NAFEMS Americas and Digital Engineering (DE) teamed up (once again) to present CAASE, the (now Virtual) Conference on Advancing Analysis & Simulation in Engineering, on June 16-18, 2020!

CAASE20 brought together the leading visionaries, developers, and practitioners of CAE-related technologies in an open forum, unlike any other, to share experiences, discuss relevant trends, discover common themes, and explore future issues, including:
-What is the future for engineering analysis and simulation?
-Where will it lead us in the next decade?
-How can designers and engineers realize its full potential?
What are the business, technological, and human enablers that will take past successful developments to new levels in the next ten years?

Resource Abstract

An end-to-end example of verification, validation and uncertainty quantification with particular reference to a mathematical model formulated for the prediction of the fatigue life of structural components in the high cycle regime will be presented. Such models are needed for the formulation of design rules and making sound condition-based maintenance (CBM) decisions.

The objective is to generalize the results of fatigue tests performed on notch-free and notched coupons under constant cycle uniaxial loading conditions to variable cycle triaxial conditions. The mathematical model comprises four sub-models:



(i) A deterministic model of linear elasticity for estimating the elastic stress field,

(ii) a deterministic predictor of fatigue failure defined on the elastic stress field that generalizes the results of experiments performed on notched coupons to arbitrary notches,

(iii) a statistical model for the generalization of fatigue data obtained from notch-free coupons to notched coupons under constant-cycle loading,

(iv) a model for the generalization of a constant-cycle fatigue model to arbitrary load spectra.



Many plausible formulations can be proposed for the sub-models (ii) to (iv). These formulations are based on subjective decisions influenced by intuition, experience and personal preferences. Therefore it is necessary to have a process established for objective ranking of candidate models based on their predictive performance. We have defined and employed such a process based on the principles and procedures of Bayesian statistics.

The choice of a mathematical model from a competing set of models is conditioned on the available experimental data. It is expected that new ideas will be proposed in the future and the available experimental data will increase over time. Therefore ranking and selection of mathematical models is an open-ended problem: It is necessary to establish a process for systematic revision and updating mathematical models. In industrial and research organizations this falls under the administration of simulation governance and simulation project management.

The presentation will introduce a new kind of predictor of fatigue failure characterized by three parameters. The predictor was calibrated and validated on the basis of experimental data available in the public domain. An interesting aspect of the validation project described in this presentation is that whereas the model was calibrated on the basis of uniaxial experimental data, it was validated on the basis of independently obtained biaxial data.

The formulation of any mathematical model must include a clear statement on the limits of admissible input data. Should a new model be proposed that will pass validation tests under less restrictive conditions, and/or have equal or higher measures of predictive performance, then that model should be preferred and the design rules and CBM decisions based on the old model should be revised and updated.

The predictor described in this paper is ranked higher than the classical predictors proposed by Neuber and Peterson as well as the more recently proposed predictors based on the theory of critical distances.