagnose a flaw from a photo or to declare a fabricated structure acceptable. Preserve the record and route unresolved technical choices to the responsible qualified personnel.

Misleading shortcut: Dropping the weld ID while preserving only the local finding number. Faithful approach: Translate location labels and copy unambiguous identifiers unchanged. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: A plant has two joints each containing an indication numbered 01. Write one precise table caption and one natural explanation to a project manager. Evidence check: Use document number, weld number and finding number together. Use a small issue log naming the weld ID, phrase, evidence and question. The reviewer should answer by citing a source, not by providing reassurance.

31. Evaluation and disposition

A final report may record accepted, rejected, repair proposed or pending investigation. Those statuses come from the applicable criteria and authority. To preserve meaning, build a line that identifies the geometry, the weld and the evidence source. Technical vocabulary is not a decorative synonym list: it is the interface between the inspector’s physical observation and the reader’s decision. If any element is missing, a polished sentence can still be wrong.

Misleading shortcut: Turning a finding into an acceptance decision when no such decision was recorded. Faithful approach: State the exact disposition and reference when present; otherwise describe the evaluation as unresolved. The distinction matters when a note is removed from the inspection sheet and placed in a multilingual fabrication dossier. A classification label should travel with its location and source context, and a finding should not silently become an acceptance verdict.

Training scenario: An NDT sheet has a finding but an empty disposition box. Write one precise table caption and one natural explanation to a project manager. Evidence check: Seek the authorised inspection or quality sign-off. Try a reverse reading without the image. Can the next reviewer identify the same feature and where it lies? If not, improve the caption rather than adding a conclusion.

Worked weld-report cases: how words change decisions

1. An indication is not automatically a crack

Ultrasonic record U-17 shows a response at a coordinate, but the inspection sheet labels it “unclassified indication.” A translator familiar with welding may believe a crack is likely. Nevertheless “crack detected” is a stronger and potentially different finding. Translate “unclassified indication at the stated location” and preserve any follow-up requirement. A later signed report may identify the nature of the imperfection, in which case that subsequent identification should be attributed to its own dated record.

The handover summary should not collapse the sequence. A useful two-sentence version is: “An unclassified indication was recorded in weld W-17. Classification and disposition are subject to the inspector’s subsequent evaluation.” Do not invent a final result to make the report feel complete. This is a transferable lesson in evidence language, but its application here is specifically to fusion-weld geometric inspection.

For practice, produce a three-line report: finding, evidence and outstanding decision. Ask a colleague whether any line converts an inspection indication into a proven imperfection or a classification into a rejection criterion without authority. Then revise the prose until the original weld identity, dimensions and result status are all recoverable. These exercises are illustrative, not substitute instructions for actual weld inspection or repair.

2. Porosity does not have one universal acceptance result

Weld W-21 has a finding described as clustered porosity, with an extent shown in millimetres. The source cites ISO 6520-1 for identification but has no stated acceptance level. A careless translation might write “ISO 6520-1 failed.” The designation alone cannot support that. A faithful report identifies the observed cluster, weld location, dimensions and source classification, and says nothing about pass/fail until the relevant criterion and disposition are available.

If the customer contract cites ISO 5817:2023 and a specific level for an applicable material and joint, that may become relevant to an authorised evaluation. The translator should not derive the decision without the full context, nor should they confuse a quality level such as B, C or D with a geometric imperfection code. Ask which project specification and reviewer issued the status.

3. Lack of fusion versus incomplete penetration

Imagine a report listing one lack-of-fusion indication on a weld sidewall and a separate note on incomplete root penetration. Both involve a joining relationship, but they identify different aspects of geometry. Translating them both as “poor weld bonding” removes the distinction the inspector deliberately recorded. The translation should retain sidewall and root references, individual indication IDs and original observations.

Write a two-column comparison: one row for the lack-of-fusion finding and one for the incomplete-penetration finding. Do not assign the same code to both unless the signed report actually does. This preserves the possibility that their measurement and acceptance treatments differ under the applicable evaluation framework. Technical clarity is sometimes less elegant than a single broad noun, but it is safer.

