Standards guide · Quality & documents
Test methods explained: destructive, non-destructive – and how much is tested
Behind every inspection certificate stand concrete tests: tensile specimens that are pulled apart, dye penetrant from the spray can, ultrasonic probes and radiographs of the weld. This page sorts the methods: what is tested destructively on specimens taken from the product, what is tested non-destructively on the component – and why the question “what percentage is radiographed?” is always a question of the standard and the purchase agreement.
- Kinds of test
- destructive (on the specimen) · non-destructive/NDT (on the component)
- NDT methods
- VT · PT · MT · UT · RT · ET
- Personnel
- qualified e.g. to ISO 9712 (levels 1–3)
- As of
- July 2026
Destructive or non-destructive: the basic difference
Destructive tests are carried out on specimens taken from the product or from the specified test unit and consumed in the process. Their results – partly characteristic values such as yield strength and toughness, partly pass statements of technological tests – stand for the whole test unit (often a cast or a lot) and end up in the inspection certificate. Non-destructive tests (NDT), by contrast, examine the individual component without damaging it – only in this way are 100 % tests possible at all, for instance of the welds of a fitting.
Destructive tests: the classics
- Tensile test (today ISO 6892-1, formerly EN 10002-1): yield strength, tensile strength, elongation.
- Impact test (ISO 148-1, formerly EN 10045-1): toughness, typical requirements 27 J or 40 J at a specified test temperature – the basis of the TR1/TR2 logic for pipes.
- Hardness test: quick spot check; in some standards it replaces the tensile test below certain limits.
- Ring tests on tubes (flattening, ring tensile, ring expanding, drift expanding test): technological tests with limits of application per diameter and wall thickness set by the standard – they show whether the tube can be deformed without cracking.
- Chemical analysis (strictly speaking not a destructive test, but part of the same package of evidence): the manufacturer’s cast analysis, on request a product analysis on the delivered piece; supplemented on site by the PMI mix-up check with a spectral or X-ray fluorescence gun. Important: handheld XRF devices do not measure carbon – for carbon-dependent distinctions (such as L grades) OES or suitable LIBS methods are needed.
The six standard NDT methods
| Code | Method | Finds | Suitable materials | Main limitation |
|---|---|---|---|---|
| VT | visual testing | surface defects | all | surface only, mandatory first step, adequate lighting needed |
| PT | dye penetrant testing (cleaner, penetrant, developer) | surface-breaking cracks and pores | non-porous materials – including non-magnetic ones (austenitic stainless steel, aluminium, nickel alloys) | surface only, clean smooth surface needed, temperature-sensitive |
| MT | magnetic particle testing | surface defects and, to a limited extent, near-surface defects | ferromagnetic materials only (carbon steel, low-alloy, ferritic, also duplex because of its ferrite content) | not for austenitic steels, two passes offset by 90°, demagnetisation where required |
| UT | ultrasonic testing | internal defects over the full wall thickness, good for planar defects (cracks, lack of fusion) | most metals (coarse-grained castings difficult) | experienced operator needed, smooth accessible surface, conventionally no image record |
| RT | radiographic testing (X-ray/gamma, today mostly digital) | volumetric internal defects (pores, slag, incomplete penetration) | most materials | radiation protection, access from both sides, misses planar defects parallel to the beam |
| ET | eddy current testing | surface and near-surface defects | electrically conductive materials only | penetration depth strongly dependent on frequency and material, sensitive to lift-off |
Dye penetrant testing in practice: the familiar three steps from the spray can – cleaner, red penetrant, white developer. The developer draws the penetrant that has entered the cracks back out and makes them visible as red lines. It is the standard method where magnetic particle testing is not possible or not practical – for instance on austenitic stainless steels, aluminium and nickel alloys. (Duplex steel, by the way, is magnetic because of its ferrite content – there the choice of method follows the test task and the test specification.)
