Large-Scale Parallel Thermal Elastic-Plastic Welding Simulation Using Balancing Domain Decomposition Method

Yusa, Yasunori; Murakami, Yuma; Okada, Hiroshi

doi:10.1115/pvp2019-93237

Cited by 2 publications

(4 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…First, this section briefly reviews the DD solvers with diagonal-scaling, BDD and BDD-DIAG preconditioners without the inactive elements and their implementation as used in the present study. 3,10,11,39,40 This implementation is almost the same as that of the ADVENTURE System. 8,9 Then, an implementation method for the DD solvers with the three preconditioners for the inactive elements is formulated.…”

Section: Balancing Domain Decomposition Methods With Inactive Elementsmentioning

confidence: 99%

See 1 more Smart Citation

Implementation of balancing domain decomposition method for parallel finite element analysis involving inactive elements

Yusa

Kobayashi

Murakami

et al. 2022

Numerical Meth Engineering

Self Cite

View full text Add to dashboard Cite

The present study developed a numerical method to implement domain decomposition (DD) solvers with the diagonal-scaling preconditioner, the balancing domain decomposition (BDD) preconditioner and the BDD with diagonal scaling (BDD-DIAG) preconditioner, for inactive elements. The inactive element, which is a finite element having zero stiffness, is used in several fields such as multi-pass welding analysis, additive manufacturing analysis, damage analysis, and topology optimization. For this sort of analysis, we adopted the one-time decomposition approach, in which the DD process is performed once at the beginning of the analysis. Based on this approach, we formulated the matrix-vector multiplication, the preconditioning and the vector operations in the algorithm of the conjugate gradient method, along with the inactive elements and floating degrees of freedom caused by the inactive elements.Consideration of the inactive elements is enabled by the slight modifications of matrices and vectors in the algorithm. Numerical examples confirmed the scalability of the BDD and BDD-DIAG preconditioners with the present implementation method. Moreover, the capability of the present method for damage analysis, topology computation, and thermal elastic-plastic analysis of metal additive manufacturing problems was demonstrated. K E Y W O R D Sadditive manufacturing analysis, damage analysis, domain decomposition method, inactive element method, parallel finite element method, topology optimization 3974

show abstract

Section: Balancing Domain Decomposition Methods With Inactive Elementsmentioning

confidence: 99%

“…Next, a thermal elastic-plastic analysis of a metal AM problem is presented. This analysis was performed by the authors' DD-based thermal elastic-plastic welding solver 40 after program modification. The problem was a plate additively manufactured on a substrate, in which only the first layer of the manufacturing process was analyzed.…”

Section: 4mentioning

confidence: 99%

Implementation of balancing domain decomposition method for parallel finite element analysis involving inactive elements

Yusa

Kobayashi

Murakami

et al. 2022

Numerical Meth Engineering

Self Cite

View full text Add to dashboard Cite

show abstract

“…There are several works in literature that demonstrate great advantage in FEM simulation with modern parallel computing systems. [10][11][12] However, with ever growing demand for high fidelity simulation that can be performed in real time, the development of efficient parallel algorithms for FEM simulation remains an active field of research. It has been found that among various expensive steps in FEM, linear solver is often the most time consuming.…”

Section: Introductionmentioning

confidence: 99%

“…The large computational time in FEM is reduced or controlled by employing a large number of compute resources in the form of parallel computing. There are several works in literature that demonstrate great advantage in FEM simulation with modern parallel computing systems 10–12 . However, with ever growing demand for high fidelity simulation that can be performed in real time, the development of efficient parallel algorithms for FEM simulation remains an active field of research.…”

Section: Introductionmentioning

confidence: 99%

Development of GPU‐based matrix‐free strategies for large‐scale elastoplasticity analysis using conjugate gradient solver

Kiran,

Sharma,

Gautam

2024

Numerical Meth Engineering

View full text Add to dashboard Cite

In recent years, matrix‐free conjugate gradient (CG) solvers with graphics processing unit (GPU) acceleration have been effectively used to reduce the execution timings of finite element method (FEM)‐based engineering simulations. However, there is not much in the literature that discusses the application of matrix‐free CG solvers for elastoplasticity. The primary challenge is the presence of both elastic and plastic states, which introduce branching issues in the parallel computation of sparse matrix‐vector multiplication (SpMV). The current work proposes an efficient split kernel strategy for elastoplasticity that segregates elements in elastic and plastic zones and allows the application of standard GPU‐based matrix‐free SpMV strategies in the literature to elastoplastic simulation. The proposed strategy avoids branching issues, facilitates coalesced memory access, allows efficient usage of on‐chip memory, and uses minimum storage to realize large‐scale simulation on a GPU with limited memory. We also propose matrix‐free SpMV strategies for GPU implementation based on an element‐by‐element technique that makes efficient use of memory hierarchy available in a GPU. The computational efficiency of the proposed strategies is demonstrated over three large‐scale benchmark examples from elastoplasticity, and the performance results are compared with GPU optimized node‐based and degrees of freedom (DOF)‐based matrix‐free strategies. The results show that the splitting of computation is most suitable for low plasticity problems, where most of the matrix‐free strategies show significant speedups with respect to single kernel strategy for elastoplasticity. The proposed element‐by‐element strategy using node‐wise thread allocation is found to have the best performance, achieving up to 7.16 and 3.35 speedup over node‐based and DOF‐based strategy, respectively. For problems with higher amounts of plasticity, the proposed strategies perform slightly better and achieve modest speedups due to advantages in the initial stages of Newton iterations. In terms of wall‐clock timings, the proposed matrix‐free solver is found to be up to 2 faster than GPU optimized assembly‐based elastoplasticity solver.

show abstract

Large-Scale Parallel Thermal Elastic-Plastic Welding Simulation Using Balancing Domain Decomposition Method

Cited by 2 publications

References 0 publications

Implementation of balancing domain decomposition method for parallel finite element analysis involving inactive elements

Implementation of balancing domain decomposition method for parallel finite element analysis involving inactive elements

Development of GPU‐based matrix‐free strategies for large‐scale elastoplasticity analysis using conjugate gradient solver

Contact Info

Product

Resources

About