• Nuclear Engineering and Technology
• /
• 제53권10호
• /
• pp.3264-3274
• /
• 2021

In a severe accident of light water reactor (LWR), molten core material (corium) can be released into the wet cavity, and a fuel-coolant interaction (FCI) can occur. The molten jet with high speed is broken and fragmented into small debris, which may cause a steam explosion or a molten core concrete interaction (MCCI). Since the premixing stage where the jet breakup occurs has a large impact on the severe accident progression, the understanding and evaluation of the jet breakup phenomenon are highly important. Therefore, in this study, the jet breakup simulations were performed using the Smoothed Particle Hydrodynamics (SPH) method which is a particle-based Lagrangian numerical method. For the multi-fluid system, the normalized density approach and improved surface tension model (CSF) were applied to the in-house SPH code (single GPU-based SOPHIA code) to improve the calculation accuracy at the interface of fluids. The jet breakup simulations were conducted in two cases: (1) jet breakup without structures, and (2) jet breakup with structures (control rod guide tubes). The penetration depth of the jet and jet breakup length were compared with those of the reference experiments, and these SPH simulation results are qualitatively and quantitatively consistent with the experiments.

• Nuclear Engineering and Technology
• /
• 제53권11호
• /
• pp.3543-3562
• /
• 2021

The methods and performance of a 3D pin-by-pin neutronics code based on the 2D/1D decoupling method are presented. The code was newly developed as an effort to achieve enhanced accuracy and high calculation performance that are sufficient for the use in practical nuclear design analyses. From the 3D diffusion-based finite difference method (FDM) formulation, decoupled planar formulations are established by treating pre-determined axial leakage as a source term. The decoupled axial problems are formulated with the radial leakage source term. To accelerate the pin-by-pin calculation, the two-level coarse mesh finite difference (CMFD) formulation, which consists of the multigroup node-wise CMFD and the two-group assembly-wise CMFD is implemented. To enhance the accuracy, both the discontinuity factor method and the super-homogenization (SPH) factor method are examined for pin-wise cross-section homogenization. The parallelization is achieved with the OpenMP package. The accuracy and performance of the pin-by-pin calculations are assessed with the VERA and APR1400 benchmark problems. It is demonstrated that pin-by-pin 2D/1D alternating calculations within the two-level 3D CMFD framework yield accurate solutions in about 30 s for the typical commercial core problems, on a parallel platform employing 32 threads.

• 디지털융복합연구
• /
• 제13권12호
• /
• pp.381-390
• /
• 2015

스마트폰과 같은 이동형 장치들은 계산 성능이나 메모리 크기, 배터리 전력량 등의 한계로 인해 엔터테인먼트 애플리케이션이 요구하는 상호작용성을 보장하기 어렵다. 이를 해결하기 위해 본 논문에서는 상호작용이 필수적인 애플리케이션의 응답 속도를 개선할 수 있는 코드 수준 병렬화 방법론을 제안한다. 이 방법을 적용하면, 스마트폰 등에서 제공하는 멀티코어 아키텍쳐를 바탕으로 기존 애플리케이션의 모노코어 알고리즘을 복잡한 재설계 없이 코드 수준에서 병렬화 할 수 있다. 특히 플랫폼 독립적인 표준 쓰레드 라이브러리인 POSIX 쓰레드를 활용하면 안드로이드나 iOS등의 다양한 스마트폰 플랫폼에서 본 방법론을 적용할 수 있다. 이의 효과적인 응용 사례로서 수백만개의 원소를 처리하는 행렬 연산 함수를 병렬화 해보았고 실사용 환경에서 약 3배가량의 성능 향상을 확인하였다.

• 한국전산유체공학회:학술대회논문집
• /
• 한국전산유체공학회 2006년도 PARALLEL CFD 2006
• /
• pp.163-166
• /
• 2006

We have introduced a new version of the 3-D lid-driven cavity problem that leads to more complicated fluid parcel trajectories and thus, enhanced mixing, but at the same time weakens corner singularities. We employed an advanced form of LES to solve this problem and presented preliminary results that show very complicated streamline structures on both large and small scales, despite a relatively low Reynolds number. Finally, we demonstrated moderate speedups via parallelization. Ongoing tests are expected to resolve the questions raised regarding possible sources of the rather poor parallel performance compared with that seen in earlier studies with the same code. Because it is expected that findings may be significant for parallel performance in general, we plan to emphasize this aspect in the oral presentation the Parrel (CFD 2006 Conference.

• Bulletin of the Korean Chemical Society
• /
• 제31권12호
• /
• pp.3639-3643
• /
• 2010

We have developed CUDA-enabled version of a general purpose molecular dynamics simulation code for GPU. Implementation details including parallelization scheme and performance optimization are described. Here we have focused on the non-bonded force calculation because it is most time consuming part in molecular dynamics simulation. Timing results using CUDA-enabled and CPU versions were obtained and compared for a biomolecular system containing 23558 atoms. CUDA-enabled versions were found to be faster than CPU version. This suggests that GPU could be a useful hardware for molecular dynamics simulation.

