What is an S-N Curve? History, Physics, FKM Estimation, and Fatigue Life Applications

About the Author

Johnny Liu is the Chief Executive Officer at Dowway Vehicle. With over twenty years of practical experience in automotive structural durability, chassis design, and vehicle safety engineering, Johnny has led numerous vehicle platform programs. His work focuses on applying fatigue-life estimation and material dynamics to ensure high-performance structural integrity in modern transportation.

Overview

For mechanical and automotive engineers, predicting when a component will fail under cyclic loading is a main focus of structural durability. The S-N Curve (also called the Wöhler Curve) serves as the main tool to map a material’s alternating stress amplitude ($S$) against its cycles to failure ($N$).

This guide connects historical basics with modern computer-aided engineering (CAE) workflows. You will learn the history behind fatigue science, how to read S-N curves, the physics of different lifespan zones, how to estimate curves using FKM Analytical Guidelines, and how to run accelerated fatigue testing using the logarithmic slope ($k$-factor).

1. The History of S-N Curves: The Versailles Train Crash

To understand why we use S-N curves today, we must look back to the mid-19th century.

In 1842, a catastrophic train derailment and subsequent fire occurred in Versailles, France. The cause was a sudden, unexpected fracture of a locomotive axle. What baffled engineers of the era was that the axle was designed to easily withstand the static weight of the train. It had not bent, buckled, or exceeded its ultimate tensile strength. Yet, under the repeated, low-level cyclic stress of daily operations, the axle broke.

German engineer and scientist August Wöhler (1819–1914) took on the task of investigating this phenomenon. Through careful laboratory recreations, Wöhler discovered that:

• Microscopic cracks initiated on the surface of the axle due to cyclic loads.
• These cracks propagated slowly over time.
• Once the crack reached a critical size, the remaining cross-sectional area of the axle could no longer support the load, leading to sudden, catastrophic failure.

Wöhler developed a custom test rig to apply repeated bending loads to axles, plotting the applied cyclic stress level against the number of cycles it took for the specimen to fail. This pioneering plot became known as the Wöhler Curve, or modernly, the S-N Curve. It established that static load capacities are insufficient for dynamic engineering; repetitive low-stress cycles can kill a component just as easily as a single massive overload.

2. Laboratory Generation of S-N Curves

Today, materials scientists and fatigue engineers generate S-N curves using standardized rotating-beam or axial fatigue testing machines.

The testing workflow works as follows:

1. Specimen Preparation: Standardized dog-bone metal specimens are polished to eliminate surface roughness, which acts as a stress concentrator.
2. Cyclic Loading: A specimen is placed into the machine and subjected to a sinusoidal alternating stress history of a designated stress amplitude ($S_a$).
3. Recording Failure ($N$): The machine runs until the specimen completely fractures or develops a macro-crack, recording the final number of cycles ($N$).
4. Multi-Point Testing: Because fatigue is highly statistical, engineers must repeat this test at multiple different stress amplitudes to build a continuous dataset.

When plotted, the data points reveal a distinct trend: higher stress amplitudes result in fewer cycles to failure, causing the S-N curve to slope downwards from the top-left to the bottom-right.

The Loading Frequency

In classical fatigue analysis, loading frequency is not considered a primary driver of fatigue failure. The damage is driven strictly by the number of stress reversals (cycles), not how quickly they occur.

However, in real-world applications, frequency does play an indirect role:

• Resonance: If the excitation frequency matches the component’s natural frequency, the stress amplitudes are highly amplified, accelerating fatigue damage.
• Hysteretic Heating: At extremely high frequencies (e.g., ultrasonic fatigue testing), polymeric or even metallic materials may experience internal heating, which degrades material properties.

3. The Three Lifespan Zones of an S-N Curve

A typical S-N curve is divided into three distinct regions based on the material behavior and stress limits: the Plastic Zone, the Elastic Zone, and the Infinite Life Zone.

To understand these zones, we first define three critical stress benchmarks, which can be derived from static tensile and stress-strain testing:

• Ultimate Tensile Strength ($S_{ut}$ or $S_{max}$): The maximum stress a material can withstand. On the S-N curve, this represents failure at a single cycle ($N = 1$).
• Yield Strength ($S_y$): The boundary stress where a material transitions from elastic (temporary) deformation to plastic (permanent) deformation.
• Endurance Limit / Fatigue Limit ($S_E$): The threshold stress below which the material can theoretically withstand infinite cycles without failing.

