Fatigue: S-N Curves and Endurance Limit

Fatigue failure is a silent threat in engineering, causing components to fracture under repeated loads far below their static strength. Understanding how materials behave under cyclic loading is crucial for designing durable structures like aircraft wings, bridges, and engine crankshafts. S-N curves and the endurance limit provide the tools to predict and prevent fatigue-related disasters.

The Fundamentals of Fatigue Failure

Fatigue is the progressive and localized structural damage that occurs when a material is subjected to cyclic loading. Unlike sudden failure from a single overload, fatigue develops over time through the initiation and growth of tiny cracks. The classic example is bending a paperclip back and forth; it doesn't break on the first bend, but after many cycles, it snaps. In engineering, this is critical because components experience millions of load cycles during their service life, from vibrating machinery to rotating shafts. The primary tool for characterizing this behavior is the stress-life approach, summarized by the S-N curve.

To analyze fatigue, engineers define stress amplitude ($S_a$), which is half the difference between the maximum and minimum stress in a cycle. The mean stress ($S_m$) is the average of these two values. For many initial analyses, a fully reversed loading cycle is assumed, where the mean stress is zero and the stress amplitude is the sole variable. The number of cycles to failure ($N_f$) is then recorded for different stress amplitudes, forming the basis of the S-N relationship.

Decoding the S-N Curve

An S-N curve, or Wöhler curve, is a graphical plot of stress amplitude ($S_a$) versus the logarithm of the number of cycles to failure ($N_f$). The vertical axis typically represents stress (in MPa or ksi), and the horizontal axis uses a logarithmic scale for cycles. To generate this curve, standardized specimens are tested under controlled cyclic loading at various stress levels until failure. Each data point represents one test, and the resulting curve shows a clear trend: as the applied stress amplitude decreases, the number of cycles to failure increases.

For many materials, the S-N curve appears as a downward-sloping line on a log-log plot. At high stress amplitudes, failure occurs in a low number of cycles (less than about $10^4$ cycles), known as low-cycle fatigue. At lower stress amplitudes, the curve may flatten out. The most important feature for design is identifying the stress level below which failure does not occur, or occurs only after an extremely high number of cycles. This concept leads directly to the endurance limit.

The Endurance Limit in Ferrous Metals

Many ferrous metals (iron-based alloys like steels) exhibit a true endurance limit (also called the fatigue limit). This is a stress amplitude below which the material can withstand an essentially infinite number of cycles without failing. For practical purposes, "infinite life" is often defined as surviving beyond $10^6$ or $10^7$ cycles. The endurance limit for many steels typically falls between 40 to 60 percent of the material's ultimate tensile strength (UTS). For example, a steel with a UTS of 600 MPa might have an endurance limit of approximately 300 MPa.

This property is a major advantage in design. If you can ensure that the operational stress amplitude remains below the endurance limit, the component can be considered safe from fatigue failure for its intended lifespan. However, the endurance limit is not an intrinsic material property; it is derived from laboratory testing under ideal conditions. It assumes a perfectly polished specimen, uniaxial loading, and no mean stress. In real-world applications, various factors reduce this ideal limit, which we will explore later.

Fatigue Behavior of Non-Ferrous Metals

In contrast to ferrous metals, most non-ferrous metals like aluminum, copper, and titanium alloys do not exhibit a true endurance limit. Their S-N curves continue to slope downward indefinitely, even at very high cycle counts. This means that for any non-zero stress amplitude, failure will eventually occur if enough cycles are applied. Consequently, design with these materials requires a fatigue life specification.

Instead of aiming for infinite life, you must specify a finite life—say, $10^7$ or $10^8$ cycles—and determine the corresponding allowable stress amplitude from the S-N curve. This is often called the fatigue strength at a given number of cycles. For instance, an aluminum alloy might have a fatigue strength of 150 MPa at $10^7$ cycles. This approach is common in aerospace and automotive industries where weight savings using aluminum outweigh the need for infinite life, and precise life calculations are mandatory.

Advanced Considerations and Modifying Factors

The textbook S-N curve and endurance limit are derived from ideal laboratory specimens. In practice, you must account for several factors that significantly reduce the fatigue strength of real components. These are often incorporated using modifying factors in the modified Goodman equation or similar models that adjust the endurance limit.

Key factors include:

• Surface Finish: Rough surfaces act as stress concentrators and initiate cracks. A polished specimen has the highest fatigue strength, while a forged or corroded surface can reduce it by 50% or more.
• Size Effect: Larger components have a greater probability of containing flaws, so their fatigue strength is lower than that of small test specimens.
• Loading Type: The endurance limit for bending is typically higher than for axial (tension-compression) loading.
• Mean Stress: Non-zero mean stress (e.g., a constant tension plus a cyclic load) alters fatigue life. Compressive mean stress is generally beneficial, while tensile mean stress is detrimental. The Goodman diagram is a common tool for analyzing this effect, using the relationship:

$$\frac{S_a}{S_e}+\frac{S_m}{S_{ut}}=1$$ where $S_a$ is the stress amplitude, $S_e$ is the endurance limit for zero mean stress, $S_m$ is the mean stress, and $S_{ut}$ is the ultimate tensile strength.

• Environmental Effects: Corrosion, high temperature, or fretting can drastically accelerate fatigue damage.

A worked example clarifies this. Suppose you have a steel shaft with an endurance limit $S_e$ of 250 MPa (from polished lab data) and an ultimate strength $S_{ut}$ of 600 MPa. In service, it has a surface finish factor of 0.8 and a size factor of 0.9. The corrected endurance limit for design is $250\times 0.8\times 0.9=180$ MPa. If the shaft experiences a stress amplitude of 150 MPa with a mean stress of 100 MPa, you check the Goodman criterion: $$\frac{150}{180}+\frac{100}{600}=0.833+0.167=1.0$$ This indicates the design is at the fatigue failure boundary, so you might need to reduce stresses or improve the surface finish.

Common Pitfalls

1. Assuming the Textbook Endurance Limit Applies Directly: The most common error is taking the endurance limit from a material data sheet and applying it directly to a real part. Always remember to apply the appropriate modifying factors for surface finish, size, and loading conditions to obtain a realistic design strength.

1. Treating Non-Ferrous Metals as Having an Endurance Limit: Designing an aluminum component for "infinite life" below an assumed endurance limit is a critical mistake. For non-ferrous metals, you must always use the S-N curve to specify a finite fatigue life and its corresponding fatigue strength.

1. Ignoring Mean Stress Effects: Analyzing only the stress amplitude while neglecting mean stress can lead to non-conservative designs. Always use a mean stress correction model, like the Goodman or Gerber criteria, when the cyclic load is not fully reversed.

1. Confusing Fatigue Strength with Ultimate Strength: Fatigue failure occurs at stress levels far below the ultimate tensile strength. A component that is perfectly safe under a single static load can fail prematurely under repeated loading. Never use static strength values to judge fatigue performance.

Summary

• S-N curves are the fundamental tool for stress-life fatigue analysis, plotting stress amplitude against the logarithm of cycles to failure.
• Many ferrous metals like steel exhibit an endurance limit, a stress amplitude below which infinite fatigue life is expected, typically 40–60% of the ultimate tensile strength.
• Most non-ferrous metals like aluminum do not have a true endurance limit; design requires specifying a fatigue life and using the S-N curve to find the allowable stress.
• Real-world fatigue strength is significantly lower than laboratory data due to factors like surface finish, size, mean stress, and environment, which must be accounted for with modifying factors.
• Always correct the endurance limit for actual conditions and use mean stress analysis to avoid non-conservative designs that could lead to catastrophic fatigue failure.