Error assessment for timeline-quantites of interest in structural dynamics
Author
Verdugo, F.; Pares, N.; Diez, P.
Type of activity
Presentation of work at congresses
Name of edition
Congreso de Métodos Numéricos en Ingeniería 2013
Date of publication
2013
Presentation's date
2013-07-25
Book of congress proceedings
Congreso de Métodos Numéricos en Ingeniería, CMN 2013: Bilbao, Spain: July 25-28, 2013: proceedings
First page
1
Last page
1
Abstract
This work presents a new approach to assess the error in specific quantities of interest in the framework of linear elastodynamics. In particular, a new type of quantities of interest (referred as timelinedependent quantities) is proposed. These quantities are scalar timedependent outputs of the transient solution which are better suited to timedependent problems than the standard scalar ones available in the literature [1]. The proposed methodology furnishes error estimates for both the s...
This work presents a new approach to assess the error in specific quantities of interest in the framework of linear elastodynamics. In particular, a new type of quantities of interest (referred as timelinedependent quantities) is proposed. These quantities are scalar timedependent outputs of the transient solution which are better suited to timedependent problems than the standard scalar ones available in the literature [1]. The proposed methodology furnishes error estimates for both the standard scalar and the new timelinedependent quantities of interest. The key ingredient is the modalbased approximation of the associated adjoint problems which allows efficiently computing and storing the adjoint solution. The adjoint solution is readily postprocessed to produce an enhanced solution, requiring only one spatial postprocess for each vibration mode and using the timeharmonic hypothesis to recover the time dependence. The recovery procedure of the vibration modes is very simmilar to the one presented in [2]. The proposed goaloriented error estimate consists in injecting this enhanced adjoint solution into the residual of the direct problem. The resulting estimate is very well suited for transient dynamic simulations because the enhanced adjoint solution is computed before starting the fordward time integration of the direct problem. Thus, the cost of the error estimate at each time step is much reduced