Testing standard
ISO 13954, Plastics pipes and fittings — Peel decohesion test for polyethylene (PE) electrofusion assemblies of nominal outside diameter greater than or equal to 90 mm; ISO 13955, Plastics pipes and fittings — Crushing decohesion test for polyethylene (PE) electrofusion assemblies
Written and technically reviewed by Dak System Inc. engineeringLast reviewed
ISO 13954 and ISO 13955 are the two decohesion tests for polyethylene electrofusion assemblies. ISO 13954 peels a strip from a socket assembly of 90 mm nominal outside diameter or more on a tensile machine; ISO 13955 crushes a sector of an assembly between platens and the fitting is then levered off. Neither reports a strength: both report the proportion of the fused plane that failed brittlely, judged against the value in the product standard. ISO 13954:2025 is the current second edition; ISO 13955 remains the 1997 first edition, confirmed on review in 2024.
Both destroy a fused joint and read the fracture surface. In the peel test a strip cut from a pipe-and-socket assembly is pulled apart on a tensile machine through a shackle-type link, peeling the fused interface progressively open along the socket. In the crushing test a sector of the assembly is compressed between platens next to the socket until the pipe walls meet, after which the fitting is levered off carefully and the exposed interface examined. Both run at 23 °C plus or minus 2 °C.
A fracture character rather than a strength. Loads are recorded — the peel test notes the maximum breaking load, and the crushing test notes the compressive force on a saddle piece just before the pipe walls touch — but neither figure decides anything. What decides it is read off the fracture surface. Along the fused plane each length is judged ductile, where the polyethylene has drawn and torn, or brittle, where the surfaces have simply come apart. On a socket the brittle-failure length is expressed as a percentage of the distance between the first and last coil winding; on a saddle the brittle-failure area is expressed as a percentage of the area of the fusion plane. That percentage is the result. A properly fused joint is one piece of polyethylene, so it should fail ductilely in the pipe or fitting wall, not cleanly at the coil plane. Brittle decohesion says the surfaces never fused: contaminated pipe, unscraped oxide skin, out-of-round pipe leaving a gap, or a fusion cycle that did not deliver its energy. The pass line is not in either method — both fail the assembly if the percentage exceeds the value in the relevant product standard, and ISO 13954:2025 adds an informative annex recommending a criterion.
Both methods test a joint somebody made, not a moulded coupon, so the preparation of the assembly is part of the test and is specified by a third standard.
ISO 13954 was reissued as a second edition in November 2025 and its procedure was reorganised. No peel rate, dimension or tolerance is quoted here for that method, because the current clause could not be read and the withdrawn 1997 value must not be carried forward as though it still applied.
Cc = (d2 / y) × 100
The reference length is the coil, not the fitting. Measuring against the socket length instead gives a percentage that is too low and looks respectable.
Cc = (SF / ST) × 100
On a saddle the quantity is an area ratio rather than a length ratio, so the two forms of the test are not interchangeable arithmetic.
Compare Cc with the value in the relevant product standard
Neither method sets the acceptance figure. ISO 13954:2025 adds an informative Annex A recommending a criterion, but it is informative and the product standard still governs.
Levering is where a sound joint is most often failed. If the fitting will not come away under small movements, the answer is patience, not a bigger lever.
Flat, rigidly fixed platens for the crushing test. What the method adds to them is a pair of stops limiting closure to twice the pipe wall thickness, which is what ends the compression before the interface is damaged.
SpecificationsA cell matched to the job rather than to the frame. On a saddle crushing test the compressive force at the moment the pipe walls meet is a reported value, so it has to be resolved properly and not read off the bottom of a large range.
SpecificationsTwo ordinary capabilities and one piece of discipline. The peel test needs a tensile frame able to pull at the specified slow rate with force enough to separate the components, plus the shackle-type link that lets the socket rotate as the strip peels. The crushing test needs a compression frame at a constant 100 mm/min plus or minus 10 per cent, with stops limiting platen closure to twice the pipe wall thickness — the stops define the end of the test. The discipline is the levering: the fitting is worked off in small movements with no impact, because a sharp blow tears ductile material and makes a good joint look bad.
Levering too hard, the commonest way to fail a joint that was sound. Testing too soon after fusion, before the joint has cooled and crystallised. Omitting the stops, so the platens crush past twice the wall thickness and damage the interface before it is read. Measuring the brittle length against the wrong reference. And reading the fracture surface without the experience to separate brittle decohesion from smooth ductile shear — the judgement is visual, and benefits most from a reference set of known joints.
| ISO 13954 peel | ISO 13955 crushing | ISO 13956 tear | ISO 21751 strip-bend | |
|---|---|---|---|---|
| Joint | Electrofusion socket | Electrofusion socket or saddle | Saddle fusion joint | Electrofusion fittings and saddles |
| Machine | Tensile frame with a shackle-type link | Compression frame with limiting stops | Tensile frame | Bend arrangement |
| Size | 90 mm nominal outside diameter and above | 16 mm to 225 mm | As that method sets | As that method sets |
| Reports | Percentage brittle-failure decohesion | Percentage brittle-failure decohesion | Ductility of the fusion interface | Ductility of the fusion interface |
All four ask the same question — did the two pieces of polyethylene actually become one — and they ask it of different joints with different apparatus. A product standard names the one it wants, and a result from another is not a substitute for it.
