›› 2016, Vol. 37 ›› Issue (11): 3309-3316.doi: 10.16285/j.rsm.2016.11.033

• Numerical Analysis • Previous Articles     Next Articles

Scalable parallel computation for finite element model with hundreds of millions of elements in geotechnical engineering

ZHANG You-liang, TAN Fei, ZHANG Li-ren, SHI Ming-ming   

  1. State Key Laboratory of Geomechanics and Geotechnical Engineering, Institute of Rock and Soil Mechanics, Chinese Academy of Sciences, Wuhan, Hubei 430071, China
  • Received:2014-12-20 Online:2016-11-11 Published:2018-06-09
  • Supported by:

    This work was supported by the National Basic Research Program of China (973 Program) (2014CB047100), the National Nature Science Foundation of China(11272330)and the 12th Five-year plan of Chinese Academy of Sciences for Informatization (XXH12503-02-02-11).

Abstract: The scalable parallel computation of finite element models with hundreds of millions of elements is discussed. The software and hardware solutions for the preprocessing, parallel computing method, software algorithms, and postprocessing are proposed. A finite element model of one hundred million elements is generated by grid encryption method. And the dual-primal finite element tearing and interconnecting method (FETI-DP) is utilized to solve the finite element system of equations. The topological connections among sub-domains are set up based on graph theories. Point to point communications are built to accelerate data exchange among sub-domains and to avoid the time-consuming global communications. Some modules are written based on an in-house code using object-oriented programming technique and message passing interface (MPI). A numerical example with more than one hundred million number of elements is carried out using 5000 cores at the same time; and super-linear speedup is obtained. The postprocessing is performed on a graphical workstation; and the interactive operations respond quickly. The study shows two big improvements: one is that the numerical model has more than one hundred million numbers of elements; the second is that 5 000 cores are used at the same time and high efficiency is obtained. The research results show that the high resolution numerical simulations provide an efficient tool for large size and complex geological conditions simulations in geotechnical engineering.

Key words: parallel finite elements, scalable parallel computation, geotechnical engineering, dual-primal finite element tearing and interconnecting method (FETI-DP)

CLC Number: 

  • O 242.21

[1] HE Jing-bin, FENG Zhong-ju, DONG Yun-xiu, HU Hai-bo, LIU Chuang, GUO Sui-zhu, ZHANG Cong, WU Min, WANG Zhen, . Dynamic response of pile foundation under pile-soil-fault coupling effect in meizoseismal area [J]. Rock and Soil Mechanics, 2020, 41(7): 2389-2400.
[2] SHI Xu-chao, SUN Yun-de. Analysis of the evolution of excess pore water pressure in soft soil under linear unloading [J]. Rock and Soil Mechanics, 2020, 41(4): 1333-1338.
[3] TIAN Wei, WANG Zhen, ZHANG Li, YU Chen. Mechanical properties of 3D printed rock samples subjected to high temperature treatment [J]. Rock and Soil Mechanics, 2020, 41(3): 961-969.
[4] ZHAO Qiang, JIAO Yu-yong, ZHANG Xiu-li, XIE Bi-ting, WANG Long, HUANG Gang-hai, . Explicit time integration based spherical DDA calculation method [J]. Rock and Soil Mechanics, 2019, 40(11): 4515-4522.
[5] FAN Wen-liang, WANG Yu-le, WEI Qi-ke, YANG Peng-chao, LI Zheng-liang, . Improved fourth-moment method for reliability analysis of geotechnical engineering [J]. , 2018, 39(4): 1463-1468.
[6] HUANG Ming-hua, ZHAO Ming-hua, CHEN Chang-fu. Influence of anchorage length on stress in bolt and its critical value calculation [J]. , 2018, 39(11): 4033-4041.
[7] XIONG Zi-ming, LU Hao, WANG Ming-yang, QIAN Qi-hu, RONG Xiao-li,. Research progress on safety risk management for large scale geotechnical engineering construction in China [J]. , 2018, 39(10): 3703-3716.
[8] FANG Yan-bing, SU Yong-hua, XIAO Wang, LIANG Bin. Non-probabilistic reliability model for implicit performance function based on subinterval method [J]. , 2017, 38(4): 1171-1178.
[9] YAN Shu-wang, LIN Shu, HUO Zhi-liang, CHU Jian, GUO Wei,. Coupled Eulerian-Lagrangian finite element analysis of suction caisson penetration processes under hydraulic pressure [J]. , 2017, 38(1): 247-252.
[10] FU Xiao-dong,SHENG Qian,ZHANG Yong-hui,LENG Xian-lun, . High efficient algorithms for solving linear equations in discontinuous deformation analysis [J]. , 2016, 37(4): 1171-1178.
[11] ZHANG Lei , TANG Xiao-song , LI Dian-qing , CAO Zi-jun , . System reliability analysis of geotechnical structures based on the Copula function [J]. , 2016, 37(1): 193-202.
[12] DONG Wei-xin, WANG Xiang-nan, WANG Yuan, YU Yu-zhen. Application of three-dimensional transitional isoparametric elements to finite element analysis of geotechnical engineering problems [J]. , 2015, 36(5): 1455-1462.
[13] WANG Hao , QIN Wei-min , JIAO Yu-yong , HE Zheng . Transitions and opportunities of geotechnical engineering monitoring in coming big data era [J]. , 2014, 35(9): 2634-2641.
[14] FU Xiao-dong, SHENG Qian, ZHANG Yong-hui. Parallel computing method for discontinuous deformation analysis using OpenMP [J]. , 2014, 35(8): 2401-2407.
[15] SHEN Hui , LUO Xian-qi , ZHENG An-xing , BI Jin-feng , WENG Yong-hong,. Three-dimensional modeling for numerical simulation of arch dam abutment rock mass structures [J]. , 2014, 35(5): 1455-1460.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!