export@ezsteelpipe.com
+86 731 8870 6116
Cyclic thermal loading is one of the most punishing service conditions that industrial piping ever has to survive. Every time a boiler ramps up to full load, a hydrocracker swings between hot and cold streams, a marine steam plant is brought out of lay-up, or a heat exchanger is taken offline for cleaning, the metal in the pipe wall is forced to expand, contract, and then do it again. Plain carbon and carbon alloy steel can be specified to handle that punishment, but only if the grade, the microstructure, and the weld procedure are chosen with the loading history in mind. This article explains what actually happens inside a steel pipe when it is heated and cooled repeatedly, how different alloying strategies change that response, and how engineers can use that understanding to specify pipe that lasts.
In design codes, "high temperature" usually starts around 300 °C and "high pressure" begins at roughly 10 MPa. In a coal-fired boiler, the pressure tubes in the superheater sit closer to 600 °C and well above 25 MPa; in a refinery hydrocracker, the reaction-effluent piping swings between 250 °C and 450 °C every time the unit is cut back; in a desalination plant, the evaporator tubes cycle between condensing steam and cold feed on every start-stop. None of these are steady-state problems. They are fatigue problems, in the broad sense of the word.
Cyclic thermal loading is therefore not one number on a data sheet. It is a combination of:
For a once-through superheater tube that is expected to run for 30 years with perhaps 100 full cold starts in that period, the dominant damage mechanism is high-temperature creep-fatigue interaction. For a furnace waterwall that sees 5 000 partial-load transients a year, low-cycle thermal fatigue is the controlling case. Two different design problems, two different steel choices.
When a pipe is heated, the inner surface wants to expand more than the outer surface. The constraint produces a compressive thermal stress on the hot face and a tensile stress on the cold face. On cooling, those stresses reverse. If the wall is restrained by a header, a support, or by its own thickness, the cyclic strain range is the same every cycle, but the damage accumulates.
The mechanisms that consume the life of the steel are well known:
Low-cycle thermal fatigue covers the 10²–10⁴ cycle range, where each cycle imposes enough plastic strain to cause measurable damage per pass. Initiation usually starts at the surface, at weld toes, at branch connections, or at any geometric feature that concentrates strain. In carbon-manganese steels, the classic fish-eye and short thumbnail cracks appear in the first 100–300 cycles if the strain range is high.
High-cycle thermal fatigue sits in the 10⁴–10⁷ range and is typical of process piping that sees many small transients a day. The strain per cycle is small, but the oxide layer that forms on the steel during the hot half-cycle changes the surface condition. Repeated oxide spalling accelerates initiation and shortens life compared with the inert-environment fatigue curve.
Creep-fatigue interaction is what most cyclic service actually experiences at peak temperature. The steel is being plastically strained in each cycle and simultaneously creeping under the mean stress. The two damage modes are not additive in any simple sense; they interact, and the interaction is what kills most high-temperature components.
Plain carbon steel, in grades such as ASTM A106 Grade B or GOST 8732 with a 0.20–0.30 % carbon range, is the cheapest option and behaves well in moderate conditions. At room temperature the yield strength is around 240 MPa, with good ductility and good weldability. As the metal heats above about 350 °C, three things change at once:
In practice, this means plain carbon steel is suitable for feedwater lines, low-pressure steam piping, and process lines that stay below 400 °C and below 100 bar, with cycling rates of a few hundred full cycles per year. It is not a candidate for superheater tubing, for any supercritical or ultra-supercritical steam circuit, or for a hydrocracker reaction loop that swings through 450 °C on every turndown.
Adding alloying elements changes the steel's response to cyclic thermal loading in three distinct ways: it changes the strength envelope, it changes the oxide that forms on the surface, and it changes the microstructure's stability over time.
Chromium is the workhorse. It slows scale growth dramatically, especially above 9 %, where a thin, adherent Cr₂O₃ layer forms in steam and in oxidizing flue gas. For cyclic service, that matters because each hot half-cycle consumes less wall thickness, and the oxide is less likely to spall on cooling. It is also the element that gives the steel its sulfidation resistance in refinery service, where H₂S is present.
Molybdenum strengthens the matrix at high temperature by solid-solution strengthening and by forming fine, stable carbides. In cyclic service, the practical effect is that the steel can hold a higher mean stress at peak temperature without creeping, which means a larger share of the cyclic strain range is taken elastically and the plastic strain per cycle is smaller. Grades with 0.5 % Mo (P11) and 1 % Mo (P22) are the standard answers for refinery and petrochemical cycling service up to about 550 °C.
Vanadium and niobium add a second tier of strengthening through fine MX-type carbonitrides that are stable well above 600 °C. Grades such as P91 (9 Cr–1 Mo–V–Nb) and P92 (9 Cr–2 W–V–Nb) use this mechanism to keep creep strength at 600 °C for 100 000 h service. In cyclic service, the same precipitates also pin dislocation movement during the plastic portion of the cycle, which reduces cyclic softening and extends life.
Tungsten, used in P92 and T92, raises the temper-resistance of the steel. For cyclic loading, the benefit shows up as less drop in hardness after many thermal cycles, which keeps the strain distribution stable over the life of the pipe.
A useful way to summarize the effect is to compare two steels on the same temperature cycle, 400–600 °C, with a 90-minute period. Plain carbon steel would reach end-of-life in low-cycle fatigue at roughly 600 cycles. A P22 tube on the same cycle can exceed 3 000 cycles before initiation, and a P91 tube can exceed 6 000. The alloying does not make the steel immune to cycling, but it shifts the damage curve to a much more useful place.