4. Length and depth are not synonyms

One inspection table lists a discontinuity 14 mm long and 1.2 mm deep. A layout conversion moves the “deep” label to the first number and changes the meaning. If a buyer copies the translated dimensions into an evaluation spreadsheet, the mistake has consequences far beyond style. The safest workflow uses a symbol map, the original inspection image and a final rendered table audit, not an assumption that numeric order remained unchanged.

Explain to a non-specialist that the two figures measure different directions of the same finding. In the translated table, keep column headers unambiguous and use a note if the source reference plane is unclear. A translator cannot fix a missing measurement by estimating from a small photograph. Uncertainty should remain visible until the inspector clarifies it.

5. Classification, quality level and repair disposition

A shop file contains three documents: geometric imperfection classification under ISO 6520-1:2007, a specified quality level under an applicable quality standard, and a repair disposition signed by the authorised reviewer. The translator must maintain the boundaries among them. It is quite possible for a feature to be classified in one document, judged in another and assigned a repair action in a third. None of those steps should be silently attributed to the wrong authority.

A reliable executive summary might read: “The indicated feature was classified under the geometric-imperfection vocabulary; its acceptance was evaluated against the referenced project criteria; the authorised disposition appears in the separate repair record.” Include exact documents and dates when available. Never substitute this generic example for an actual project conclusion. The point is to show readers how an evidence chain remains intact across languages.

Final publication and technical-review checklist

Copy ISO 6520-1:2007 and all reported classification identifiers exactly.
Distinguish an observation, an indication, a confirmed geometric imperfection and a defect judgement.
Preserve weld ID, joint, position, surface or internal status when established, and report reference.
Check every crack orientation, root/face/toe location and fusion-versus-penetration distinction.
Reconcile length, depth, height, angle, dimensions and units against the original sketch.
Keep NDT technique and the basis of the identification separate from the result.
Do not treat classification as pass/fail; identify the actual applicable acceptance criteria.
Do not confuse ISO 5817 quality letters with ISO 6520-1 classification codes.
Keep repair, rework, accepted, rejected and pending dispositions attributed to the authorised source.
Test PDF layout for superscript, symbols, numbering, drawing callouts and detached notes.

The final check should include a person who knows the target language and a person qualified to assess the welding evidence. The language expert asks whether a foreign reader would understand the same facts; the welding reviewer asks whether the chosen nouns, locations and sizes truly describe the source record. Store any signed clarification. A translation that has passed spelling review but not source-based technical review can still put the right number beside the wrong phenomenon.

Four-week learning plan

Week 1 — learn the geometry

Create labelled sketches for root, face, toe, fusion interface, weld axis and plate alignment. Associate each with a source vocabulary entry and an ordinary-language explanation. Practise distinguishing a crack from porosity, an inclusion from a cavity and lack of fusion from incomplete penetration without turning the exercise into an inspection diagnosis. Record one successful correction and one unresolved technical question each week. The exercise is complete when a reader can recover the same geometry, evidence and stated level of certainty without being told the intended conclusion in advance.

Week 2 — learn the documentary chain

Read three mock documents: a weld map, an NDT indication list and a final evaluation record. Mark the authority and status of each statement. Translate ten short labels, then write the exact question you would ask when a required dimension or disposition is missing. Keep codes and IDs unchanged. Record one successful correction and one unresolved technical question each week. The exercise is complete when a reader can recover the same geometry, evidence and stated level of certainty without being told the intended conclusion in advance.

Week 3 — learn units and acceptance boundaries

Prepare a measurement grid for length, depth, height and angle. Deliberately transpose two units, then design a reverse-check that catches the error. Compare the purpose of ISO 6520-1 classification with that of a relevant quality-level standard, without inventing any acceptance criterion or numeric tolerance. Record one successful correction and one unresolved technical question each week. The exercise is complete when a reader can recover the same geometry, evidence and stated level of certainty without being told the intended conclusion in advance.

