More about Advanced Modelling and Analysis of Structures

Published on June 8, 2020 Updated on June 8, 2020
Numerical simulation is used for the vast majority of the engineering problems. Its purpose is to "experiment on a model" [1]. It allows us to artificially reproduce, i.e. through the prism of a model, the phenomena we wish to observe. Simulation can be seen as an extension of theory, giving flesh to equations, making its concepts speak, and demonstrating their concrete applications. It is indeed this virtue that makes it very attractive, and sometimes indispensable. However, simulation "only" allows equations to speak, with approximation errors as part of the bargain. Simulation thus deals with what is possible, and not only with what is [1].

It is therefore essential to create some interfaces between the simulation and what we think we know about reality, more precisely to establish a link between the data that can be measured/extracted from a test on a structure with its response simulated by a numerically solved model. Recent advances in the last two decades in experimental methods based on imaging in solid mechanics (correlation of digital images, microtomography, etc.), sometimes coupled with model reduction techniques, or enriched by taking into account uncertainty, have made it necessary to adopt a dual analysis approach combining characterization/observations generating increasingly rich data, with the numerical resolution of models that require validation through experience. Even more so in recent years, with the increase in computing capacity, experimental data can be used to generate databases, to teach a metamodel some elements of the behaviour of a material or to estimate quantities that were previously inaccessible to measurement, which can then be used to conduct numerical simulations. The data, whether it comes from experiments or simulations, partly drives the analysis or other simulations, and is therefore referred to as data-driven. The recent significant development of numerical methods for experimental analysis has made it possible to replace the few known interfaces between theory and experience with a much wider exchange area, where the richness of experimental data can more easily help validate models or partially drive simulations, and where the simulation can be used to reuse, drive the test differently or dynamically enrich a database. The final purpose of this test/calculation dialogue is to save time in the design of new products, to reduce costs, by moving more quickly from the specimen to the structure.

The "Advanced Modeling and Analysis of Structures" specialisation is part of this dual approach to structural analysis in solid mechanics, and aims to train mechanical engineers with both strong skills in modelling and numerical simulation in mechanics, but also trained in new experimental field measurement techniques.

This analytical approach combining simulation and image-based experimental methods for solid mechanics is likely to play a major role in the solutions that science and technology can provide to the challenges of tomorrow (factory of the future, sustainable development, health, energy and mobility). More than preparing the student for a specific profession, the purpose of this specialisation is to instruct the engineer in this dual approach to analysis, which can then be applied to numerous industrial fields (automotive, aeronautics, space, energy, railway, naval, environment).

[1] E. Klein, Comprendre, concevoir, agir : les trois finalités de la simulation. CLEFS CEA -N°47- Hiver 2002-2003.
Published on June 8, 2020 Updated on June 8, 2020