Visual testing of welds to ISO 17637

Visual testing is the first and most widely used form of non-destructive testing on welds. Under EN 1090-2 it applies to one hundred percent of welds in every execution class. ISO 17637 sets out how to carry out visual testing of fusion welds: under what conditions, with which tools, at what point and how to record the results.

Because visual testing seems so obvious, it is often underestimated. A quick glance at a weld in poor light is not an examination to the standard. Applying ISO 17637 properly finds many imperfections early, limits supplementary testing to where it is really needed and produces a dossier that stands up to audit.

What ISO 17637 covers and what it does not

ISO 17637 is a test standard. It defines how visual testing is performed, not when a weld is accepted. Acceptance criteria come from another standard named in the specification or execution standard: usually ISO 5817 for steel and ISO 10042 for aluminium. General rules for selecting methods and acceptance levels are given in ISO 17635.

The standard covers fusion welds in metallic materials, in plate and pipe, and can also be applied to joint preparation before welding. Repair welds follow the same rules as the original weld.

Conditions: lighting, viewing distance and angle

The reliability of visual testing depends heavily on conditions, so ISO 17637 sets clear requirements:

• Lighting: illuminance at the surface under examination of at least 350 lux, preferably 500 lux. Workshop lighting often falls short, especially inside sections or beneath a structure, and a light meter makes this objective.
• Viewing distance: for direct visual testing the eye must be able to get within 600 mm of the surface.
• Viewing angle: no less than about 30 degrees to the surface.
• Surface condition: weld and adjacent area free of slag, spatter, oil and other contamination that could mask imperfections.

Where a weld cannot be reached this way, use remote visual testing with mirrors, borescopes or cameras, with a resolution at least equivalent to direct examination.

Tools for visual weld inspection

Visual testing is more than looking. Depending on the weld, the inspector uses:

• Weld gauges for throat and leg size, excess weld metal and undercut.
• A rule and callipers for length, alignment and linear misalignment.
• A magnifier of two to five times for fine cracks and surface porosity.
• A good lamp, ideally giving oblique light to reveal surface irregularities.
• Mirrors, borescopes or a camera for areas that are hard to reach.

Measuring tools belong in the company's calibration or verification system. A worn weld gauge gives consistently wrong readings, and with them wrongly accepted or rejected welds.

Timing of the examination

Final examination takes place on the completed weld, in the condition in which it will be assessed. If a weld is still to be ground or blasted, agree in advance whether testing happens before or after, since grinding can hide or create imperfections. Always examine before coating, or the weld can no longer be assessed.

With higher strength grades and greater thicknesses, allow for delayed hydrogen cracking. EN 1090-2 sets a minimum hold time between welding and final examination depending on steel grade, thickness and heat input.

Visual testing also pays off earlier: when checking joint preparation and during welding, for instance between runs. A defect found then is far cheaper to repair than after completion.

Reporting visual testing

Where a report is required, ISO 17637 lists what it should contain, and in practice a report is always wise. A useful report includes at least:

• identification of the object and the welds examined, for example by weld numbers from a weld map;
• material, joint type, thickness and welding process;
• the acceptance criteria and quality level applied;
• the extent of testing and the tools used;
• imperfections exceeding the acceptance limit, with their location;
• the result, the examiner's name and the date.

Also record any unusual conditions, such as restricted access or remote examination. At audit or in a handover dispute, that often makes the difference between a watertight dossier and one open to challenge.

How ISO 17637 relates to ISO 5817

ISO 17637 and ISO 5817 work as a pair: the first defines how you look, the second what you assess against. ISO 5817 sets limits for imperfections such as undercut, excess weld metal, surface porosity, incompletely filled grooves and misalignment for quality levels B, C and D. Under EN 1090-2 the level follows from the execution class, as explained in our article on ISO 5817 quality levels.

Visual testing only sees the surface. Internal defects such as lack of fusion or incomplete penetration need supplementary testing, such as ultrasonic or radiographic methods. Our article on choosing an NDT method covers which method suits which weld. For independent visual testing or structural assurance, see welding inspection and QA/QC.