Oct 18, 2024 Leave a message

Why Steel Pipes Crack When Heated: Causes and Prevention

What the Question Really Asks: Cracking During Heating or Cracking in Hot Service

When a pipe cracks in a furnace, a boiler or a heat exchanger, the report usually says the pipe cracked when heated. In practice there are two distinct situations behind that sentence, and they have different root causes. The first is failure during a heating operation: stress relief, normalising, bending with heat, welding preheat, or a shutdown and restart thermal cycle. The second is failure in sustained high temperature service, where the wall has been above its design range for months or years. A crack that opens in fifteen minutes is almost always a stress or ductility problem; a crack that opens after eight thousand hours is almost always a microstructure or creep problem. Getting that distinction right at the start saves the cost of replacing the wrong component twice.

The second useful distinction is where the crack sits. A longitudinal crack in a seamless tube points to a manufacturing defect or to hoop stress; a circumferential crack points to axial restraint or to a weld; a network of fine surface cracks points to a thermal fatigue or corrosion mechanism; and a crack that follows the weld toe points to a fabrication or heat treatment issue.

Cause One: Material and Manufacturing Defects

Defects already present in the tube are amplified by heating, because thermal expansion converts a small geometric discontinuity into a stress concentration. Four defect populations matter:

Non-metallic inclusions and segregation - stringers and brittle inclusion clusters act as crack initiators. In plate and heavy section the relevant acceptance route is ultrasonic examination and, for critical service, inclusion rating by microscopic method to ASTM E45.

Porosity and internal fissures - typical of poor casting practice or of a pierced shell with an off-centre bore. They are usually detectable by ultrasonic wall examination to ASTM E213 and by the hydrostatic test only when they are through-wall.

Laminations and laps - rolling defects that lie roughly parallel to the surface. They open when the wall is heated and the surface oxide spalls, or when the tube is expanded into a tube sheet.

Weld defects - lack of fusion, incomplete penetration and crater cracks in ERW or welded pipe. These are orientation-specific and are the reason the longitudinal weld in ERW pipe is examined by ultrasonic or electromagnetic methods per the product specification and its supplementary requirements.

The procurement consequence is that heat service orders should carry the non-destructive test requirements of the product specification as a stated line item, with a stated acceptance level. A tube bought to a dimensional and tensile specification only has no verified internal quality.

Cause Two: Residual Stress from Cold Work and Fabrication

Cold drawing, cold pilger rolling, cold bending and even heavy machining leave a residual stress field in the wall. The stress alone rarely cracks a sound tube, but it adds directly to the thermal stress when the tube is heated, and it drives stress corrosion cracking when the environment supplies the second ingredient. That is why cold formed bends in stainless lines are solution annealed or at least stress relieved before service, and why the stress relief requirement appears explicitly in codes for services where cracking is credible.

The fabrication operations that add the most residual stress are: cold field bending without stress relief, welding with excessive restraint and no stress relief, forced alignment of flanges, and tube-to-tube sheet expansion. In each case the crack appears after the first significant thermal cycle, not during fabrication, which is why the failure is often attributed to heating rather than to the fabrication method.

Cause Three: Overheating, Burning and Scale

Temperature is the variable that most directly changes a steel pipe into a cracked pipe. Around 300 C carbon steel reaches the tempering range where a cold worked or quenched structure loses strength. Above roughly 400 C creep becomes a design consideration and the allowable stress in ASME B31.3 or in the boiler code has to be taken at temperature rather than at ambient. Above roughly 540 C oxidation becomes significant for carbon steel, and above about 700 C the material is in a range where short term exposure alone can cause permanent damage.

Burning is a separate and more severe event. If a tube is heated locally with an oxy-fuel torch beyond the point where the surface oxide becomes liquid, the grain boundaries oxidise preferentially, the steel loses ductility locally, and the burnt zone will crack on cooling with no external load. The signature is a thick, dark, irregular scale with a decarburised layer beneath it, visible on a metallographic section. A burnt tube cannot be repaired by re-heat treatment, because the grain boundary oxidation is permanent.