They are the two decohesion tests for polyethylene electrofusion assemblies. ISO 13954 is the peel decohesion test, for socket assemblies of 90 mm nominal outside diameter and above, in which a strip is pulled apart on a tensile machine through a shackle-type link. ISO 13955 is the crushing decohesion test, in which a sector of the assembly is compressed between platens and the fitting is then levered off. Both are maintained by ISO/TC 138/SC 5.
ISO 13954:2025 is the second edition, published on 19 November 2025, and it withdrew both ISO 13954:1997 and that edition's Amendment 1 of 2020 on the same day. ISO 13955 is still the 1997 first edition, confirmed on systematic review in 2024, with Amendment 1:2020 alongside it. The two are commonly cited together and it is easy to assume they moved together; they did not.
The percentage of the fused plane that failed brittlely. On a socket that is the total brittle-failure length divided by the distance from the first to the last coil winding; on a saddle it is the brittle-failure area divided by the area of the fusion plane. Loads are recorded — the peel test notes the maximum breaking load, the saddle crushing test notes the compressive force just before the pipe walls meet — but the percentage is what the assembly is judged on.
That the two surfaces never fused. A properly made electrofusion joint is one continuous piece of polyethylene, so it should fail ductilely, in the wall of the pipe or the fitting, with the material drawing and tearing. A clean separation at the coil plane says something stopped the fusion: contamination on the pipe, an oxide skin that was not scraped off, out-of-round pipe leaving a gap, or a fusion cycle that did not deliver its energy.
From the product standard, not from either test method. Both say the same thing: if the percentage brittle-failure decohesion is higher than the value stated in the relevant product standard, the assembly has failed. ISO 13954:2025 adds an informative Annex A recommending a criterion, but an informative annex recommends and does not require, and the product standard for the pipe system still governs.
Because it is the step that most often fails a joint that was sound. The method says to use only small movements of the lever and to subject the test piece to no impacts, and that instruction is doing real work: a sharp blow tears through ductile material and leaves a surface that reads as brittle decohesion afterwards. There is no way to tell the two apart on the finished fracture face, so the discipline has to be in the doing.
A compression frame able to hold a constant platen approach of 100 mm/min plus or minus 10 %, fitted with stops that limit closure to twice the pipe wall thickness. Capacity is rarely the constraint — a pipe sector collapses well below what a general-purpose frame provides — but the stops are not optional, and on saddle pieces the compressive force just before the walls touch is a reported figure, so the load cell has to be chosen to resolve it.
Because it has not finished becoming a joint. The procedure is carried out at least twelve hours after fusion, with at least six hours at 23 °C plus or minus 2 °C between fusion and cutting and at least six hours more at the test temperature after the pieces are cut. A joint tested warm has not cooled and crystallised through the fusion zone, and it will not behave the way the same joint behaves in the ground.
Dak verifies against whichever standard the method names, and where a class applies our frames sit a class tighter than it asks.
| The method asks for | Dak supplies | |
|---|---|---|
| Capacity | Not prescribed. The peel test needs force enough to separate the components of the strip; the crushing test collapses a pipe sector, well within a general-purpose frame. On saddle crushing pieces the compressive force just before the pipe walls meet is itself a reported value. | Load cells from 1 kg to 60 ton on the Series 7200, and 0.5 to 100 kN on the Series 9000 |
| Force accuracy | unknown — no accuracy class could be confirmed in either method | ISO 7500-1 Class 0.5, verified to ASTM E4, DIN 51221 and BS 1610 |
| Gripping | Peel: shackle-type link with a tensioning chain or wire rope of at least 300 mm. Crushing: flat platens with stops limiting closure to twice the pipe wall thickness, plus a lever for separating the fitting. | Our compression anvils or a fixture built for this method, built to the specimen |
| Environment | 23 °C plus or minus 2 °C throughout; at least 6 h at that temperature between fusion and cutting, at least 6 h more after cutting, and at least 12 h from fusion to test | 3009 series chambers, −150 °C to +400 °C — temperature only |
This page describes the method as practised. The governing text is the current edition from the issuing body. Tell us what you are testing and we will answer with the machine, the fixture and a quotation.