• 한국수자원학회논문집
• /
• 제43권8호
• /
• pp.747-755
• /
• 2010

지금까지 분포형 모형은 개념적 모형에 비해 운동역학적인 이론에 근거하여 물의 흐름을 수치해석으로 추적해 나가기 때문에 많은 컴퓨터 메모리용량이 요구되고 계산수행시간이 상대적으로 오래 걸리는 단점이 있었다. 그래서 분포형모형의 적용은 주로 소유역에 국한되어 적용되어왔으며, 대유역에 적용하기 위해서는 격자의 해상도를 낮추거나 소유역 적용과 동일한 해상도에서는 계산 시간이 많이 소요되어 실무적용을 어렵게 하는 요인이 되어왔다. 이에 대해 본 연구에서는 MPI (Message Passing Interface)기법을 적용하여 물리적이고 복잡한 수치계산을 하는 분포형모형의 단점 중 하나인 대유역 적용시의 계산소요 문제를 해결하고자 하였다. 또한 병렬화 기법의 적용효과를 분석하기 위하여 단일 영역에서 계산된 결과와 다수의 소영역으로 분할하여 수행되어진 계산 결과를 비교하였다. 용담댐 유역에 2006년 태풍 '에위니아' 사상을 대상으로 병렬화 계산을 수행한 결과 단일 프로세서에서 분포형모형을 수행하는 것보다 병렬코드에서 수행한 경우가 유출계산 값의 변동 없이 최고 10배에 해당되는 계산시간이 단축되었다.

• 한국전산구조공학회논문집
• /
• 제30권2호
• /
• pp.127-135
• /
• 2017

멀티코어 CPU와 BLAS, LAPACK을 구현한 최적 수치라이브러리, 직접 희소 솔버의 대중화 등 PC나 워크스테이션 수준에서도 대규모 유한요소 모델을 해석할 수 있도록 컴퓨팅 환경이 급속도로 변화되었다. 이 논문에서는 멀티코어 CPU를 갖는 공유 메모리 구조에 대한 병렬 유한요소 프로그램 설계시 고려사항으로 (1) 최적화된 수치라이브러리의 사용, (2) 최신 직접 희소 솔버의 사용, (3) OpenMP를 이용한 병렬 요소 강성 행렬의 계산, (4) 희소행렬 저장방식의 일종인 triplet을 이용한 어셈블 기법 등을 제시하였다. 또한 대규모 수치모델을 통해 많은 시간이 소요되는 작업을 기준으로 병렬화 효과를 검토하였다.

• 한국전산유체공학회지
• /
• 제13권1호
• /
• pp.35-40
• /
• 2008

Generally flight vehicles have many cavities such as wheel wells, bomb bays and windows on their external surfaces and the flow around these cavities makes separation, vortex, shock and expansion waves, reattachment and other complex flow phenomenon. The flow around the cavity makes abnormal and three-dimensional noise and vibration even thought the aspect ratio (L/D) is small. The cavity giving large effects to the flow might make large noise, cause structural damage or breakage, harm the aerodynamic performance and stability, or damage the sensitive devices. In this study, numerical analysis was performed for cavity flows by the unsteady compressible three dimensional Reynolds-Averaged Navier-Stokes (RANS) equations with Wilcox's $\kappa-\omega$ turbulence model. The MPI(Message Passing Interface) parallelized code was used for calculations by PC-cluster. The cavity has the aspect ratios of 2.5, 3.5 and 4.5 with the W/D ratio of 2 for three-dimensional cavities. The Sound Pressure Level (SPL) analysis was done with FFT to check the dominant frequency of the cavity flow. The dominant frequencies were analyzed and compared with the results of Rossiter's formula and Ahuja& Mendoza's experimental datum.

• 대한기계학회논문집B
• /
• 제21권8호
• /
• pp.1024-1033
• /
• 1997

The past several years have witnessed an ever-increasing acceptance and adoption of parallel processing, both for high performance scientific computing as well as for more general purpose applications. Furthermore with increasing needs to perform the complex flow calculations in an efficient manner, the use of the message passing model on distributed networks has emerged as an important alternative to the expensive supercomputers. This work attempts to provide a generic framework to enable the parallelization of all CFD-related works using the master-slave model. This framework consists of (1) input geometry, (2) domain decomposition, (3) grid generation, (4) flow computations, (5) flow visualization, and (6) output display as the sequential components, but performs computations for (2) to (5) in parallel on the workstation clustering. The flow computations are parallized by having multiple copies of the flow-code to solve a PDE on different spatial regions on different processors, while their flow data are exchanged across the region boundaries, and the solution is time-stepped. The Parallel Virtual Machine (PVM) is used for distributed communication in this work.

• 대한전기학회논문지:전력기술부문A
• /
• 제48권4호
• /
• pp.388-394
• /
• 1999

This paper is concerned with an application of Parallel Genetic Algorithms(PGA) to optimal econmic load dispatch(ELD) in power systems. The ELD problem is to minimize the total generation fuel cost of power outputs for all generating units while satisfying load balancing constraints. Genetic Algorithms(GA) is a good candidate for effective parallelization because of their inherent principle of evolving in parallel a population of individuals. Each individual of a population evaluates the fitness function without data exchanges between individuals. In application of the parallel processing to GA, it is possible to use Single Instruction stream, Multiple Data stream(SIMD), a kind of parallel system. The architecture of SIMD system need not data communications between processors assigned. The proposed ELD problem with C code is implemented by SIMSCRIPT language for parallel processing which is a powerfrul, free-from and versatile computer simulation programming language. The proposed algorithms has been tested for 38 units system and has been compared with Sequential Quadratic programming(SQP).