MCRE-based finite element model updating: Cast3M implementation. (November 2022)

• Record Type: Journal Article
• Authors: Oliveira, Hugo Luiz; Louf, François; Gatuingt, Fabrice
• Is Part Of: Advances in engineering software. Volume 173(2022)
• Publication Date: 2022-11
• Subjects: Optimization -- Modified constitutive relation error -- Vibrating regime -- Inverse analysis; Computer-aided engineering -- Periodicals; Engineering -- Computer programs -- Periodicals; Engineering -- Software -- Periodicals; Periodicals; 620.0028553
• DOI: 10.1016/j.advengsoft.2022.103220
• Languages: English
• ISSNs: 0965-9978

Abstract

Highlights: MCRE techniques show good convergence properties in the absence of measurement noise. MCRE modular implementation is important to expand the set of objective functions that can be used. This study presents educational value in both academia and industry. Abstract: In structural engineering, the search for computational models capable of making realistic predictions capturing varied physical phenomena while maintaining simplicity and reliability has gained wide attention of the scientific community. Some outcomes of this effort are the finite element model updating techniques, which consists of using inverse analysis to find realistic values for the set of input parameters. The Modified Constitutive Relation Error (MCRE) has proved to be efficient for model updating in many engineering applications. The central idea of this technique is to formulate parameter identification as a well-posed optimization problem. In general, for each new MCRE application, new formulations need to be deduced and implemented. Despite the specifics of each case, there are standard features that can be abstracted and placed in a single framework. In the present work, we abstract and materialize those features into a new computational framework using Cast3M (an open-source high-level toolbox for finite elements computation). Using typical examples from structural engineering, we test the framework's efficiency and highlight its limitations. The results show that the proposed open-source code allows the identification of physical parameters for finite element models, such as stiffness, as long as physical and geometric linearity conditions remain valid. In addition, elastic inter-element interfaces can be included without significant changes. The benefits of the present study are twofold: (i) source codes are made available and can be used as learning support, in particular for newcomers, and (ii), it can be used as a starting point for new developments in the field of MCRE-based approaches.