› Thermo-Mechanical Fatigue (TMF) Twin Test Simulator Back
Material Fatigue / Turbine

Thermo-Mechanical Fatigue (TMF) Twin Test Simulator

Predict the thermo-mechanical fatigue life of gas turbine blades and other high-temperature components where temperature and mechanical stress cycle together. Switch among Inconel 718, single-crystal CMSX-4 and other Ni-base superalloys, vary the maximum temperature, strain amplitude and IP/OP phasing, and read the Miner-summed life from LCF, creep and oxidation in real time.

Parameters
Material
Auto-sets Young's modulus, thermal expansion and creep activation energy
Max temperature T_max
°C
Min temperature T_min
°C
Mechanical strain amplitude Δε_m/2
%
Phasing
Relative phase of temperature and mechanical strain cycles
High-temp hold t_hold
s
Time held at peak temperature per cycle. Longer hold makes creep damage dominant.
Cycle rate
cyc/h
Results
—
Thermal strain amp. (%)
—
Total strain amp. (%)
—
LCF life N_LCF (cyc)
—
Creep contribution (cyc)
—
Total life N_total (cyc)
—
LMP (Larson-Miller)
—
Turbine blade section — temperature / strain cycle

Left: blade cross-section with cooling holes. Right: temperature (orange) and mechanical strain (blue) waveforms with the IP/OP/CD phase marker. The bottom counter tracks the cumulative cycle / life ratio.

IP / OP / CD life vs. temperature
Larson-Miller parameter — material comparison
Theory & Key Formulas

$$\frac{1}{N_{\text{total}}} = \frac{1}{N_{\text{LCF}}} + \frac{1}{N_{\text{creep}}} + \frac{1}{N_{\text{ox}}}, \qquad LMP = \frac{T\,(20 + \log_{10} t)}{1000}$$

Miner's rule combines the LCF, creep and oxidation single-mechanism lives N. LMP is the equivalent temperature-time index used to extrapolate accelerated tests. IP/OP/CD phasing changes the damage factor substantially — the signature of TMF.

$$\Delta\varepsilon_{\text{th}} = \alpha\,(T_{\max}-T_{\min}), \qquad \Delta\varepsilon_{\text{tot}} = \Delta\varepsilon_m + \Delta\varepsilon_{\text{th}}$$

Thermal strain equals the temperature span ΔT times the linear expansion coefficient α. The total strain Δε_tot, mechanical plus thermal, drives the Coffin-Manson LCF life.

Thermo-Mechanical Fatigue (TMF) — Twin Testing and Turbine Blade Life

🙋
How is "thermo-mechanical fatigue" actually different from ordinary metal fatigue? In both cases the material is just being cycled to failure, isn't it?
🎓
The big thing in TMF is that the temperature is cycling at the same time as the mechanical load. Ordinary LCF tests pull and push the sample at room temperature. TMF cycles between, say, 100°C and 800°C every 60 seconds while simultaneously oscillating mechanical strain at ±0.5%. A gas turbine blade does exactly this: every flight is one cycle — engine start to 800°C, steady cruise, descent to cool down — and an airliner racks up thousands of those cycles. That is why the TMF test, which reproduces both the temperature and the stress cycles together, is the gold standard for life prediction.
🙋
I switched the phasing from IP to OP and the life dropped by more than half. What is so different between In-Phase and Out-of-Phase?
🎓
Good observation. IP (in phase) keeps the peak tensile strain at the peak temperature, which is what the hot outside surface of a blade sees in steady cruise. OP (out of phase) flips it: hot when compressed, cold when in tension. Once the material yields in compression at high temperature, cooling it locks in a large tensile residual stress, and the oxide scale that grew during the hot half cracks off, exposing fresh metal that re-oxidises. That cycle drives intergranular cracks much faster. The ASTM E2368 community has repeatedly shown OP to be 2-3x more punishing than IP, which is why our damage factor jumps from 2.5 (IP) to 5.0 (OP).
🙋
The result cards show separate LCF, creep and total lives. Are you really adding the three together?
🎓
Yes — via Miner's linear damage rule: 1/N_total = 1/N_LCF + 1/N_creep + 1/N_ox. Intuitively, the shortest of the three dominates. At default conditions LCF is about 45000 cycles, creep about 30000, oxidation about 2 million, so creep is the bottleneck. That is mostly the 60-second hold. Set the hold to zero and you get pure TMF (LCF-dominated). Set it to an hour and creep takes over completely. Real cruise conditions sit hours at temperature, so industrial gas turbines almost always live in the creep-dominated regime.
🙋
Switching from Inconel 718 to CMSX-4 gave an order-of-magnitude longer life. Is the single-crystal effect really that strong?
🎓
Single-crystal (SX) alloys are genuinely revolutionary. In a polycrystal the grain boundaries are the weak link — creep cavities preferentially grow there and form cracks. An SX casting deliberately has no grain boundaries: a single seed is grown into the whole blade by directional solidification through a helix selector. That removes intergranular creep as a mode entirely. CMSX-4 also has γ' (ordered Ni₃Al) at about 70% volume fraction, which physically pins dislocations. Every modern large-fan engine — GE9X, Rolls-Royce Trent XWB, PW1100G — uses CMSX-4 or its cousins (CMSX-10, René N5) in the first HPT blade row, with a 200 μm YSZ thermal barrier coating on top and cooling-hole film air underneath, knocking surface temperature down by 100-200°C.
🙋
What does the Larson-Miller Parameter actually tell me? Just seeing a number doesn't help much.
🎓
LMP = T(20 + log t)/1000 is "one number that summarises the punishing-ness of a temperature-time combination". 800°C × 1000 h and 900°C × 10 h can give very similar LMP, meaning the material will fail in much the same way under both. That is what makes it possible to predict 30-year service life from a few hundred hours of lab testing. For Ni-base superalloys you find LMP=22 gives roughly 100 MPa rupture strength, 25 gives 60 MPa and 27 gives 30 MPa — that's the Master Curve. The default 24.4 puts Inconel 718 in its typical operating window. CMSX-4 still survives above 26, and that gap shows up clearly in LMP space.

