An automated approach for parallel adjoint-based error estimation and mesh adaptation
- 59 Downloads
In finite element simulations, not all of the data are of equal importance. In fact, the primary purpose of a numerical study is often to accurately assess only one or two engineering output quantities that can be expressed as functionals. Adjoint-based error estimation provides a means to approximate the discretization error in functional quantities and mesh adaptation provides the ability to control this discretization error by locally modifying the finite element mesh. In the past, adjoint-based error estimation has only been accessible to expert practitioners in the field of solid mechanics. In this work, we present an approach to automate the process of adjoint-based error estimation and mesh adaptation on parallel machines. This process is intended to lower the barrier of entry to adjoint-based error estimation and mesh adaptation for solid mechanics practitioners. We demonstrate that this approach is effective for example problems in Poisson’s equation, nonlinear elasticity, and thermomechanical elastoplasticity.
KeywordsAutomated Adjoint A posteriori Error estimation Adaptation Finite element
The authors acknowledge the support of IBM Corporation in the performance of this research. The computing resources of the Center for Computational Innovations at Rensselaer Polytechnic Institute are also acknowledged. The development of tools used in this work was partly supported by the U.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing Research, under award DE-SC00066117 (FASTMath SciDAC Institute). The authors would like to thank Li Dong for providing the discrete geometric model used for the microglial cell example and Max Bloomfield for the geometric model used for the solder ball example.
This work was performed at the time at Rensselaer Polytechnic Institute before the authors moved to different organizations. This paper describes objective technical results and analysis. Any subjective views or opinions that might be expressed in the paper do not necessarily represent the views of the U.S. Department of Energy or the United States Government.
Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.
This work is supported by the U.S. Army grants W911NF1410301 and W911NF16C0117.
- 6.Bangerth W (2017) Deal ii step 14. https://www.dealii.org/current/doxygen/deal.II/step_14.html
- 7.Bavier E, Hoemmen M, Rajamanickam S, Thornquist H (2012) Amesos2 and Belos: direct and iterative solvers for large sparse linear systems. Sci Progr 20(3):241–255Google Scholar
- 13.Diamond G, Smith CW, Shephard MS (2017) Dynamic load balancing of massively parallel unstructured meshes. In: Proc. of the 8th Workshop on Latest Advances in Scalable Algorithms for Large-Scale Systems, Denver, CO, USA. Denver, CO, USAGoogle Scholar
- 15.Eriksson K, Estep D, Hansbo P, Johnshon C (1996) Computational differential equations, 2nd edn. Cambridge University Press, New YorkGoogle Scholar
- 20.Ghorashi SS, Amani J, Bagherzadeh AS, Rabczuk T (2014) Goal-oriented error estimation and mesh adaptivity in three-dimensional elasticity problems. In: WCCM XI-ECCM V-ECFD VI, Barcelona, Spain, Barcelona, SpainGoogle Scholar
- 23.Granzow BN (2017) Goal GitHub repository. https://github.com/bgranzow/goal
- 27.Griewank A, Walther A (2008) Evaluating derivatives: principles and techniques of algorithmic differentiation. Soc Ind Appl Math, Philadelphia, PA, USA, 2 editionGoogle Scholar
- 29.Heroux MA, Willenbring JM (2012) A new overview of the Trilinos project. Sci Program 20(2):83–88Google Scholar
- 34.Li Z, Bloomfield MO, Oberai AA. Simulation of finite-strain inelastic phenomena governed by creep and plasticity. Comput. Mechanics (to be published)Google Scholar
- 36.Nemec M, Aftosmis MJ (2007) Adjoint error estimation and adaptive refinement for embedded-boundary Cartesian meshes. In: 18th AIAA Computational Fluid Dynamics Conf., Miami, FL, USA, Miami, FL, USAGoogle Scholar
- 38.Pawlowski RP, Phipps ET, Salinger AG (2012) Automating embedded analysis capabilities and managing software complexity in multiphysics simulation, Part I: Template-based generic programming. Sci Program 20(2):197–219Google Scholar
- 39.Pawlowski RP, Phipps ET, Salinger AG, Owen SJ, Siefert CM, Staten ML (2012) Automating embedded analysis capabilities and managing software complexity in multiphysics simulation, Part II: application to partial differential equations. Sci Program 20(3):327–345Google Scholar
- 40.Phipps E, Pawlowski R (2012) Efficient expression templates for operator overloading-based automatic differentiation. In: Recent advances in algorithmic differentiation. Springer, Berlin, pp 309–319Google Scholar
- 41.Prokopenko A, Hu JJ, Wiesner TA, Siefert CM, Tuminaro RS (2014) MueLu user’s guide 1.0. Technical Report SAND2014-18874, Sandia Nat. Lab., Albuquerque, NM, USAGoogle Scholar
- 51.Salinger AG, Bartett RA, Chen Q, Gao X, Hansen G, Kalashnikova I, Mota A, Muller RP, Nielsen E, Ostien J et al (2013) Albany: a component-based partial differential equation code built on trilinos. Technical Report SAND2013-8430J, Sandia Nat. Lab., Albuquerque, NM, USAGoogle Scholar
- 53.Smith CW, Granzow B, Ibanez D, Sahni O, Jansen KE, Shephard MS (2016) In-memory integration of existing software components for parallel adaptive unstructured mesh workflows. In: Proc. of the XSEDE16 Conf. on Diversity, Big Data, and Science at Scale, Miami, FL, USA. Miami, FL, USAGoogle Scholar