Fatigue is the reduction in the ultimate stress of a component due to repeated loading and unloading. For many applications, loading is not constant, and the materials and design must reflect the need to withstand potential fatigue from thousands or millions of cycles or more over time.
Defining Fatigue
In its simplest state, a bridge can be envisioned as a beam that is supported at each end. As a car passes over the bridge, the bridge is loaded by the weight of the car. The center of the bridge deflects, and the returns to its original state as the car leaves the bridge. This can be considered one loading cycle.
Generally, if a component is loaded, but the stress remains in the elastic range of a stress-strain diagram, no permanent deformation occurs. However, under significant cyclic loading, failure can occur at a much lower stress level than a freshly loaded component. Fatigue life is also influenced by environmental conditions, surface quality, and material properties.
There are two different types of fatigue. Low cycle fatigue damage occurs in systems with 1,000 cycles or less, and is more highly influenced by plastic deformations brought on by higher stresses. High cycle fatigue damage occurs in system with 1,000 or more cycles. Deformations are primarily elastic, but result in material changes over the long term.
Results of Fatigue
The primary result of fatigue in an engineered structure is crack growth and eventual fracture. Crack growth generally initiates at geometric weak points, such as a bolt hole or notch, or at microscopic impurities in the material. Cracks begin as very small points of damage, but each loading cycle extends the crack a bit more. When the crack length reaches a critical magnitude, the component will break.
Depending on the stress level, number of cycles, and material type, the mechanical properties of the material can remain unchanged or can undergo stress-related changes such as softening or hardening.
Displaying Fatigue Effects Graphically
For a given material, a graph can be generated that shows the relationship between the stress level (S) and the number of cycles to failure (N). An S-N curve displays the cycles to failure on a logarithmic scale, and stress on a standard scale.
S-N curves are generated experimentally by subjecting precise test specimens to cyclic loading. These test specimens are often called coupons. A series of coupon tests are performed using many different maximum stress levels to build an S-N curve. A large sample is also required because there can be significant variability in the performance of the coupons under high stress and/or high cycle conditions.
Sources
Beer, F., Johnston, E.R., Mechanics of Materials, Second Edition, McGraw-Hill, 1992.
 |
 |
