Parallel Strategy of FMBEM for 3D Elastostatics and its GPU Implementation Using CUDA

Author(s):  
Zhaohui Xia ◽  
Qifu Wang ◽  
Yunbao Huang ◽  
Wei Yixiong ◽  
Wang Yingjun

Finite Element Method (FEM1) is pervasively used in most of 3D product design analysis, in which Computer Aided Design (CAD) models need to be converted in to mesh models first and then enriched with some material features and boundary conditions data, etc. The interaction between CAD models and FEM models is intensive. Boundary Element Method (BEM) has been expected to be advantageous in large-scale problems in recent years owing to its reduction of the dimensionality and its reduced complexity in mesh generation. However, the BEM application has so far been limited to relatively small problems due to the memory and computational complexity for matrix buildup are O(N2). The fast multipole BEM (FMBEM) combined with BEM and fast multipole method (FMM) can overcome the defect of the traditional BEM, and provides an effective method to solve the large-scale problem. Combining GPU parallel computing with FMBEM can further improve its efficiency significantly. Based on the three-dimensional elastic mechanics problems, the parallelisms of the multipole moment (ME), multipole moment to multipole moment (M2M) translation, multipole moment to local expansion (M2L) translation, local expansion to local expansion (L2L) translation and near-field direct calculation were analyzed respectively according to the characteristics of the FMM, and the parallel strategies under CUDA were presented in this paper. Three main major parts are included herein: (1) FMBEM theory in 3D elastostatics, (2) the parallel FMBEM algorithm using CUDA, and (3) comparison the GPU parallel FMBEM with BEM, FEM and FMBEM respectively by engineering examples. Numerical example results show the 3D elastostatics GPU FMBEM not only can speed up the boundary element calculation process, but also save memory which can be effective to solve the large-scale engineering problems.

Author(s):  
Yijun Liu ◽  
Milind Bapat

Some recent development of the fast multipole boundary element method (BEM) for modeling acoustic wave problems in both 2-D and 3-D domains are presented in this paper. First, the fast multipole BEM formulation for 2-D acoustic wave problems based on a dual boundary integral equation (BIE) formulation is presented. Second, some improvements on the adaptive fast multipole BEM for 3-D acoustic wave problems based on the earlier work are introduced. The improvements include adaptive tree structures, error estimates for determining the numbers of expansion terms, refined interaction lists, and others in the fast multipole BEM. Examples involving 2-D and 3-D radiation and scattering problems solved by the developed 2-D and 3-D fast multipole BEM codes, respectively, will be presented. The accuracy and efficiency of the fast multipole BEM results clearly demonstrate the potentials of the fast multipole BEM for solving large-scale acoustic wave problems that are of practical significance.


Author(s):  
Yingjun Wang ◽  
Qifu Wang ◽  
Gang Wang ◽  
Yunbao Huang ◽  
Yixiong Wei

Finite Element Method (FEM) is pervasively used in most of 3D elastostatic numerical simulations, in which Computer Aided Design (CAD) models need to be converted into mesh models first and then enriched with semantic data (e.g. material parameters, boundary conditions). The interaction between CAD models and FEM models stated above is very intensive. Boundary Element Method (BEM) has been used gradually instead of FEM in recent years because of its advantage in meshing. BEM can reduce the dimensionality of the problem by one so that the complexity in mesh generation can be decreased greatly. In this paper, we present a Boundary Element parallel computation method for 3D elastostatics. The parallel computation runs on Graphics Processing Unit (GPU) using Computing Unified Device Architecture (CUDA). Three major components are included in such method: (1) BEM theory in 3D elastostatics and the boundary element coefficient integral methods, (2) the parallel BEM algorithm using CUDA, and (3) comparison the parallel BEM using CUDA with conventional BEM and FEM respectively by examples. The dimension reduction characteristics of BEM can dispose the 3D elastostatic problem by 2D meshes, therefore we develop a new faceting function to make the ACIS facet meshes suitable for Boundary Element Analysis (BEA). The examples show that the GPU parallel algorithm in this paper can accelerate BEM computation about 40 times.


