Search

Numerous tsunami numerical models have been proposed, but their prediction accuracies have not been directly compared. For quantifying the modeling uncertainties, the authors statistically analyzed the prediction results submitted by participants in the tsunami blind contest held at the 17th World Conference on Earthquake Engineering. The reproducibility of offshore water level generated due to the tsunami with soliton fission significantly decreased when the nonlinear shallow water equation models (NSWE) was used compared to three-dimensional (3D) models. The inundation depth was reproduced well in 3D models. However, the reproducibility of wave forces acting on the structure and velocities over land was lower in 3D models than that in NSWE models. For cases where the impulsive tsunami wave pressure generated could not be calculated based on the hydrostatic assumption, the prediction accuracy of the NSWE models was higher than that of the 3D models. The prediction accuracies of both models were not improved at small grid-cell sizes. The NSWE model cannot simulate the short-wave component and vertical pressure distribution. Therefore, further developments in 3D models and smoothed particle hydrodynamics methods (SPH) are needed. The presented results contribute to the future development of tsunami numerical simulation tools. [ABSTRACT FROM AUTHOR]

Highlights • Presentation of Alya computational mechanics code in the fusion field. • Including a neutronics module within an HPC multiphysics framework. • First benchmark results of the neutron transport module. Abstract Integrated modelling is an important element of tokamak fusion research. It contributes to the interpretation and planning of experiments, validation of theory against experimental results, development of plasma control techniques as well as the design of next step devices. Large efforts continue to be devoted to the integrated modelling of plasma physics effects in a fusion reactor core and have led to significant progress in physics understanding, while tools to model multiphysics phenomena taking place in fusion reactor components such as tritium breeding blankets have not yet reached a similar level of detail and completeness. We are developing a new high performance computing tool for simulating the latter effects. It is based a parallel computational mechanics code Alya which is specially designed for running with high efficiency standards in large scale supercomputers and is capable of solving different physics problems in a coupled way. As first step towards multiphysics simulations of fusion reactor components, we are developing a new neutron transport module based on the deterministic approach for coupling with the existing, implemented physics modules in Alya. This module, together with existing physics implemented in Alya will allow to solve future high fidelity multiphysics and multiscale problems. We present the capabilities of the Alya system and the latest version of the new neutron transport module currently under development. [ABSTRACT FROM AUTHOR]