Fatigue Failure Theories

The Invisible Killer of Metals

A material might safely carry a 100kg load statically forever. But apply that same 100kg load, remove it, and reapply it a million times... and the component suddenly snaps. This is Fatigue: failure due to cyclic loading at stresses well below the yield strength.

Understanding Fluctuating Stresses

In a typical fatigue scenario (like a rotating axle), the stress goes through maximum ($\sigma_{max}$) and minimum ($\sigma_{min}$) peaks. We characterize this cycle using two key metrics:

Alternating Stress ($S_a$)

$$S_a = \frac{\sigma_{max} - \sigma_{min}}{2}$$

The amplitude of the fluctuation. This is the primary driver of fatigue crack growth.

Mean Stress ($S_m$)

$$S_m = \frac{\sigma_{max} + \sigma_{min}}{2}$$

The average background stress. Tensile mean stress accelerates failure, while compressive retards it.

The Wöhler (S-N) Curve

Developed by August Wöhler, the S-N curve plots alternating stress ($S_a$) against the number of cycles to failure ($N_f$) on a logarithmic scale.

  • Low Cycle Fatigue ($N < 10^3$): Governed by plastic deformation. High stress, few cycles.
  • High Cycle Fatigue ($10^3 < N < 10^6$): Elastic deformation dominates. Classic fatigue regime.
  • The Endurance Limit ($S_e$): Notice the "knee" in the graph for ferrous metals (like steel) around $10^6$ cycles. If the stress is kept below $S_e$, the material has an theoretically infinite life! Non-ferrous metals (like Aluminum) do not have this horizontal knee.

Mean Stress Failure Envelopes

The S-N curve assumes Mean Stress ($S_m = 0$)—a fully reversed cycle. But in reality, components often operate with a non-zero mean stress. To predict failure under combined $S_a$ and $S_m$, we use the Haigh Diagram.

Soderberg (Most Conservative)

$$\frac{S_a}{S_e} + \frac{S_m}{S_{yt}} = \frac{1}{N_f}$$

Guards against both fatigue and static yielding.

Modified Goodman (Standard for Design)

$$\frac{S_a}{S_e} + \frac{S_m}{S_{ut}} = \frac{1}{N_f}$$

Connects endurance limit to ultimate strength. widely used for brittle materials and conservative design.

Gerber (Parabola)

$$\frac{S_a}{S_e} + \left(\frac{S_m}{S_{ut}}\right)^2 = \frac{1}{N_f}$$

Fits experimental failure data best for ductile metals.

✨ AI Fatigue Design Analyzer

Describe a cyclically loaded component and let the Gemini AI act as your Lead Reliability Engineer.