2010 ◽  
Vol 20-23 ◽  
pp. 76-81 ◽  
Author(s):  
Hai Lian Gui ◽  
Qing Xue Huang

Based on fast multipole boundary element method (FM-BEM) and mixed variational inequality, a new numerical method named mixed fast multipole boundary element method (MFM-BEM) was presented in this paper for solving three-dimensional elastic-plastic contact problems. Mixed boundary integral equation (MBIE) was the foundation of MFM-BEM and obtained by mixed variational inequality. In order to adapt the requirement of fast multipole method (FMM), Taylor series expansion was used in discrete MBIE. In MFM-BEM the calculation time was significant decreased, the calculation accuracy and continuity was also improved. These merits of MFM-BEM were demonstrated in numerical examples. MFM-BEM has broad application prospects and will take an important role in solving large-scale engineering problems.


2010 ◽  
Vol 439-440 ◽  
pp. 80-85
Author(s):  
Hai Lian Gui ◽  
Qing Xue Huang ◽  
Ya Qin Tian ◽  
Zhi Bing Chu

Based on fast multipole boundary element method (FM-BEM) and mixed variational inequality, a new method named mixed fast multipole boundary element method (MFM-BEM) was presented in this paper. In order to improve calculation time and accuracy, incompatible elements as interpolation functions were used in the algorithm. Elements were optimized by mixed incompatible elements and compatible elements. On the one hand, the difficult to satisfy precise coordinate was avoided which caused by compatible elements; on the other hand, the merits of MFM-BEM were retained. Through analysis of example, it was conclusion that calculation time and accuracy were improved by MFM-BEM, calculation continuity was also better than traditional FM-BEM. With increasing of degree of freedom, calculation time of MFM-BEM grew slower than the time of traditional FM-BEM. So MFM-BEM provided a theoretical basis for solving large-scale engineering problems.


2005 ◽  
Vol 13 (01) ◽  
pp. 71-85 ◽  
Author(s):  
Y. YASUDA ◽  
T. SAKUMA

The fast multipole boundary element method (FMBEM) is an advanced BEM, with which both the operation count and the memory requirements are O(Na log b N) for large-scale problems, where N is the degree of freedom (DOF), a ≥ 1 and b ≥ 0. In this paper, an efficient technique for analyses of plane-symmetric sound fields in the acoustic FMBEM is proposed. Half-space sound fields where an infinite rigid plane exists are typical cases of these fields. When one plane of symmetry is assumed, the number of elements and cells required for the FMBEM with this technique are half of those for the FMBEM used in a naive manner. In consequence, this technique reduces both the computational complexity and the memory requirements for the FMBEM almost by half. The technique is validated with respect to accuracy and efficiency through numerical study.


2013 ◽  
Vol 813 ◽  
pp. 387-390
Author(s):  
Hai Lian Gui ◽  
Qiang Li ◽  
Yu Gui Li ◽  
Xia Yang ◽  
Qing Xue Huang

In this paper, a new fast multipole boundary element method is presented. By using Taylor series expansion and a new mapping in boundary cell, the efficiency of calculation about influence coefficients has been improved. Compare with the old fast multipole boundary element method, this new method is easier to be suitable for the large-scale numerical calculus request.


Author(s):  
Chunxiao Yu ◽  
Cuihuan Ren ◽  
Xueting Bai

To solve large scale linear equations involved in the Fast Multipole Boundary Element Method (FM-BEM) efficiently, an iterative method named the generalized minimal residual method (GMRES)(m)algorithm with Variable Restart Parameter (VRP-GMRES(m) algorithm) is proposed. By properly changing a variable restart parameter for the GMRES(m) algorithm, the iteration stagnation problem resulting from improper selection of the parameter is resolved efficiently. Based on the framework of the VRP-GMRES(m) algorithm and the relevant properties of generalized inverse matrix, the projection of the error vector rm+1 on rm is deduced. The result proves that the proposed algorithm is not only rapidly convergent but also highly accurate. Numerical experiments further show that the new algorithm can significantly improve the computational efficiency and accuracy. Its superiorities will be much more remarkable when it is used to solve larger scale problems. Therefore, it has extensive prospects in the FM-BEM field and other scientific and engineering computing.


Sign in / Sign up

Export Citation Format

Share Document