Fatigue Resistance of Piping Materials

The ability of a metal to resist crack initiation and further propagation under repeated cyclic loading is a measure of its fatigue resistance. Everything with Piping & Fabrication will continuing the explanation of this post and this is full explanation of fatique resistance. 

Several standardized test methods have been developed to test metals, machined to particular geometries, where applying a repeating load range. Loads are generally applied through bending, cantilevered, or push-pull load application in suitably outfitted testing machines. Either constant applied stress or strain ranges can be employed to determine material response.

The most common representation of fatigue test data is an S-N curve, relating stress (S) required to cause specimen failure in a given number of cycles (N) (Fig. A3.10a). These tests are generally performed on smooth specimens, but they can also be run with stress-concentrating mechanisms such as notches machined into the specimen surface. The effect of stress concentrations on fatigue life cycles can also be estimated from the smooth specimen S-N curve by calculating the intensified stress due to the particular geometry, and intersecting the curve at that point on the stress axis.

FIGURE A3.8 Transition temperature range and transition temperature in Charpy impact test

FIGURE A3.9 Drop-weight test specimen with brittle weld deposit
on specimen face; machined notch to act as crack starter.
Impact load applied from side opposite weld deposit

As the applied load range decreases, ferritic steels exhibit a point at which an infinite number of cycles can be absorbed without causing failure. This level of stress is called the endurance limit. Many of the other metals do not exhibit this behavior, but rather exhibit an increasing, but finite, number of cycles to failure with decreasing cyclic load (Fig. A3.10b).

FIGURE A3.10 S-N curves that typify fatigue test results (a) for testing medium-strength
steels and (b) showing typical curve shape for ferrous and nonferrous materials. SL is the
endurance limit

The fatigue resistance of a material at a given applied stress or strain range is a function of a number of variables, including material strength and ductility. Results may vary significantly for different surface finishes, product forms of the same material (Fig. A3.11), material internal cleanliness, test specimen orientation, and levels of residual stress, among other factors. Variations in the test environment can also have a profound effect on test results (Fig. A3.12). Therefore, fatigue test results characteristically exhibit significant scatter.

Fatigue design curves are generated from test data by applying large safety margins to the average property curve. In U.S. design codes, the fatigue design curve is commonly generated by taking the lesser of 1/20 times the cycles to failure, or 1/2 of the stress to cause failure. A new curve is constructed taking the lower bound of these two factored curves.

FIGURE A3.11 Fatigue characteristics (S-N curve) for cast and wrought 1040 steel in the normalized
and tempered condition, both notched and unnotched. R. R. Moore rotating beam tests, Kt 2.2

When considering metal fatigue in design, a further safety margin is often also applied against the cycles-to-failure at a given stress amplitude. For example, if a component is continuously cycled over the same stress range, a design limit on allowable cycles may correspond to the cycle life multiplied by a factor such as 0.8. This is a common safety margin employed in vessel and piping design. As is normally the case, components may experience a wide variety of cyclic stress ranges, at various temperatures, over their life. The effect of this array of cyclic parameters on fatigue life can be estimated by an approach referred to as life fraction summation. In this design practice, the percentage of life used up in cycling at a certain stress range is calculated, corresponding to the ratio of the number of actual service duty cycles to the total number of cycles to failure at that stress range. This calculation is performed for all of the various stress ranges/duty cycles anticipated. The fractions thereby calculated are summed and compared to the design limit (1.0 with no safety margin, or 0.8 or some other value depending on the design safety factor that applies).