The correct control is measured temperature, not colour of the scale. Furnace thermocouples with a recorded chart, or contact thermocouples on local heating, plus a defined maximum interpass temperature, are what make a heat operation auditable after the fact.

Cause Four: Thermal Expansion Under Restraint

A pipe heated uniformly and free to expand develops no thermal stress at all. The stress appears only when expansion is restrained by anchors, guides, rigid nozzles, cold spring that has been incorrectly installed, or by a temperature gradient across the wall or along the line. The classic mechanism is a stainless line heated rapidly while the outside surface reaches temperature before the inside: the outer fibres want to expand more than the inner ones, and with the ends fixed the result is compressive yield on the outside and tensile stress at the inner surface. On cooling the stress reverses, and enough cycles produce low cycle thermal fatigue cracks that start at the inner bore.

The engineering response is a flexibility analysis to the applicable code, with the actual support arrangement and the actual operating and design temperatures, followed by correct installation of the specified expansion loops, guides and spring hangers. A crack that appears during start-up is frequently a support or guide problem, not a material problem.

Cause Five: Metallurgical Ageing Mechanisms

Several mechanisms change the material itself during exposure at temperature, and each has a characteristic temperature band and a characteristic inspection method.

Mechanism Temperature band Where it applies Failure signature Verification
Sensitisation and intergranular attack about 425-815 C Austentitic stainless 304, 316, 321 weld heat affected zones Cracks along grain boundaries in the weld zone after a thermal cycle ASTM A262 practice E on a sample; metallography
Sigma phase formation about 600-900 C Duplex 2205 and 2507, super-austenitic grades Loss of toughness and corrosion resistance, brittle cracking on cooling ASTM A923 impact and metallography
Blue brittleness and strain ageing about 250-350 C Cold worked or rimmed carbon steel Sudden loss of ductility, brittle fracture at low strain Tensile and impact testing of the affected heat
Graphitisation roughly 450-650 C Carbon and carbon-molybdenum steels in long term service Chains of graphite nodules along the heat affected zone, then brittle fracture Metallographic replication on site
Creep and stress rupture above the code creep range, commonly above 400 C for carbon steel Boiler and reformer tubing under sustained stress Bulging, diametral growth, longitudinal fissures near the hot side Diametral measurement and replication; remaining life assessment
Hydrogen attack and embrittlement process dependent, elevated temperature and partial pressure Carbon and low alloy steel in hydrogen or wet H2S service Internal fissuring, decarburisation, blistering, loss of ductility Ultrasonic examination to detect blistering; hardness and metallography
Thermal fatigue cyclic, whatever the absolute temperature Any grade subject to repeated heating and cooling under restraint Multiple parallel cracks starting at the bore or at a stress concentration Penetrant testing per ASTM E165 and metallography

Cause Six: Welding and Heat Treatment Errors

A welded joint heated without the specified preheat, or cooled too fast after welding, can transform into a hard martensitic structure in the heat affected zone. That zone has high hardness, low toughness and high residual stress, and it cracks either immediately as hydrogen assisted cold cracking or on the first thermal cycle in service. The controls are the ones written into the qualified welding procedure: preheat and interpass temperature, heat input range, hydrogen controlled consumables where required, and a post weld heat treatment at the temperature and hold time the code specifies for the material and thickness.

The opposite error also occurs: excessive post weld heat treatment temperature or hold time in austenitic stainless, which causes sensitisation or sigma formation and leaves a joint that is dimensionally perfect and metallurgically damaged. This is why the post weld heat treatment chart belongs in the quality record with actual measured temperature, not only a nominal set point.

How to Find the Real Cause After a Failure

A failure investigation is only as good as the samples taken. A practical sequence for a cracked pipe is:

Document before disturbing - photograph the crack orientation, location relative to the weld and to the hot side, and measure the crack length and the wall thickness at the crack.

Dimensional survey - outside diameter and wall at several points, to distinguish creep bulging from a mechanical overload.

Surface examination - penetrant testing to ASTM E165 or magnetic particle testing to ASTM E709 for ferritic material, to establish whether the visible crack is the whole population or one of many.

Volumetric examination - ultrasonic examination to ASTM E213 or radiographic examination to ASTM E94 to establish the depth and to find internal damage.

