A failed hose is a piece of evidence. Where it cracked, what the braid looks like, and whether it wept or burst each point at a different cause — and most of those causes are not the hose.
Stainless steel hose rarely fails at random. It fails in a small number of recognisable ways, and each one leaves a distinct signature on the part. Read that signature and you learn what to change; miss it and you fit an identical hose into the same conditions and wait for the same outcome.
What follows is the seven modes we see most often, what each looks like on a hose in hand, and what actually caused it.
1. Flex fatigue
The most common failure by a wide margin, and almost always a specification error rather than a manufacturing one.
What it looks like: fine cracks at the crest of a convolution, usually clustered within the first few convolutions next to a fitting. The hose weeps rather than bursts — a slow drip that appears days before anything dramatic happens.
What caused it: the hose was bent tighter than its dynamic rating, or it was flexing at a point where it cannot flex. The convolutions immediately behind a ferrule are restrained by the fitting, so movement that should be spread over a long length gets concentrated into two or three convolutions. Leave a straight run of at least one and a half hose diameters before any bend begins.
The other frequent cause is using the static bend radius on a line that moves. Those are two different numbers answering two different questions, and the gap between them is often a factor of two or more.
2. Torsion
What it looks like: the braid's diamond pattern is visibly distorted — stretched on one side, bunched on the other. The hose holds a twist when it is taken off the line. Failure, when it comes, is usually the braid parting rather than the core cracking.
What caused it: almost always installation. The fitting was rotated while being tightened and the twist went into the hose instead of the spanner. Corrugated hose tolerates very little torsion — it is the one direction its geometry has no answer for, and the braid transmits the load straight into the core.
Two spanners, always: one holding the hose fitting still, one turning the nut.
3. Chloride stress corrosion cracking
The mode that surprises people, because the hose looks perfectly clean right up until it leaks.
What it looks like: a pinhole leak with no rust, no thinning and no visible corrosion anywhere on the part. Under magnification the cracks branch like a river delta.
What caused it: three things present at once — an austenitic stainless steel, a tensile stress, and chlorides. Susceptibility is generally taken to begin around 60 °C and rises sharply with temperature. The stress does not have to be applied load; residual stress from forming the convolutions is enough.
The chloride source is usually mundane: coastal air, chlorinated wash-down water, or — the classic — chlorides concentrating under wet thermal insulation, where a lagged hose sits in a warm damp sleeve that dries and re-wets for months.
The important correction: moving from 304 to 316, or to 321, does not solve this. All three are austenitic and all three are susceptible. 316 buys some margin, and 321 buys none at all for this particular mode — its advantage is intergranular resistance after welding and high-temperature service, which is a different problem. Keeping chlorides off the outside of the hose is the control that works.
4. Pitting and crevice corrosion
What it looks like: rust staining on the braid, and braid wires parting one at a time. The braid almost always goes before the core — the wires are thin, they sit outboard of everything else, and each crossover point is a crevice.
What caused it: stagnant liquid trapped in the convolutions of a hose that cannot drain, weld spatter landing on the braid during nearby fabrication, or iron contamination picked up from carbon steel tooling or a carbon steel wire brush. That last one is worth watching: a few embedded iron particles will rust and pit an otherwise sound stainless surface.
If the line must drain dry between batches, that is a geometry decision made at specification time, not something to fix later with drain points.
5. Erosion
What it looks like: a leak with no cracking at all. Sectioned, the convolution crests are visibly thinned from the inside, usually worst just downstream of a bend.
What caused it: high flow velocity, entrained solids, or both. Corrugations are turbulence generators by design, and every convolution crest is an obstacle presented to the flow. For abrasive slurries or high-velocity gas, an interlocked liner inside the corrugated core takes the wear and leaves the pressure-containing layer alone.
6. Overpressure — which is usually a temperature problem
What it looks like: a burst rather than a weep. The braid opens over a length and the core balloons through it.
What caused it: nine times out of ten, not a pressure spike. The working pressure stamped on the tag is the figure at 20 °C, and it falls a long way as the line gets hot.
From our own derating table, the fraction of the 20 °C rating that survives at 200 °C is:
- SS304 — 0.60
- SS316 — 0.66
- SS321 — 0.74
So a 304 assembly rated 45 bar cold is a 27 bar assembly at 200 °C. Run it at 40 bar hot and it is not marginally over — it is nearly fifty per cent past its rating, and it will fail without anyone having exceeded the number written on the tag.
Grade limits matter here too: 304 and 316 are held to 425 °C for continuous service, while 321 runs to 700 °C. Above a grade's limit, no derating factor makes it the right choice.
7. Termination failure
What it looks like: a leak at the ferrule or the weld itself, with the hose body untouched.
What caused it: either the joint was made wrong — wrong procedure, contaminated surface, an arc strike beside the weld that became a corrosion initiation site — or the hose was flexing hard right at the fitting, which is failure mode one arriving at a different address. If the crack is in the parent metal beside a sound weld, the cause is movement, not welding.
Reading the part
Three questions usually narrow it to one or two candidates:
- Did it weep or burst? Weeping points at fatigue, cracking or erosion. Bursting points at pressure — and therefore at temperature.
- Where on the hose? Within a few convolutions of a fitting means flex or termination. Mid-length means erosion, corrosion or environment.
- What state is the braid in? Distorted means torsion. Rust-stained and parting means corrosion. Clean and intact over a leaking core means the failure started inside or was stress corrosion.
Keep the failed hose
The single most useful thing you can do after a failure is not throw the part away. Note which way up it was installed and how it was routed, photograph the crack location against the fittings, and keep it. A hose that failed at 500 hours and one that failed at 50,000 look similar in a photograph of the leak alone; the position of the damage relative to the fittings is what separates them.
Send us that, along with the media, the operating temperature and how the line moves, and we will tell you which of these seven it was rather than simply quoting a replacement.
For the numbers behind two of the modes above: the bend radius finder lists every hose that will make a given turn in static or dynamic service, and the temperature correction factors give the derating for each grade across the full range.