Week 4 — produce an auditable multilingual report

Create a sample four-page report with drawings, findings, method labels, disposition references and an issue log. Invite one bilingual reviewer and one welding specialist to check it independently. Use their comments to revise the captions, then confirm the final HTML or PDF still carries the drawing references and measurement units correctly. Record one successful correction and one unresolved technical question each week. The exercise is complete when a reader can recover the same geometry, evidence and stated level of certainty without being told the intended conclusion in advance.

Frequently asked questions

What is ISO 6520-1:2007 used for?

It supports precise classification and description of geometric imperfections in fusion-welded metallic materials, not a universal acceptance verdict. If the source does not document a classification, size or disposition, leave it as a clearly recorded question. A truthful translation need not manufacture a complete answer.

Does an imperfection mean the weld failed?

Not automatically. An evaluation depends on the applicable specification, actual geometry and authorised assessment. If the source does not document a classification, size or disposition, leave it as a clearly recorded question. A truthful translation need not manufacture a complete answer.

How is a crack different from porosity?

They describe different physical forms. Preserve the actual inspector’s identification, rather than choosing the most familiar label. If the source does not document a classification, size or disposition, leave it as a clearly recorded question. A truthful translation need not manufacture a complete answer.

What is the difference between lack of fusion and incomplete penetration?

They refer to different joining and joint-geometry situations. The translation should retain the precise named feature and its location. If the source does not document a classification, size or disposition, leave it as a clearly recorded question. A truthful translation need not manufacture a complete answer.

Should I change an illegible numeric code to a plausible one?

No. Keep the source uncertainty, seek the original digital record or ask the document owner. Guessing creates a potentially false classification. If the source does not document a classification, size or disposition, leave it as a clearly recorded question. A truthful translation need not manufacture a complete answer.

Does ISO 5817 replace ISO 6520-1?

They serve different purposes: classification versus quality-level requirements within their respective scopes. Preserve the standards actually cited. If the source does not document a classification, size or disposition, leave it as a clearly recorded question. A truthful translation need not manufacture a complete answer.

What does a quality level B mean?

Under ISO 5817:2023, within its scope, level B corresponds to the highest of its specified quality requirements. It is not an ISO 6520-1 geometric code. If the source does not document a classification, size or disposition, leave it as a clearly recorded question. A truthful translation need not manufacture a complete answer.

Can I assess a weld from the photograph in the translation file?

No. Photographs can help locate terms, but actual inspection and fitness-for-purpose decisions require the applicable methods and qualified personnel. If the source does not document a classification, size or disposition, leave it as a clearly recorded question. A truthful translation need not manufacture a complete answer.

Why does the material type matter?

Acceptance standards have different material and process scopes. A claim about steel joints should not automatically be transferred to aluminium or another joining process. If the source does not document a classification, size or disposition, leave it as a clearly recorded question. A truthful translation need not manufacture a complete answer.

Who signs off the translation?

Have a bilingual reviewer verify meaning and a responsible welding specialist confirm technical statements that affect inspection, acceptance or repair. If the source does not document a classification, size or disposition, leave it as a clearly recorded question. A truthful translation need not manufacture a complete answer.

Official sources and wider learning routes

Primary references: ISO 6520-1:2007 classification — ISO catalogue; ISO 5817:2023 quality levels — ISO catalogue; ISO 10042:2018 aluminium arc-welded joints — ISO catalogue. The public abstracts support the purpose, edition and scope distinctions used here. The full licensed standards, actual inspection records and applicable contracts must be consulted for authoritative code detail and acceptance decisions. No proprietary tables or numerical acceptance limits are reproduced.

Vocabulary Mastery | The Complete System — general terminology acquisition and memory, maintained as the protected vocabulary owner.
How English Works | Translation and Meaning Across Languages — grammar and semantic clarity across languages.
Master Art of Translation | Evidence and Operational Records — an example of the wider evidence-preservation architecture, not a competing weld terminology page.

A useful final exercise is to write two lines: “The inspector observed …” and “The authorised evaluation concluded …”. If the source supports only the first, stop there. ISO 6520-1 terminology is most useful when it helps readers name a geometric finding precisely while leaving acceptance decisions to the evidence and authorities that actually govern them.