"Fatigue is caused by repeated application of stresses below the ultimate tensile strength of the material. A crack is initiated, grows, and leads to final failure when the remaining cross-section of material can't support the load anymore. One example might be something like an airplane wing that repeatedly flexes during takeoff, landing, and flying. Another example could be parts on a car going over bumps, stressing the engine mounts, exhaust system, drivetrain and anything which is not completely isolated from loading. Random vibration and thermal cycling can also produce fatigue damage.

High-cycle fatigue is when the cyclic stress is below the yield strength of the material, and failure occurs after more than ten thousand cycles. The majority of the component's lifetime is spent initiating cracks, which then slowly propagate until they become unstable, leading to failure. Delaying crack initiation prolongs life, and that means reducing stress concentrations that help cracks form - things like sudden changes in cross-section, sharp radii, notches, holes, and part defects that create 'crack starters.' To avoid high-cycle fatigue, parts should be designed with no sharp radii. Increasing the thickness or strength of the part to decrease the stress amplitude can help. Another common solution is shot peening or case hardening to produce compressive residual stresses in the part surface, which reduces the effective tensile stress under cyclic loading.

When we observe fatigue failure in less than ten thousand cycles, we call it low-cycle fatigue. The applied load results in stresses higher than the yield strength of the material, so that low-cycle fatigue cracks grow by plastic deformation. Thermal cycling is a common source of low-cycle fatigue. Crack initiation occurs early in the component's lifetime, and most of the low-cycle fatigue life will be spent in the crack growth phase. The rate of crack growth determines whether the crack remains harmless or leads to premature failure. In low-cycle fatigue, stress concentrations are still important, but not as important for high-cycle fatigue, and there is no benefit from compressive stresses at the surface, because the first few strain cycles wipe out the residual stresses. Simply using a thicker section or a stronger material can make low-cycle fatigue worse. A more ductile and tougher material, one that can accommodate more strain, can be helpful. Finding ways to reduce the load on the sensitive area is also effective. One way to do this is to provide sliding seals and bellows in high temperature machinery. Expansion joints in bridges is another example."