Simulation Technology R&D Office

Research Topics

  • Simulation of Compricated Phenomena
  • Modeling of Embrittlement in Nuclear Materials
  • Fukushima Environmental Research
  • R&D of Quantum Simulation Techniques

| Simulation of Compricated Phenomena

In decommissioning activities at the Fukushima Daiichi Nuclear Power Station and the development of advanced reactors, physical properties of nuclear fuel materials and structure materials are required. However, it is not easy to reproduce extreme situations inside nuclear reactors in laboratories and to measure various properties of nuclear materials there. In such cases, simulations are effective for evaluation of their properties. CCSE performs state-of-the-art numerical simulations to model various factors and to evaluate the properties of nuclear materials.

cement_e.jpg

・This figure shows cesium adsorption in cement to investigate contamination of concrete by radiocesium in reactor containment vessels. We perform machine-learning molecular dynamics with a force field trained to reproduce results of first-principles calculations.

| Modeling of Embrittlement in Nuclear Materials

We use large-scale atomistic simulations such as ab-initio and molecular dynamics calculations to elucidate the mechanism of irradiation embrittlement in nuclear materials, by investigating properties of grain boundaries, dislocations, and irradiation defects in pico-meter and femto-second scales, which are inaccessible by experimental observa tions.

 
Al-GB-e.jpg

A new mechanism of grain boundary fracture, in which hydrogen atoms slip into expanded interstitial space and further expand the space, allowing more hydrogen atoms to slip in (M Yamaguchi, et al. ,Computational Materials Science 156, 368-375).

HCP-disl.jpg

A new mechanism of dislocation cross-slip in which stacking fault meanders and slip-plane changes (M.Itakura et al., Phys.Rev.Lett. 116, 225501 (2016)).

| Fukushima Environmental Research

By using the latest computational techniques, CCSE is evaluating the redistribution of radioactive cesium in Fukushima Prefecture, and its effect on radiation exposure levels.

3d.jpg

3D Air Dose Rate Evaluation System (3D-ADRES) is a system developed by CCSE that uses satellite imagery, terrain data and geographic information systems (GIS) to create 3D models of sites in Fukushima Prefecture. These models are used in state-of-the-art calculations for air radiation dose rates. The rightmost image above shows how the results of 3D-ADRES are able to reproduce the high dose rates measured around the small forest, as well as the low dose rates on paved roads around the buildings. Map imagery © DigiGlobe Inc., NTT Data, Google & Zenrin 2018

| R&D of Quantum Simulation Techniques

In strongly correlated electron systems such as nuclear fuel or superconductors, we have to consider many-body quantum mechanical effect to study physical properties. However, a quantum simulation, simulations for correlated many-body quanta, has huge computational complexity. We develop method to speed up simulations with the use of the machine-learning techniques.

fig-e.jpg

・Speed up of quantum simulations with the use of the machine-learning. The schematic figures of the sampling in the phase space. Each sampling has a huge computational complexity. With the use of the effective model constructed by machine-learning, we can bypass links with huge computational complexity in phase space. The number of the exact quantum simulation is reduced (NAGAI et al. Physical Review B, 96 161102.). Right panel: Phonon density of states for superconductor YNi2B2C, which is obtained by the machine-learning molecular simulations.

Member