3.1. The Plastic Zone (Low-Cycle Fatigue – LCF)

• Cycle Range: Typically $N < 10^3$ or $10^4$ cycles.
• Stress Level: Exceeds the material’s yield strength ($S > S_y$).

In this region, the component undergoes macroscopic plastic deformation during each load cycle. S-N curves are not recommended for calculating fatigue life in the plastic zone. Because strain is no longer linearly proportional to stress, engineers instead utilize the Strain-Life (E-N) method, which plots strain amplitude ($\epsilon$) against cycles. E-N methods are highly sensitive to the order in which loads are applied (sequence effects).

3.2. The Elastic Zone (High-Cycle Fatigue – HCF)

• Cycle Range: Typically $10^3 < N < 10^6$ cycles.
• Stress Level: Well below the yield strength but above the endurance limit ($S_E < S < S_y$).

Within the elastic zone, the macroscopic stress-strain relationship remains linear and reversible. When the load is removed, the component returns to its original geometry. However, micro-plasticity still occurs at localized stress concentrations (micro-voids, surface scratches, or sharp geometric transitions).

• Critical Drivers: High-cycle fatigue life is highly dictated by surface roughness, residual stresses (e.g., tensile residual stresses from welding accelerate crack propagation, while compressive residual stresses from shot-peening delay it), and geometric stress concentration factors ($K_t$).

3.3. The Infinite Life Zone (Endurance Limit)

• Cycle Range: Typically $N > 10^6$ to $10^7$ cycles.
• Stress Level: Below the Endurance Limit ($S < S_E$).

For critical components that run continuously at high speeds (e.g., automotive engine crankshafts, connecting rods, and transmission gears), design engineers target the Infinite Life Zone. All operational cyclic stresses must remain strictly below the endurance limit $S_E$ to prevent fatigue cracking over millions of cycles.

Do All Materials Have an Endurance Limit?

No. There is a fundamental metallurgical difference between ferrous and non-ferrous metals:

• Ferrous Metals (Steel, Cast Iron): Exhibit a distinct horizontal plateau (Endurance Limit) on the S-N curve. If stress is kept below this limit, the component can theoretically run infinitely.
• Non-Ferrous Metals (Aluminum, Magnesium, Copper Alloys): Do not exhibit a distinct endurance limit. Their S-N curves continue to slope downward indefinitely, albeit at a gentler slope. For these materials, engineers define a pseudo-endurance limit at a specific high cycle count (typically $N = 10^7$ or $10^8$ cycles).

When Infinite Life Fails

Even if a component is designed below its theoretical endurance limit, infinite life can be negated by:

• Corrosion: Environmental chemical attack continuously degrades the surface, preventing the protective oxide layer or dislocation barriers from halting crack initiation.
• High-Temperature Creep: Elevating temperatures lowers the yield and fatigue strength thresholds.
• Extremely High Cycles (Gigacycle Fatigue): Beyond $10^9$ cycles, subsurface inclusions can still initiate cracks even under very low stress amplitudes.

To determine if a real-world component with non-zero mean stress sits safely in the infinite life zone, engineers plot stress amplitudes against mean stresses using the Goodman-Haigh Diagram.

4. Analytical S-N Curve Estimation (FKM Guidelines)

Physical fatigue testing is exceptionally time-consuming and expensive. The minimum standard to construct a reliable experimental S-N curve requires testing at 5 different stress levels with at least 3 repeat trials per level.

When physical test data is unavailable (e.g., during early design phases of a new alloy), engineers use the Universal Slope Method or the FKM Guideline (Analytical Strength Assessment) published by the German Mechanical Engineering Industry Association (VDMA).

Based on the FKM guideline, the fatigue limit ($S_E$) is analytically derived from the material’s ultimate tensile strength ($S_{max}$ or $S_{ut}$). Along with the transition cycle limit ($N_E$) and the logarithmic slope exponent ($k = 5$), the entire S-N curve can be mathematically estimated.

Below is the complete mathematical library of standard vs. FKM conservative estimation formulas for various engineering metals:

