Quantitative control of idealized analysis models of thin designs

When preparing a design model for engineering analysis, model idealization is often used, where defeaturing, and/or local dimension reduction of thin regions, are carried out. This simplifies the analysis, but quantitative estimates of the idealization error, the analysis error caused by this idealization, are necessary if the results are to be of practical use. The paper focuses on a posteriori estimation of such idealization error, via both a theoretical analysis and practical algorithms. Our approach can compute bounds for the errors induced by dimension reduction, defeaturing or both in combination. Performance of our error estimate is demonstrated using examples.

► Idealization error: engineering analysis error caused by detail removal or/and local dimension reduction for a complex model. ► We proposed a unified approach for a posteriori estimation of goal-oriented idealization error. ► Approach’s performance is demonstrated using examples.