Selection rules of thumb: magnetisable material → magnetic particle testing (fast, captures near-surface defects to a limited extent in addition to the surface); non-magnetisable → dye penetrant testing. Volumetric defects (pores, slag) → radiography; planar defects (cracks, lack of fusion) → ultrasonics. Modern variants such as phased-array UT and TOFD can supplement radiography or – where code and specification permit – replace it, because they need no radiation protection zones and deliver recordable images.
Leak testing: classically as a hydrostatic test at the manufacturer’s works; in gas tests a foaming leak detection agent (“leak detection spray”) makes leaks visible by bubbling. For smaller tube dimensions many standards permit electromagnetic testing (eddy current) in place of the hydrostatic test.
“What percentage is tested?” – the standard decides
There is no blanket percentage: the extent of testing is set in the product standard, in the code and in the purchase agreement – and it can range from random samples to complete testing. Three examples of this logic:
- Fittings (EN 10253-2): the product standard specifies which production welds are to be tested by which method and to what extent – depending on test category, manufacturing method and the options ordered.
- Welded tubes: the usual ultrasonic weld test to the ISO 10893 series with reference defect sizes as a percentage of the wall thickness, graded in acceptance levels (U levels) – the level is specified in the order.
- ASME B31.3 (process piping): the extent depends on fluid service category, type of weld and the stipulations of the engineering design – from random radiography to complete testing, with progressive examination: if a random sample fails, further welds by the same welder are examined.
Two practical rules apply throughout: crack indications are fundamentally not acceptable under the usual acceptance criteria for pressure-bearing components and regularly lead to rejection – binding are the criteria of the standard or specification agreed in each case. And: without traceable documentation (report, operator qualification, calibration) a test cannot be used towards clients and inspection representatives. The operator qualification itself is standardised (e.g. ISO 9712, three levels, certificates limited in time).
Frequently asked questions about test methods
What is the difference between destructive and non-destructive testing?
Destructive tests (tensile test, impact test, ring tests) consume a specimen taken from the product; the results stand for the specified test unit (often a cast or a lot) and appear in the inspection certificate. Non-destructive tests (visual, dye penetrant, magnetic particle, ultrasonic, radiographic and eddy current testing) examine the component itself without damage and can therefore cover up to 100 % of the parts or welds.
Magnetic particle or dye penetrant testing – when do I use which?
The rule of thumb follows the material: ferromagnetic steels (carbon steel, low-alloy, ferritic) are preferably tested with magnetic particles – this is fast and, in addition to the surface, captures near-surface defects to a limited extent. Non-magnetisable materials such as austenitic stainless steel or nickel alloys are tested with dye penetrant (cleaner, penetrant, developer), which makes surface-breaking defects visible. Duplex steel is magnetic because of its ferrite content – here the choice of method follows the test task and the test specification.
Are welds always 100 % radiographed?
No – the extent is set in the respective product standard, in the code (for instance ASME B31.3 by fluid service category) and in the purchase agreement. The range runs from random samples to complete testing. Whoever needs a particular extent of testing specifies it in the order – and crack indications are not acceptable under the usual acceptance criteria regardless of the extent of testing.
What happens in dye penetrant testing with the three spray cans?
First the surface is degreased with the cleaner, then the usually red penetrant is sprayed on and seeps into surface-breaking cracks. After the dwell time the excess is removed and the white developer applied: it draws the penetrant out of the defects so that cracks appear as red indications on a white background. A clean, smooth surface is a prerequisite – grinding marks, oil or paint distort the result.
Testing and evidence thought of as one
From the right type of certificate to third-party inspection organised for you – the neighbouring page and our export team lead on.
Technical basis: ISO 6892-1, ISO 148-1, the ISO 10893 series, ISO 9712 and the extent-of-testing logic of the product standards and codes (among them EN 10253-2, ASME B31.3), BDS trade literature and Zickwolff’s own trading practice (as of July 2026). Specific extents of testing always follow from the edition of the standard or code agreed in each case. All information without guarantee – the current standard texts and the manufacturers’ technical documentation always prevail.