Table 1: Material S-N Curve Estimation Library

Material Type	Standard Estimation Formula	Transition Cycles ($N_E$)	Slope ($k$)	FKM Conservative Formula	Transition Cycles ($N_{E, FKM}$)	Slope ($k$)
Steel	$S_E = 0.5 \cdot S_{max}$ (if $S_{max} \le 1400 \text{ N/mm}^2$)$S_E = 700 \text{ N/mm}^2$ (if $S_{max} > 1400 \text{ N/mm}^2$)	$5 \times 10^6$	$5$	$S_E = 0.385 \cdot S_{max} + 30 \text{ N/mm}^2$	$1 \times 10^6$	$5$
Cast Iron	$S_E = 0.4 \cdot S_{max}$ (if $S_{max} \le 500 \text{ N/mm}^2$)$S_E = 200 \text{ N/mm}^2$ (if $S_{max} > 500 \text{ N/mm}^2$)	$5 \times 10^6$	$5$	$S_E = 0.32 \cdot S_{max}$	$1 \times 10^6$	$5$
Aluminum	$S_E = 0.4 \cdot S_{max}$ (if $S_{max} \le 325 \text{ N/mm}^2$)$S_E = 130 \text{ N/mm}^2$ (if $S_{max} > 325 \text{ N/mm}^2$)	$1 \times 10^8$	$5$	$S_E = 0.22 \cdot S_{max}$	$1 \times 10^8$	$5$
Magnesium	$S_E = 0.35 \cdot S_{max}$ (if $S_{max} \le 325 \text{ N/mm}^2$)$S_E = 130 \text{ N/mm}^2$ (if $S_{max} > 325 \text{ N/mm}^2$)	$1 \times 10^8$	$5$	$S_E = 0.22 \cdot S_{max}$	$1 \times 10^8$	$5$
Titanium Alloy	$S_E = 0.55 \cdot S_{max}$ (if $S_{max} \le 1130 \text{ N/mm}^2$)$S_E = 620 \text{ N/mm}^2$ (if $S_{max} > 1130 \text{ N/mm}^2$)	$5 \times 10^6$	$5$	N/A	N/A	N/A
Copper / Nickel	$S_E = 0.4 \cdot S_{max}$ (if $S_{max} \le 750 \text{ N/mm}^2$)$S_E = 300 \text{ N/mm}^2$ (if $S_{max} > 750 \text{ N/mm}^2$)	$1 \times 10^8$	$5$	N/A	N/A	N/A

5. The Logarithmic Nature of S-N Curves: The 32x Damage Phenomenon

The most critical and counter-intuitive aspect of fatigue design is that S-N curves are highly non-linear. While they appear as straight lines on engineering charts, this is only because both axes (Stress $S$ and Cycles $N$) are displayed on a log-log (logarithmic-logarithmic) scale.

To demonstrate the massive real-world impact of this logarithmic relationship, let us analyze three distinct load histories using a standard steel material with a slope factor ($k = 5$):

Case Study: Evaluating Three Load Histories

• Load History #1: $300\text{ MPa}$ cyclic stress amplitude applied for $1,000\text{ cycles}$.
• Load History #2: $300\text{ MPa}$ cyclic stress amplitude applied for $2,000\text{ cycles}$ (Cycles are doubled).
• Load History #3: $600\text{ MPa}$ cyclic stress amplitude applied for $1,000\text{ cycles}$ (Stress amplitude is doubled).

Comparing History #1 vs. History #2 (Linear Relationship of Cycles)

Using the Palmgren-Miner rule of linear damage accumulation, cumulative damage ($D$) is calculated as: $$D = \sum \frac{n_i}{N_i}$$

Since the stress amplitude remains identical at $300\text{ MPa}$, the allowable cycles to failure ($N_i$) is the same. Doubling the applied cycles ($n_2 = 2 \cdot n_1$) directly doubles the accumulated damage: $$\text{Damage of History \#2} = 2 \times \text{Damage of History \#1}$$

Comparing History #1 vs. History #3 (Exponential Relationship of Stress)

Now, we keep the cycle count constant at $1,000\text{ cycles}$ but double the stress amplitude from $300\text{ MPa}$ to $600\text{ MPa}$.

Due to the log-log scale of the S-N curve, the damage scale factor is governed by the exponent $k$ (where $k = 5$ for steel): $$\text{Damage Scale Factor} = (\text{Stress Amplitude Ratio})^k$$$$\text{Damage Scale Factor} = \left(\frac{600\text{ MPa}}{300\text{ MPa}}\right)^5 = (2)^5 = 32$$

Therefore, doubling the stress amplitude does not double the damage—it increases the accumulated structural damage by exactly 32 times!

This non-linear relationship is why a minor miscalculation in operational load estimation or structural weight can lead to premature structural failures in field operations.

6. S-N Curve Slope: The “k-Factor”

The slope of the S-N curve plotted in a log-log coordinate system is mathematically represented by the $k$-factor (Wöhler slope exponent).

The mathematical relationship linking stress amplitude ($S$) and fatigue life ($N$) is given by: $$N \cdot S^k = C \quad \text{or} \quad N_1 \cdot S_1^k = N_2 \cdot S_2^k$$

Where $C$ is a material-specific constant.