Hardness survey - a rapid discriminator between an incorrect heat treatment, cold work and local overheating.

Metallographic section - the decisive test. Grain size per ASTM E112, decarburisation depth, carbide morphology, sigma content, and the presence of grain boundary oxidation together identify the mechanism.

Chemical verification - positive material identification to ASTM E1476 to confirm that the installed tube is the grade that was specified.

A crack report that stops at visual inspection cannot distinguish between overheating, sensitisation and thermal fatigue, and will usually lead to replacing the tube without removing the cause.

Prevention: What Belongs in the Purchase Order and in the Procedure

Most of these failures are preventable at the specification stage rather than at the failure stage. For a pipe intended for heat service, the purchase order should state the product specification and its grade, the design temperature, the required heat treatment condition and whether it is to be supplied in the as-rolled, normalised, quenched and tempered or solution annealed condition, the non-destructive test method and acceptance level, and the test certificate content including actual chemical analysis and mechanical results at the specified condition.

On the operating side, three disciplines prevent the majority of heating related failures: keep the recorded temperature history so that an overheating event can be identified rather than argued about, verify that the piping flexibility and support arrangement matches the design analysis, and inspect the known damage locations at defined intervals instead of only when a leak appears. Replication and hardness checks on high temperature lines cost a fraction of a forced outage.

Finally, correct grade matching matters more than any single inspection. A carbon steel tube in a 600 C stream is outside its oxidation range; an austenitic grade in a chloride bearing stream at 200 C is a stress corrosion cracking candidate regardless of how well it was heat treated; and a duplex grade held in the sigma band loses the toughness that justified its purchase. The right material, correctly heat treated, correctly fabricated and correctly supported, does not crack simply because it was heated.

FAQ

Q: Can a pipe crack from heating even without any applied load?
Yes, in three situations: residual stress from cold work or welding is released by heating and the release is concentrated at a defect; local overheating causes grain boundary oxidation and loss of ductility so the material cracks on cooling; and a temperature gradient across the wall under restraint generates thermal stress without any external force. All three are load paths that exist before the pipe is pressurised.

Q: How can I tell overheating from a material defect?
Examine the fracture surface and a metallographic section. Overheating leaves thick irregular scale, a decarburised layer and grain boundary oxidation, with no inclusion origin at the crack. A material defect leaves a flat facet or an inclusion cluster at the initiation point with no decarburisation around it. A hardness survey and a chemical check by positive material identification complete the discrimination.

Q: Why do austenitic stainless pipes crack near welds after a thermal cycle?
The most common cause is sensitisation: chromium carbides precipitate on the grain boundaries in roughly 425 to 815 C and deplete the adjacent matrix, so the boundaries corrode or crack. The fix at the design stage is a low carbon grade such as 316L or a stabilised grade such as 321 or 347, and at the fabrication stage a controlled heat input with rapid cooling and no unnecessary post weld heat treatment. ASTM A262 Practice E verifies the delivered condition.

Q: Does stress relieving prevent cracking during heating?
Stress relief reduces the residual stress that adds to thermal stress, so it lowers the risk substantially but does not change the material metallurgy. It will not help a tube that has already been sensitised, sigma embrittled or burnt, and for austenitic stainless a badly chosen stress relief hold can make the microstructure worse. Select the heat treatment from the material, the thickness and the applicable code rather than as a general precaution.

Q: What is the difference between thermal fatigue and creep damage?
Thermal fatigue is cyclic and produces multiple parallel cracks, usually starting at the bore where the temperature swings first, and the crack faces are often oxidised without significant deformation. Creep is time dependent and produces diametral growth, bulging and longitudinal fissures on the hot side without the cyclic crack pattern. Measurement of diameter and wall, plus replication, separates the two.

Q: What should be recorded to make a heat treatment auditable?
Actual temperature against time from calibrated thermocouples attached to the component, the heating and cooling rates, the hold temperature and hold time, the furnace identification and the chart or data file. A certificate that states only that the pipe was stress relieved cannot be used to rule out either under-treatment or overheating when a failure occurs.

Send Inquiry

whatsapp

Phone

E-mail

Inquiry