A Parallel Method of Finite Element Domain Decomposition Improved ssorpcg for Solving Temperature Field

Abstract

The invention provides a parallel method for improving SSORPCG to solve temperature field by finite element area decomposition, comprising steps: (1) establishing a solution model, performing finite element discretization on the model, and obtaining the structural information of the model; (2) serially calculating The model is divided into a certain number by the area decomposition method, and the unit information and message transfer index of each partition are allocated; (3) The unit heat conduction matrix in each partition and the contribution matrix of the heat release boundary to the heat conduction matrix and the heat conduction matrix are independently formed in each processor. The element temperature load array is formed to form the finite element governing equation; (4) The improved SSORPCG solver is called to solve the temperature value of each unknown node, and the temperature field of each partition is obtained; (5) The calculation result of the overall model is obtained. Advantages: the invention can effectively improve the ability of solving large-scale sparse coefficient matrix equations of temperature field, and can improve the quality and efficiency of parallel calculation of temperature field by finite element analysis.

Description

technical field [0001] The invention relates to a parallel method for improving SSORPCG to solve temperature field by finite element area decomposition, and belongs to the technical field of finite element engineering numerical simulation calculation and large-scale parallel computing. Background technique [0002] In recent years, computers are developing and progressing at an astonishing speed, and have achieved great progress and remarkable results in terms of computing performance. However, with the improvement of computing scale and computing accuracy, especially in the data-intensive and computing-intensive numerical analysis application fields, the more frequently finite element analysis is used, the performance of ordinary computers can no longer meet the needs of these fields. Efficient parallel algorithms have become one of the indispensable ways of current computing. [0003] In order to solve the differential equation of heat conduction and obtain the temperatur...

AI Technical Summary

Problems solved by technology

Han Lin and others published the article "Application of improved algorithm for symmetric step-by-step over-relaxation preconditioning conjugate gradient method for mass concrete simulation calculation" in the Journal of Hohai University (Natural Science Edition), 2010(03):278-283. Chunyu et al. published an article "The Practice of Improved SSOR-PCG Fast Solution Method in Solving Efficiency and Saving Memory of High Concrete Face Rockfill Dam" in Hydropower Energy Science, 2014(09):90-93. The improved SSORPCG solution method, but these improved methods are not suitable for parallel solution

Images