Search Results

The overstress probe fatigue testing method, although codified to characterize fatigue strength, is poorly understood. While it yields data confirming whether minimum fatigue strength may be met, it does not directly reveal the mean fatigue strength. Procedures for conducting the test are somewhat arbitrary and rely on fitting a 3-parameter Weibull model. In this paper, a Monte Carlo procedure is developed to simulate the overstress probe test. The effect of various parameters used in the test is also discussed. A comparison is made between Weibull and Gaussian models. Suggestions for conducting the overstress probe test are provided.

The staircase fatigue test method is a well-established, but poorly understood probe for determining fatigue strength mean and standard deviation. The sensitivity of results to underlying distributions was studied using Monte Carlo simulation by repeatedly sampling known distributions of hypothetical fatigue strength data with the staircase test method. In this paper, the effects of the underlying distribution on staircase test results are presented with emphasis on original normal, lognormal, Weibull and bimodal data. The results indicate that the mean fatigue strength determined by the staircase testing protocol is largely unaffected by the underlying distribution, but the standard deviation is not. Suggestions for conducting staircase tests are provided.

Springback is one of the main detrimental factors affecting the dimensional quality of stamped parts in automotive industry. Accurate determination of springback is vital to the design of tools used in the automotive sheet stamping operations. Generally the least distance from a point on a stamped part to the die surface is used as a measurement of the amount of springback. This paper presents a reversed finite element method for assessing the springback of stamped parts. First, the point cloud (coordinate measurement data) scanned from actual stamped parts is triangulated to generate a finite element mesh with triangles. Contrary to the traditional finite element mesh from an existing CAD model, this is a reversed process in which the actual CAD data for the surface is unknown. Then finite element method is utilized to determine the deformation the stamped part needed to conform to the die surface using the minimum amount of work.