The value of the $k$-factor determines how sensitive a material’s fatigue life is to variations in loading. As $k$ increases, even minor increases in stress amplitude result in massive drops in fatigue life.

Table 2: Typical $k$-Factors across Engineering Applications

• Aluminum Alloys: $k \approx 7$ (Lower sensitivity to high-stress peaks, but fatigue life continues to decay).
• Structural/Cast Steels: $k \approx 5$ (Standard metallic behavior in the elastic HCF zone).
• Welded Joints: $k \approx 3$ (Highly sensitive. Because welds contain micro-voids and crack-like geometries, crack propagation dominates the fatigue life immediately, leading to a steep slope).

Utilizing $k$-Factors for Accelerated Durability Testing

In automotive engineering, testing a vehicle chassis for $500,000\text{ km}$ of simulated road travel is economically unfeasible. By leveraging the $k$-factor, testing engineers can compress testing time through load amplification.

If we increase the testing load (stress amplitude) by a mere 15% under a $k = 5$ steel component: $$\text{Damage Increase} = (1.15)^5 \approx 2.01$$

A 15% load increase doubles the fatigue damage, which means the required physical test cycles can be cut in half (0.5x test duration) while maintaining the exact same failure mode. This enables fast, reliable validation of structural components.

7. S-N Curve Implementation in Simcenter Testlab Neo

In modern engineering laboratories, manually calculating cumulative fatigue damage from complex, random road-load time histories is impossible. Engineers use professional DAQ and analysis suites like Siemens Simcenter Testlab Neo to process strain gauge and load cell data.

The software utilizes S-N curve data within several core structural durability workflows:

1. Stress-Life (S-N) Analysis: Combines rainflow cycle counting with the component’s designated S-N curve to output expected life and cumulative damage.
2. Mean Stress Correction: Corrects the S-N curve based on static tensile mean stresses, utilizing built-in mathematical models like Goodman, Gerber, Soderberg, or Morrow.
3. Pseudo-Damage Calculation: Converts raw strain/force time histories into a standard “damage” metric using a reference S-N curve. This allows engineers to compare the severity of different test tracks directly.
4. Range vs. Damage Analysis: Identifies which specific load cycle amplitudes (ranges) contributed to the majority of the cumulative damage, helping target structural redesigns.
5. Rotating Torque Pseudo-Damage: Evaluates cyclic torsional stress profiles in rotating drivetrains and axles.
6. Damage-Based Time Compression: Automatically synthesizes long, non-damaging periods out of real-world load histories, creating an accelerated, highly damaging “shaker test profile” for lab validation.

8. Frequently Asked Questions (FAQ)

Q1: What is the main difference between an S-N curve and an E-N curve?

The S-N curve is used for elastic, high-cycle fatigue, while the E-N curve is used for plastic, low-cycle fatigue. S-N curves operate in the elastic region where stress is proportional to strain ($N > 10^4$). E-N (Strain-Life) curves measure localized plastic strains at hot-spots or notch boundaries where local yielding occurs.

Q2: Why do aluminum alloys lack a true endurance limit?

Aluminum lacks a chemical mechanism to lock crystal dislocations at low stresses, meaning damage accumulates at any load level. Unlike steel, where carbon atoms block micro-slip paths under low loads, aluminum experiences microscopic dislocation glide even under tiny cyclic stresses. Because of this, it will eventually crack if cycled long enough.

Q3: How do compressive residual stresses affect the HCF zone of an S-N curve?

They force micro-cracks to stay closed, pushing the S-N curve upward to improve high-cycle life. Compressive forces (from shot-peening or cold work) physically counteract the tensile loads that pull cracks open. This successfully extends the material’s fatigue limit.

References

• [1] Radaj, D.; Vormwald, M.: Fatigue Strength (3rd Edition); Springer-Verlag Berlin Heidelberg, 2007.
• [2] FKM-Guideline: Analytical Strength Assessment of Components in Mechanical Engineering (5th Edition), VDMA Verlag, Frankfurt/M., 2003.
• [3] Fuchs, H.O.; Stephens, R.I.: Metal Fatigue in Engineering; John Wiley & Sons, New York, 1980.
• [4] Juvinall, R.C.: Stress, Strain, and Strength: Engineering Considerations; McGraw-Hill, New York, 1967.
• [5] Sonsino, C.M.; Hanselka, H.: Operational Design of Magnesium Components; Konstruktion 53 (2001) 12, pp. 55-58.