Validation and Verification

Publications

Finite Element Analysis of a Pipe Elbow Weldment Creep-Fracture Problem

P. V. Marcal, J. Fong, R. Rainsberger, and L. Ma

This paper highlights the superiority of the 27-node (hexa-27) element for accurate FEA simulations of complex problems like creep-fracture. It explains that unlike the traditional hexa-20 element, the hexa-27 element's fully quadratic formulation reduces error and provides more precise results.

Uncertainty of FEM Solutions Using a NLLS and a Design of Experiments Approach

This paper introduces and applies a two-pronged approach to quantify uncertainty in Finite Element Method (FEM) simulations, particularly within COMSOL. It uses a nonlinear least squares logistic fit for mesh-induced errors and a Design of Experiments (DOE) approach for model parameter uncertainties. The methodologies are illustrated with examples from structural mechanics and RF coil design, demonstrating their broad applicability and effectiveness in improving simulation accuracy.

AC 2008-2725: DESIGN OF EXPERIMENTS APPROACH

J. T. Fong, J. J. Filliben, N. A. Hecker, R. DeWit

This paper proposes a Design of Experiments (DOE) approach for the verification and uncertainty estimation of simulations based on the Finite Element Method (FEM). It aims to provide a systematic methodology for addressing the reliability of FEM predictions by managing computational resources and analyzing output variability. The approach is illustrated with examples involving software like ABAQUS and ANSYS.

Design Creep Rupture Time vs. Stress Curve

This paper proposes a statistical method to improve finite element analysis (FEA) for creep-fracture damage in pressure vessels and piping. Using a non-linear logistic function, the approach estimates a 95% confidence lower limit for the time-to-failure versus stress curve, accounting for both data and measurement uncertainties.

FEM SOLUTION UNCERTAINTY, ASYMPTOTIC SOLUTION, AND A NEW APPROACH TO ACCURACY ASSESSMENT

This paper presents an approach to verifying Finite Element Method solutions by addressing the various sources of errors and uncertainties in FEM computations. It introduces three mathematical methods and corresponding metrics to assess the accuracy of FEM solutions, allowing for the quantification of uncertainty and the identification of a "best" estimate. The methods and metrics are calibrated with known analytical solutions and applied to problems where theoretical solutions are not available