000013256 001__ 13256
000013256 005__ 20161114160333.0
000013256 04107 $$aeng
000013256 046__ $$k2009-06-22
000013256 100__ $$aWaeytens, J.
000013256 24500 $$aModel verification in dynamics through strict upper error bounds

000013256 24630 $$n2.$$pComputational Methods in Structural Dynamics and Earhquake Engineering
000013256 260__ $$bNational Technical University of Athens, 2009
000013256 506__ $$arestricted
000013256 520__ $$2eng$$aThis paper deals with verification in dynamics. To provide information on the reliability of the approximate solution, which is here obtained with a finite element discretization in space and a Newmark scheme integration in time, a global error estimator based on the constitutive relation error is used. Thus, reconstruction of admissible fields verifying the kinematic equations and the dynamic equilibrium is needed. For the linear acceleration Newmark scheme, the CFL parameter is chosen such as it minimizes the error estimator. Then, a method introducing an adjoint problem is employed to compute guaranteed error bounds on a given quantity of interest, such as the mean stress over a prescribed space-time domain. First illustrations are shown on a 2D academic problem in dynamics along with a linear viscoelastic Maxwell model.

000013256 540__ $$aText je chráněný podle autorského zákona č. 121/2000 Sb.
000013256 653__ $$aError Estimation, Output of interest, Upper bound, Dynamics. Abstract. This paper deals with verification in dynamics. To provide information on the reliability of the approximate solution, which is here obtained with a finite element discretization in space and a Newmark scheme integration in time, a global error estimator based on the constitutive relation error is used. Thus, reconstruction of admissible fields verifying the kinematic equations and the dynamic equilibrium is needed. For the linear acceleration Newmark scheme, the CFL parameter is chosen such as it minimizes the error estimator. Then, a method introducing an adjoint problem is employed to compute guaranteed error bounds on a given quantity of interest, such as the mean stress over a prescribed space-time domain. First illustrations are shown on a 2D academic problem in dynamics along with a linear viscoelastic Maxwell model.

000013256 7112_ $$aCOMPDYN 2009 - 2nd International Thematic Conference$$cIsland of Rhodes (GR)$$d2009-06-22 / 2009-06-24$$gCOMPDYN2009
000013256 720__ $$aWaeytens, J.$$iChamoin, L.$$iLadeveze, P.
000013256 8560_ $$ffischerc@itam.cas.cz
000013256 8564_ $$s164060$$uhttps://invenio.itam.cas.cz/record/13256/files/CD382.pdf$$yOriginal version of the author's contribution as presented on CD, section: Inverse problems and system identification.
000013256 962__ $$r13074
000013256 980__ $$aPAPER