Alloying only pays off if the steel ends up with the right microstructure. For cyclic service, three heat-treatment conditions matter:
Normalized and tempered (N&T) gives a fine ferrite-pearlite or ferrite-bainite structure in carbon and low-alloy steels. It is the standard condition for P11, P22, and most ASTM A192/A179 tubes used in heat exchangers. N&T tubes handle cyclic service well below 500 °C but soften progressively above 540 °C, so the cyclic endurance curve drops.
Quenched and tempered (Q&T) refines the grain and produces a tempered martensite or bainite structure with a much better combination of strength and toughness. The 9 Cr family (P91, P92, T91, T92) is always supplied in Q&T condition, and that is what gives these grades their high creep-rupture strength and their good low-cycle fatigue life at 600 °C.
Solution-annealed condition is the choice for austenitic stainless and copper-nickel grades used in heat-exchanger service, where corrosion resistance and thermal conductivity matter more than high-temperature strength. Solution annealing removes precipitates from grain boundaries, which gives a stable response to repeated thermal cycling in the 200–400 °C range, typical of U bend tube service in condensers and feed heaters.
A practical rule: if the cyclic service is above 500 °C and the design life is more than 100 000 h, the steel should be specified in the Q&T condition with full tempering above the maximum service temperature. If it is below 500 °C, N&T is usually sufficient.
Most cyclic failures initiate in the weld zone, not in the parent pipe. Three reasons explain this.
First, the heat-affected zone (HAZ) is a metallurgical gradient. The grain size, hardness, and precipitate distribution change over a few millimetres, so the strain compatibility between the weld metal, the HAZ, and the parent pipe is imperfect. Under cyclic thermal strain, that incompatibility concentrates at the fusion line and at the weld toe.
Second, residual stress from welding adds a tensile mean stress to the cyclic thermal stress. The effective stress range that the steel sees is shifted upward, and the mean stress is exactly the term that accelerates creep-fatigue damage.
Third, any geometric feature — a reinforcement, a branch connection, a mismatch in thickness — creates a local stress concentration that amplifies the thermal strain at the surface.
The mitigation is well established. For cyclic service in the creep range, post-weld heat treatment (PWHT) is mandatory. For P91 and P92, the standard practice is to hold the joint at 760 °C for a minimum of two hours, which tempers the as-welded martensite and brings the HAZ back to the Q&T condition. Without PWHT, the HAZ softens during the first few hundred hours of service, and the soft zone concentrates strain in every subsequent cycle.
Weld profile matters as well. A high, narrow cap with a sharp toe is a crack initiator; a smooth, low cap ground flush at the toe is not. For any piping that will see significant cyclic loading, the weld should be specified with a controlled cap height, a ground or blended toe, and a documented PWHT cycle.
Most of the major pipe and tube standards now include language that explicitly covers cyclic service:
Specifying to one of these standards is not, on its own, enough for cyclic service. The procurement document has to add the conditions that matter: the heat-treatment condition, the PWHT cycle, the cap-height and toe-blending limits on welds, the sampling plan for hardness surveys on the HAZ, and the NDT scope (typically 100 % radiographic plus ultrasonic on critical welds).
Case A — subcritical boiler, 540 °C main steam, 200 cold starts in 30 years. The best fit is P22 in the normalized-and-tempered condition, supplied to ASTM A335. The cyclic endurance at 540 °C with a 50 °C range is well above the design cycle count, and P22 welds easily. PWHT at 705 °C for one hour is the standard finishing step.
Case B — ultra-supercritical boiler, 600 °C reheat, 200 cold starts in 30 years. P91 or P92, in the Q&T condition, is required. P91 is more widely available and easier to weld; P92 allows thinner walls and lower metal temperatures but needs tighter control of the welding procedure and the PWHT cycle. Either way, the procurement specification has to mandate weld-cap grinding, full-bore radiographic examination, and a hardness survey on every HAZ.
Case C — refinery hydrocracker, 250–450 °C, 1 000 turndown cycles per year. A 1.25 Cr–0.5 Mo (P11) tube in N&T condition is the standard answer. The temperature range is moderate, the cycles are numerous but shallow, and the resistance to sulfidation is the deciding factor, not peak strength. P11 also has the ductility to be bent into U-bend configurations for the feed-effluent exchangers without cracking.
Material choice only sets the potential life of a pipe under cyclic service. Operating practice determines how much of that potential is actually used. A few rules consistently pay off:
Two trends are putting more cyclic demand on carbon and carbon-alloy piping than ever before. The first is the move to flexible operation in coal-fired power plants, which is driven by the growth of wind and solar on the grid. Plants that used to run at base load now cycle daily, and the cyclic damage on the high-temperature header piping is starting to dominate the inspection programme.
The second is the shift to electric and hydrogen-fired process heat, which replaces the steady radiant heat of a fired heater with rapid electric heating or with rapid burners cycling on renewable power. Both approaches demand pipe that can take thousands of fast cycles without losing creep strength.
The industry response is to extend the 9 Cr family to higher temperatures, to develop creep-resistant austenitic alternatives for the most aggressive cycles, and to build damage-tolerant design rules into the next revision of the major piping codes. For engineers specifying pipe today, the practical message is unchanged: pick the right grade for the peak temperature, weld it correctly, heat-treat the joint, and control the start-up and shutdown rates. Carbon and carbon-alloy steel will do the job if it is given the chance.
For more on how to apply these grades to power plant, petrochemical, and marine service, the product range at EZ Steel Industrial covers the relevant ASTM, EN, GOST, JIS, and GB/T specifications with full mill test certificates.
Related Products