FAQ

In-Phase (IP) means the maximum tensile strain occurs at the maximum temperature, which represents the steady-state condition of a hot, tensioned turbine blade surface. Out-of-Phase (OP) puts maximum compression at the maximum temperature, typical for cooled-hole edges and start-stop transients at the blade leading edge. OP is generally 2-3 times more severe than IP, because the material compressed plastically while hot develops large tensile residual stress upon cooling, which accelerates oxide-scale spallation and intergranular cracking. This tool uses damage factors IP=2.5, OP=5.0 and CD=3.5 to reproduce the trends observed in ASTM E2368 standard TMF tests.
LMP = T(20 + log t)/1000 collapses temperature T and time-to-failure t into a single "equivalent" index. For example, 800°C / 1000 h and 900°C / 10 h give similar LMP values, allowing engineers to extrapolate accelerated tests to in-service life. The default conditions in this tool yield LMP ≈ 24.4, which matches the typical creep-rupture LMP band (24-26) for Inconel 718. Smaller LMP means harsher conditions; LMP > 26 territory is only realistic for single-crystal superalloys such as CMSX-4. Turbine designers plot LMP against stress to build the Master Curve used for life estimation.
Because several damage mechanisms run simultaneously under TMF, the linear damage rule of Miner is applied: 1/N_total = 1/N_LCF + 1/N_creep + 1/N_ox. LCF captures plastic cyclic damage from mechanical + thermal strain, creep accounts for stress relaxation and cavity growth during high-temperature holds, and oxidation describes intergranular oxide-scale formation and spallation. Default conditions here yield N_LCF ≈ 45000, N_creep ≈ 30000, N_ox ≈ 2x10⁶, combining to N_total ≈ 17800 cycles. Miner's rule ignores mechanism interactions in the strict sense, but is accurate enough for design use and is the starting point for advanced methods such as Strain Range Partitioning (SRP).
CMSX-4 is a directionally solidified single-crystal (SX) Ni-base superalloy with no grain boundaries and γ' (Ni₃Al) volume fraction near 70%. The fastest failure mode in polycrystalline alloys — intergranular creep and oxidation — does not exist in SX, extending IP/OP TMF life by 3-10x. CMSX-4 is used in the first-stage high-pressure turbine blades of GE9X, Rolls-Royce Trent XWB and PW1100G, paired with thermal barrier coatings (YSZ TBC) and cooling-hole film cooling to drop surface temperature by 100-200°C. Selecting CMSX-4 in this tool with ε=0.3% and a 30s hold gives TMF lives above 10⁵ cycles, making the gap to Inconel 718 visually obvious.

Real-world applications

Aero high-pressure turbine blades: The first-stage HPT blades of GE9X (Boeing 777X), Rolls-Royce Trent XWB (A350) and PW1100G (A320neo) operate in 1600°C-class combustor exit gas while spinning at several thousand rpm. One flight equals one TMF cycle, and 20-year service life demands 50,000 - 100,000 cycles. Single-crystal CMSX-4 or René N5 blades combined with film-cooled passages and YSZ thermal barrier coatings are the standard solution.

Heavy-duty industrial gas turbines: Class-H machines such as Mitsubishi Heavy Industries M701JAC, GE 9HA.02 and Siemens SGT5-9000HL push turbine inlet temperatures of 1600-1700°C in 400 MW combined-cycle plants. Unlike aero engines, these run for thousands of hours per start, so the long hold makes creep the dominant damage mechanism. The Master Curve (LMP vs. stress) sets the inspection interval, and periodic fluorescent penetrant inspection (FPI) checks for β-phase spallation and intergranular cracks.

