This page provides a number of benchmark studies for testing Fatlab and the applied theory. Each study is build around published experimental results and include; a CAD model of the specimen/component, loading and experimental results. In some cases also the FE model (small models only).

SAE keyhole

The following SAE studies are adapted from The specimens were investigated by the Society of Automotive Engineers in the 1970-80’s. The user is encouraged to compare calculations with Fatlab to those provided at the above webpage.

The specimen is modeled in 2D under plane stress assumptions. The loading is applied symmetrically in the 3 holes in each side. The model is constrained in the middle of the top edge. The maximum notch stress at a unit loading of F=1kN is determined to 33.8MPa. The model- and FE stress file is included in the download.

The material is high strength steel: RQC-100, with an ultimate tensile strength of approximately 820MPa. A local stress SN curve has been derived from constant amplitude fatigue tests as shown below. The necessary inputs for Fatlab is also shown in the figure. Both constant- and variable amplitude tests were conducted.

Two checks can be performed for this specimen; one for constant amplitude and one for variable amplitude loading, respectively, using the supplied loads files. In either case, the goal is to obtain a utilization ratio of UR = 1.0. This means, that the estimated life is equal to the experimentally determined life.

Specimen ID Load Life to crack initiation
CR15 (CA) F = ±15.6kN
200,000 cycles
BR2 (VA) SAE bracket, Fmax = 35.6kN
47-113 blocks (of 2968 cycles),
mean: 83 blocks = 246344 cycles

The first study using the CR15 specimen contains a single cycle load. This is difficult for the cycle counters to extract and it is therefore recommended to use the the “Single cycle” option for cycle counting. Try different ones and see how much the results are affected.

The second study (BR2 specimen) is more complicated due to the variable amplitude loading. Here, the user also needs to select a mean stress handling scheme. In lack of specific fatigue data for other mean stresses, the “Modified Goodman” approach could be used, based purely on the tensile strength of the material.

In both cases, utilization ratios very close to 1.0 can be obtained.