Viscous Effects Investigated by Applied Numerics and Experimental Nanoindentation - VEIANEN
Institution
CRP Henri Tudor
Autres partenaire(s)
Goodyear S.A. ,
Romanian Academy of Sciences (RO) ,
University of Liège
Du : 01/07/2006
Au : 30/06/2008
Budget : 250 000,00€
Contact(s) :
Rauchs Gaston
Summary
In this project, material parameter identification using nanoindentation testing together with the subsequent solution of an inverse problem is investigated. In this method, the difference between experimental measurements and finite element modeling results of the indentation test are minimized using gradient-based numerical optimization. After successful implementation of a sensitivity analysis based on direct differentiation, required for an efficient calculation of the gradients used in the optimization algorithm, the reliability of the method has been assessed through so-called virtual experimental curves, where the experimental input data are replaced by finite element modeling results of the same experiment. In that case, the exact material parameters are known, and the accuracy of the material identification procedure can be assessed. This also gives the possibility to analyze noisy data generated by superposition of a controlled random noise on the virtual experimental curves in order to assess the effect of noise on the material parameters obtained. Using an approximation of the second order gradient, the stability of the identification procedure was monitored in order to define appropriate experimental load histories. It was found that a suitable load history, which includes periods of constant load, can decouple the response of some material parameters. However, a strong coupling between several parameters remains.
The material constitutive laws considered are elasto-viscoplasticity, representative of metals, and viscous hyperelastic material behaviour, representative of rubber. Some results on experimental indentation curves have been obtained for three different indenter tip radii. It was found that residual imprint measurements do not improve material parameter identification because they are prone to large measurement errors. In fact, they are highly depending on the geometry of the indenter tip, which deviates significantly from the nominal shape and size because of the small dimensions involved in nanoindentation. In the case of small indenter tip radii, very good agreement was found between the experimental indentation curves and those modeled with the identified material parameters. For larger tip radii, the results are less reliable, because the indentation depth is limited by the maximum applicable load, and surface roughness significantly influences the indentation curves. The method developed enables the determination of a reasonably good estimate of material parameters. However, a comparison with stress-strain curves recorded from uniaxial tensile test shows significant discrepancies. Whereas the maximum hardening stress and the Young’s modulus can be determined with good accuracy, the initial slopes of yielding are generally of poor quality.
Refereed Scientific Publications:---
Other Publications
- D. Dumitriu, G. Rauchs, V. Chiroiu, Optimisation procedure for parameter identification in inelastic material indentation testing, Revue Roumaine des Sciences Techniques - Série de Mecanique Appliquée, ISSN 0035-4074, Vol. 53, Nº 1, pp. 43-54, Bucarest, 2008.
- G. Rauchs, D. Dumitriu, Parameter Identification in Inelastic Material Indentation Testing, to appear in Research Trends in Mechanics
- G. Rauchs, D. Dumitriu, Optimization Procedure for Material Parameter Identification based on Indentation Testing, submitted to Proceedings of the Romanian Academy, Series A.
- G. Rauchs, Optimization-based Elasto-Viscoplastic Material Parameter Identification from Nanoindentation Testing, to appear in Cahiers de l’Innovation