Rocket combustion chambers and regenerative-cooled nozzles: The SpaceX Raptor main chamber, Aerojet Rocketdyne RS-25 (SLS Core Stage) and AR1 expose their inner wall to 3300°C combustion gas while cooling the outer wall down to 20 K with liquid methane or LH2. A 1 mm wall thus carries a 3000°C drop — a brutal OP-type TMF case that produces the classic "doghouse effect" cracks in Cu-Cr-Nb (GRCop-84) and Inconel 718 walls. Reusable rockets target 100 flights as 100 TMF cycles.

Nuclear and fusion structural materials: LWR vessel nozzles, sodium-cooled fast-reactor fuel assemblies and the tungsten divertor of ITER all experience TMF coupled with neutron-irradiation-induced microstructural change. The W-Cu interface in a fusion divertor receives intermittent 10 MW/m² heat flux, evaluated as IP TMF. Design codes such as RCC-MRx and ASME III Sec. NH formally combine creep and fatigue damage via a Miner-style rule.

Common misconceptions and pitfalls

The first major pitfall is extrapolating room-temperature LCF data directly to high temperature. The Coffin-Manson fatigue ductility exponent c is typically around -0.6 at room temperature, but at 800°C and above the creep-fatigue interaction shifts it effectively to -0.7 to -0.8, making strain amplitude bite harder. Hold time effects are absent from LCF tests altogether. The N_LCF here is only the pure fatigue contribution from mechanical + thermal strain; predicting service life requires adding creep and oxidation via Miner's rule. Do not design directly from LCF test data.

The second is assuming the Larson-Miller constant C = 20 fits every material. The constant inside the LMP equation is experimentally fitted: 20 for many Ni-base superalloys, but 15-17 for stainless steels, 20-25 for Ti alloys, and 30 for ferritic creep-resistant steels (P91 family). The same temperature and time then give LMP values that differ by 2-3 units, throwing life predictions off by orders of magnitude. Real design must use the C value matched to the alloy, microstructure and heat treatment from databases such as NIMS. The values in this tool are educational defaults, not certification-grade numbers.

Finally, "if there is a TBC, the base alloy TMF doesn't matter" is a dangerous misconception. The YSZ thermal barrier does drop substrate surface temperature by 100-200°C, but the TBC itself has a thermal cycling life — spallation typically at 3000-5000 cycles, driven by oxidation stress at the bond coat (NiCoCrAlY) / TGO (thermally grown Al₂O₃) interface. That is often shorter than the base alloy TMF life, and many fleets in fact retire blades on TBC spallation. This simulator covers only the base alloy; in service, use the shorter of TBC spallation and base alloy TMF lives.

How to Use

  1. Enter maximum temperature (°C) for the thermal cycle—typical gas turbine blade operating range is 850–1100°C
  2. Enter minimum temperature (°C); for air-cooled blades, use 200–400°C to simulate cooling phases
  3. Input mechanical strain amplitude (%) applied during cycling; aerospace applications commonly use 0.3–0.8%
  4. Set hold time (seconds) at peak temperature to capture dwell-induced creep damage; 60–300 s is standard for jet engines
  5. Click Calculate to generate LCF life, creep contribution, and total TMF life using Larson-Miller parameter correlation

Worked Example

Nickel-superalloy turbine blade (IN718): T_max=950°C, T_min=350°C, strain amplitude=0.5%, hold time=120 s. Thermal strain amplitude calculated as 0.38% (CTE difference between 950–350°C ≈ 12×10⁻⁶/K). Total strain amplitude becomes 0.65%. LCF life N_LCF ≈ 8,500 cycles (Coffin-Manson exponent −0.7). Creep contribution over 120 s dwell ≈ 2,100 cycles (stress-rupture model). Total life N_total ≈ 5,200 cycles, LMP ≈ 22,400 (log-stress normalized parameter). Blade expected to survive ~8.7 operating hours at 10 Hz cycling.

Practical Notes

  1. Thermal strain dominates for large ΔT; a 700°C excursion in cast superalloy generates ~0.42% inelastic strain independent of applied mechanical load
  2. Hold time at peak temperature triggers stress relaxation and accelerated creep cavitation—doubling dwell from 60 to 120 s typically reduces total life by 30–45%
  3. LMP values above 24,000 indicate creep-controlled failure; below 20,000, low-cycle fatigue governs—design accordingly for mission profiles
  4. Out-of-phase TMF (temperature minimum during peak strain) is more damaging than in-phase; this simulator uses in-phase by default for conservative estimates