Accuracy of Predicting Stress Concentration Factors

This presentation was made at CAASE18, The Conference on Advancing Analysis & Simulation in Engineering. CAASE18 brought together the leading visionaries, developers, and practitioners of CAE-related technologies in an open forum, to share experiences, discuss relevant trends, discover common themes, and explore future issues.

Resource Abstract

Master Degree Students in Civil and Mechanical Engineering were given specific geometries to model in a finite element code to capture theoretical stress concentration factors. Their results were reviewed after some level of support from the instructor. These results may guide managers and practitioners in the accuracy of finite element results in a typical stress analysis study.

The author takes the unixial, isothermal results and presents a subset of data for composite materials, biaxiality of stress, and thermal gradients. The expectations should be lowered accordingly as the data suggests. The author speculates in his conclusions regarding level of difficulty, user experience and what a manager should expect.

The anlayses presented include the Kt of a hole where the theoretical Kt approaches 2.9 for a finite width plate. This represents the simplest of analyses in that it is effectively two-dimensional, isothermal, uniaxial, and isotropic. Next the Kt of an embedded spheroid where the Kt approaches 2.02 had the same limitations except that it is a three-dimension stress concentration factor.

The analytical verification and validation protocol is reviewed. The author presents souces of scatter observed from decades of analyzing and testing isotropic materials. Sources of scatter are expanded to non isotropic materials and should include orthotropy, loading variation, loading alignment, grain boundaries, variation in elastic constants, variation in thermal coefficients of expansion, variation in density, and the like.

The type of analysis plays a role in accuracy as well. Buckling prediction presents a challenging example of analytical prediction vs test data.

The testing sources of error are listed and can be just as malevolent as finite elment predictions.

The author can only present anecdotal data for the more difficult problems found in the real world. Sources of scatter or error are listed and should provoke discussion. An industry can adopt coefficient of variation assumptions in the analytical design cycle to ameliorate structural integrity.