Variable amplitude fatigue analysis using S-N curves and Miner's rule. Real-time calculation of cumulative damage, safety factor, and remaining life for steel, aluminum, and CFRP with Goodman mean stress correction.
What exactly is "fatigue damage"? I know metal can break from repeated bending, but how do we measure the "damage" before it fails?
🎓
Basically, fatigue damage is a measure of how much of a component's useful life has been "used up" by cyclic loading. We quantify it by comparing the number of applied cycles at a given stress level to the number of cycles the material can withstand at that stress before failure. In this simulator, you can see this directly: try setting a high Stress Amplitude S₁ and watch the Damage Meter jump—it shows how a few high-stress cycles consume a lot of life.
🙋
Wait, really? So if a part experiences different stress levels, like a car suspension on smooth roads and potholes, how do we add up the damage?
🎓
Exactly! That's where Miner's rule comes in. It's a simple but powerful idea: we sum the damage fractions from each stress level. For instance, if potholes (high stress, S₂) use 0.4 of the life and smooth roads (lower stress, S₁) use 0.2, the total damage is 0.6. Try it: use the two load blocks in the simulator. Set different combinations of S₁, n₁ and S₂, n₂ and see how the cumulative damage D changes in real-time.
🙋
I see the "Mean Stress" and "Ultimate Strength" parameters. What's their role? Does a constant pulling force make fatigue worse?
🎓
Great observation! Yes, a constant tensile mean stress (like pre-tension) makes the alternating stress more damaging. We correct for this using the Goodman rule. It gives us an "equivalent" alternating stress that would cause the same damage if the mean stress were zero. Slide the Mean Stress σₘ up while keeping S₁ constant—you'll see the Equivalent Stress increase and the predicted life N₁ drop, raising the damage meter faster.
Physical Model & Key Equations
The core model is the S-N (Stress vs. Cycles to failure) curve, which describes the material's fatigue strength. It's often plotted on a log-log scale and can be represented by a power law.
Here, $N_i$ is the number of cycles to failure at the equivalent stress amplitude $S_{a,eq}$. $\sigma_f'$ is the fatigue strength coefficient (intercept at 1 cycle), and $b$ is the fatigue strength exponent (negative slope). The simulator uses typical values: steel ($\sigma_f'$ = 1000 MPa, $b$ = −0.085) and aluminum ($\sigma_f'$ = 800 MPa, $b$ = −0.102).
To handle real-world loading with a non-zero mean stress, we use the Goodman correction to find an equivalent fully-reversed stress amplitude.
$$S_{a,eq}= \dfrac{S_a}{1 - \sigma_m / S_u}$$
$S_a$ is the applied stress amplitude, $\sigma_m$ is the mean stress, and $S_u$ is the material's ultimate tensile strength. This equation shows how tensile mean stress ($\sigma_m \gt 0$) increases the damaging effect ($S_{a,eq}\gt S_a$). Finally, Miner's linear damage rule sums the damage from all load blocks:
$$D = \sum_i \frac{n_i}{N_i}$$
$D$ is the total cumulative damage, $n_i$ is the number of applied cycles at a given stress level, and $N_i$ is the life at that level from the S-N curve. Failure is predicted when $D \geq 1$, though in practice a safety factor is applied.
Frequently Asked Questions
Use it when analyzing fluctuating loads with non-zero mean stress (e.g., preloaded bolts or welded joints with residual stress). When the mean stress is tensile, fatigue life decreases, so without correction, the evaluation becomes non-conservative. This tool applies it automatically.
Yes, the fatigue strength coefficient σf' and fatigue strength exponent b for each material are freely editable by the user. By adjusting them to match actual measurement data or literature values, more accurate life predictions are possible. Default values are set to representative values.
The remaining life is the number of cycles until the current cumulative damage D reaches 1. The safety factor is the ratio of the current D to the allowable damage value (typically 1.0); a value below 1 indicates safety, while a value of 1 or above indicates a risk of fatigue failure. In design, a safety factor of 0.5 to 0.7 is often targeted.
Please correctly pair each stress amplitude with its corresponding number of occurrences. Ensure the stress amplitude unit is unified to MPa, and set the number of occurrences based on the actual operating cycles of the machine. Additionally, data where the stress amplitude is below the material's fatigue limit can be ignored.
