Computational & Technology Resources
an online resource for computational,
engineering & technology publications 

CivilComp Proceedings
ISSN 17593433 CCP: 83
PROCEEDINGS OF THE EIGHTH INTERNATIONAL CONFERENCE ON COMPUTATIONAL STRUCTURES TECHNOLOGY Edited by: B.H.V. Topping, G. Montero and R. Montenegro
Paper 233
An Identification Strategy for Highly Corrupted Measurements in NonLinear Transient Dynamics H.M. Nguyen^{1}, O. Allix^{1} and P. Feissel^{2}
^{1}Laboratoire de Mecanique et Technologie (LMT), ENS Cachan, France
H.M. Nguyen, O. Allix, P. Feissel, "An Identification Strategy for Highly Corrupted Measurements in NonLinear Transient Dynamics", in B.H.V. Topping, G. Montero, R. Montenegro, (Editors), "Proceedings of the Eighth International Conference on Computational Structures Technology", CivilComp Press, Stirlingshire, UK, Paper 233, 2006. doi:10.4203/ccp.83.233
Keywords: identification, inverse problem, corrupted measurements, transient dynamic, viscoplasticity, constitutive relation error, LATIN method.
Summary
The motivation of this study is the identification
of the parameters of a rate dependent damage model [1]
from crash tests
which are characterized by highly scattered measurements.
In this case, contrary to other experimental
situations, there is no a priori information available on
the nature or on the level of the perturbation, which is why we
were a priori unable to rely on a Tikhonov
regularization approach [3,4] or on a Kalman filter
approach [5,6].
Therefore, a specific method has been studied.
It is based on the concept of modified constitutive relation error
(CRE) [7]
which has been proven to be effective in the case of
model updating in vibrations [8].
Its principle consists in splitting
the experimental relations as well as the theoretical ones
into two groups:
the reliable relations and the less reliable relations.
The verification of the reliable ones are enforced throughout the
identification process, whereas the less reliable ones
are relaxed and verified globally, at best,
throughout the identification strategy by the minimization of
the sum of a modeling error term and an experimental error term.
Since the identification of the above damage model appeared to be too complex to be studied at first, the method was first developed in the elastic case. In [2], the method exhibits a very robust "behavior" with respect to the corruption of the measurements. The present work deals with the extension of this method to the case of viscoplasticity. In the case of viscoplastic material, the reliable quantities appear to be the initial conditions, the dynamic equilibrium equation and the state law. The model distance is chosen, among other possibilities, as the quadratic distance to the evolution law. This choice allows the minimization of a quadratic function under nonlinear constraints. The use of the adjoint method leads us to solve a non linear wave propagation problem with both initial and final conditions. The large time increment (LATIN) method [9] is therefore applied in order to linearize this problem over the whole time domain. Thus, the problem is treated by an iterative procedure, where at each iteration we solve successively:
The resolution of the second one is done by using the method developed for the case of elasticity [2]. This identification strategy has been applied in the case of a viscoplastic model with isotropic hardening. The results obtained show that the iterative resolution strategy based on the LATIN method converges with different measurements and different initial guesses of material parameters; and the identification method appears to be robust with respect to perturbations on the measurements up to 40%. The next step will be to apply the methodology to the damage case. References
purchase the fulltext of this paper (price £20)
go to the previous paper 
