Skip to Main content Skip to Navigation
Conference papers

Automatic Matrix Partitioning Algorithm Underlying Iterative Substructuring Methods for Finite Element Analysis

Abstract : This paper deals with a non-overlapping domain decomposition method involving additional augmented terms along the interface between the sub-domains. The goal of these additional terms is to prevent the singularity of the sub-domain matrix problems, such as rigid body motion in mechanics or resonance frequencies in acoustics. If the definition of these augmented terms requires minor intrusion inside an existing computational code, the sub-domain matrices are however assembled in parallel and the sub-domain meshes are required for this purpose. To cope with these requirements, a small modification of the algorithm is proposed, which helps to partition automatically every global assembly matrix and helps in the development of the method as a 'black box' package. Details of the algorithms, implementation aspects and computational cost are clearly described in order to help the reader to implement this new algorithm. New numerical results performed on interior and exterior acoustics problems illustrate the performance and efficiency of the proposed methods. © 2016 IEEE.
Document type :
Conference papers
Complete list of metadatas

https://hal-centralesupelec.archives-ouvertes.fr/hal-02426673
Contributor : Delphine Le Piolet <>
Submitted on : Thursday, January 2, 2020 - 4:47:32 PM
Last modification on : Thursday, July 2, 2020 - 9:12:02 AM

Identifiers

Citation

Hanyu Zhang, Peter Ivanyi, Frédéric Magoulès. Automatic Matrix Partitioning Algorithm Underlying Iterative Substructuring Methods for Finite Element Analysis. 19th IEEE International Conference on Computational Science and Engineering, 14th IEEE International Conference on Embedded and Ubiquitous Computing and 15th International Symposium on Distributed Computing and Applications to Business, Engineering and Science, CSE-EUC-DCABES 2016, Aug 2016, Paris, France. pp.308-319, ⟨10.1109/CSE-EUC-DCABES.2016.202⟩. ⟨hal-02426673⟩

Share

Metrics

Record views

44