The use of fatigue data, generated using a simple periodic overload test sequence, can provide several advantages over conventionally measured fatigue data. These include more realistic and accurate fatigue life predictions for variable amplitude loading histories and shorter testing times for prototypes and fatigue test articles provided by editing non-damaging cycles from measured loading spectra. The technique involves testing uniaxial fatigue coupons using a fully reversed periodic overload of near-yield magnitude with a fixed period in an otherwise constant amplitude sequence of high stress ratio smaller cycles. The tests can be performed in either load control or strain control with the former offering the advantage of speed and the latter the advantage of greater precision and control of the tests. The net outcome of this procedure is an effective strain life curve for a given material. When used as the basis for fatigue damage calculations, this type of data can give realistic fatigue life estimates for variable amplitude loading conditions. In addition, the endurance limit exhibited by the effective strain life data, which is referred to as the intrinsic fatigue limit, provides a material-based criterion for non-damaging cycles in fatigue. This intrinsic fatigue limit can be of large magnitude in typical engineering alloys and has a practical application as a parameter for filtering non-damaging cycles from rainflow-counted fatigue service load histories. It is shown empirically that the magnitude of the intrinsic fatigue limit can be estimated from the modulus of elasticity of the material.
This paper describes a procedure for measuring the effective strain life curve for a material and demonstrates how to apply the measured data to calculate fatigue life for variable amplitude loading spectra with the aid of computer algorithms. Finally, a technique for editing service load histories to remove non-damaging cycles from service load spectra is described.