• home Home
  • facebook
  • youtube
  • OVERVIEW
    • Reactor Physics & its Vision
    • Research Scope
    • Resources
  • RESEARCH
    • Direct Whole Core Calculation System Development
    • Monte Carlo Code Development
    • Computational Science
    • Multi-Physics Simulation
    • Practical Two-Step Method Development
  • CODES
    • Overview
    • nTRACER
    • RENUS
    • PRAGMA
    • SPHINCS
    • ESCOT
    • EXUS
  • MEMBERS
    • Professor
    • Ph.D. Students
    • MS Students
    • Alumni
  • PUBLICATIONS
    • Journals
    • Conferences
    • Dissertations
  • GALLARY
  • CONTACTS

CODES

  • Overview
  • nTRACER
  • RENUS
  • PRAGMA
  • SPHINCS
  • ESCOT
  • EXUS

nTRACER

Home CODES

Recommended Citation

Jung, Yeon Sang, et al. "Practical numerical reactor employing direct whole core neutron transport and subchannel thermal/hydraulic solvers." Annals of Nuclear Energy 62 (2013): 357-374.

Direct whole Core Transport Code

nTracer is direct whole core transport calculation code. For high-fidelity neutronics calculation, the core calculation paradigm is being shifted from the diffusion theory based two-step core calculation to the direct whole core transport calculation. In contrast to the two-step calculation, the direct whole core calculation requires no prior homogenization of the fuel assemblies or the fuel pins. Instead, the detailed nuclide and temperature distributions inside each fuel pellet are explicitly modeled to represent the actual core condition realistically.

Planar MOC based 3-D CMFD formulation

A 3-D transport calculation without any approximation for a real core problem with sub-pin level heterogeneity is not possible with the present day computing power. Therefore the planar method of characteristics(MOC) based 3-D coarse mesh finite difference(CMFD) formulation in which the axial coupling is resolved by 1D-SP3 nodal solution is employed as the basic transport calculation methodology of nTracer.

Assembly Based Modular Ray Tracing

Assembly based modular ray tracing (AMRT) is essential for practical reactor calculation since most reactors involve either rectangular or hexagonal fuel assemblies. nTracer is equipped with two AMRT modules which can handle the rectangular and hexagonal geometries separately.
A modular ray can be constructed such that a ray passing through the center of the rectangle at a certain position in the core passes another center point at a different core position after traveling n pitches horizontally and m pitches vertically.
In order to improve the efficiency of memory usage, a hierarchical structure in the ray was devised. It consists of core rays, modular rays, cell rays and ray segments. The ray starting from one exterior boundary point and ending at the other boundary point is called a core ray. The modular rays are the part of a core ray dissected by each assembly. The ray inside a pin cell forms a cell ray. The cell ray is further divided into segments by the internal structure of the pin cell. These four levels of ray are constructed efficiently considering various symmetries of the rays.

Axial SP3 Solver

The one-node NEM solution for the SP3 equation was employed in code DeCART. However, it was noted that if the intra-nodal flux shape varies severely, the NEM solver did not represent it properly. Thus a two-node SENM (Source Expansion Nodal Method) solution for the SP3 equation is developed to improve the efficiency over the one-node NEM solution. Also, the CMFD method for the SP3 equation is developed and implemented in the SENM solver.

Decoupled Planar MOC Method

A decoupled planar method of characteristics (MOC) solution based direct three-dimensional (3-D) core calculation scheme is introduced as an effort to enhance the solution stability and accuracy of a hybrid transport solution scheme employing the MOC only to the radial solution. Each planar MOC problem is solved as an eigenvalue problem with the axial leakage converted into the pseudo absorption cross section. The MOC-generated cell homogenized group constants and the current correction factors which are to be used in the subsequent coarse mesh finite difference (CMFD) formulation are not iteratively updated unlike the coupled scheme. The resulting 3-D CMFD problem is thus solved as an isolated problem.

GPU Acceleration

GPU-based massive parallelization was implemented to nTRACER. To achieve maximum performance of GPU, optimized threading scheme was developed and CPU-GPU concurrency and mixed-precision strategy was applied. With consumer-grade GPUs, nTRACER takes about 3 minutes for criticality calculation of 3D commercial PWR core.

GPU based Ray Tracing Threading Scheme

Validation of nTRACER

nTRACER has been applied to the core follow calcullations of two Korean OPR1000 pressurized water reactor cores, YeongGwang Unit 3 and Ulchin Nuclear Unit 5. APR1400 and AP1000 core were analysed with nTRACER, too. nTRACER was also applied to the realistic core benchmark calculations such as BEAVRS and VERA.

Flux Distributions and Pin Power Distribution of BEAVRS Core

Also, nTRACER has transient calculation capability and core depletion calculation capabilities. nTRACER succeeded to perform core depletion calculation of BEAVRS cycle 1 and cycle 2. Currently, transient capability of nTRACER is validated with reacitivity accident experiments in SPERT III E-core.

CBC Curve of BEAVRS Cycle 1 (left) & Cycle 2 (right)

SPERT III E-core test 86 Reactivity Accident Simulation

Rm 32-205(Department of Nuclear Engineering), Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul, Rep. of Korea(08826)  |  TEL: +82-2-880-9241
Copyright © 2020 원자로 물리 연구실