A goal-oriented field measurement filtering technique for the identification of material model parameters
- 168 Downloads
The post-processing of experiments with nonuniform fields is still a challenge: the information is often much richer, but its interpretation for identification purposes is not straightforward. However, this is a very promising field of development because it would pave the way for the robust identification of multiple material parameters using only a small number of experiments. This paper presents a goal-oriented filtering technique in which data are combined into new output fields which are strongly correlated with specific quantities of interest (the material parameters to be identified). Thus, this combination, which is nonuniform in space, constitutes a filter of the experimental outputs, whose relevance is quantified by a quality function based on global variance analysis. Then, this filter is optimized using genetic algorithms.
KeywordsParameter identification Local field measurements Global sensitivity analysis
Unable to display preview. Download preview PDF.
- 2.Herakovich CT (1997) Mechanics of fibrous composites. Wiley, New YorkGoogle Scholar
- 3.Iosipescu N (1967) New accurate procedure for single shear testing of metals. J Mater 2(3): 537–566Google Scholar
- 19.Sobol IM (1969) Multidimensional quadrature formulas and Haar functions. NaukaGoogle Scholar
- 23.Owen A (1995) Monte Carlo and Quasi-Monte Carlo methods in scientific computing. Chapter randomly permuted (t,m,s)-Nets and (t,s)-Sequences. Springer-Verlag, New YorkGoogle Scholar