Real-World Applications
Aerospace Component Life Prediction: Aircraft wings and landing gear experience a complex spectrum of loads during takeoff, turbulence, and landing. Engineers use tools like this to analyze recorded flight data, summing damage from millions of cycles to schedule mandatory inspections and part replacements before critical damage accumulates.
Automotive Suspension & Chassis Design: Cars are subjected to variable amplitude loading from road surfaces. A common case is using a standardized "proving ground" load spectrum to simulate years of driving in a short test. This calculator helps designers choose materials and dimensions to ensure the chassis survives the warranty period with an adequate safety factor.
Wind Turbine Blade & Tower Analysis: These structures face random, high-cycle loading from wind gusts and rotation. Fatigue analysis is often the governing design criterion. Engineers define load spectra for different wind speeds and use Miner's rule to estimate a 20+ year service life, crucial for the economic viability of the turbine.
Railway & Bridge Infrastructure Monitoring: For railway axles and steel bridges, each passing train applies a distinct load block. By monitoring traffic (stress cycles n_i) and knowing the S-N curve of the steel, maintenance teams can calculate cumulative damage in real-time, moving from fixed-interval inspections to condition-based, predictive maintenance.
Common Misconceptions and Points to Note
When starting to use this tool, there are several pitfalls that beginners in particular tend to fall into. First and foremost is the misconception that the calculation results guarantee an absolute lifespan. As experienced engineers often say, "Simulations are just 'paper calculations.' Reality is much harsher." For example, just because you select a steel material in the tool and it shows a safety factor of 1.5, you shouldn't proceed with the design as is. Actual products always have factors not included in the calculation, such as corrosion, manufacturing defects (like fine cracks in weld beads), and unexpected overloads. In practice, it's common to multiply the safety factor obtained from the calculation by an additional "experience factor."
Secondly, there's the quality of the load spectrum input. While the tool allows you to choose from pre-prepared spectra, in actual design work, you often need to create your own based on measured data or standards. Here, it's a major mistake to "ignore" low-stress level cycles. For instance, subjecting a material to 10 MPa of small vibration for 1 million cycles might seem to only slightly increase the D value, but microscopic cracks are definitely initiating and propagating inside the material. If a large load is then applied once in this "pre-conditioned" state, a "pre-conditioning effect" can occur, leading to fracture sooner than predicted.
Finally, consider the reliability of material constants. The values for σ_f' and b registered in the tool are representative values. In actual materials, fatigue strength varies due to fluctuations in heat treatment and lot-to-lot differences. For example, even for the same steel grade like "S45C," it's not uncommon for the fatigue life to vary by several times depending on the quenching and tempering conditions. For reliable design, it's ideal to obtain fatigue test data for the materials you procure and customize the tool's parameters accordingly.
Enter stress amplitude S₁ (MPa) and applied cycle count n₁ for the first load level, then S₂ and n₂ for the second stress amplitude.
Input material S-N curve parameters: slope exponent m and fatigue strength coefficient σ_f from your material datasheet (e.g., aluminum 7075-T73: m=8, σ_f=450 MPa).
The calculator computes allowable cycles N₁ and N₂ using the power-law S-N relationship, calculates partial damage fractions D₁=n₁/N₁ and D₂=n₂/N₂, then sums cumulative damage D via Miner's linear damage rule.
Review Safety Ratio (1/D): values >1.0 indicate no failure; <1.0 predicts fatigue crack initiation.
Worked Example
Steel shaft (E=210 GPa, S-N curve: m=9, σ_f=800 MPa) subject to variable loading: 15,000 cycles at 400 MPa stress amplitude, then 8,000 cycles at 250 MPa. First level yields N₁=200,000 cycles (D₁=0.075); second level yields N₂=5,200,000 cycles (D₂=0.0015). Cumulative damage D=0.0765, Safety Ratio=13.07, indicating robust margin. If operating at 2×10⁶ cycles/year, estimated remaining life=67 years before 100% damage accumulation.
Practical Notes
Miner's rule assumes no load sequence effects; conservative for variable-amplitude helicopter rotor blade spectra where high-stress cycles occurring early reduce subsequent crack growth resistance.
Validate S-N exponent m using material test data; typical steel m=8–12, aluminum m=5–8, composites m=3–6.
Input actual service duty cycles from flight data recorders or mission profiles to avoid non-conservative life predictions in offshore wind turbine gearboxes.
Account for environmental factors (corrosion, temperature cycling) by reducing σ_f by 10–30% for seawater-exposed components.