RIKEN Center for Computational Science

OVERVIEW 計算科学研究機構とは

Computational Disaster Mitigation and Reduction Research Team

Development of Large-scale Numerical Simulations of Multi-hazard Natural Disasters

We aim to develop and conduct large-scale numerical simulations of multi-hazard natural disasters, including earthquakes, tsunamis, floods, landslides and mud slides.

Simulation of a natural disaster that can devastate an existing city requires the representation of existing buildings and infrastructures as equations and numerical data. These equations and data are defined as numerical city structure objects. Cities have many kinds of structures. They are made up of not only buildings whose design drawings are available, but also underground facilities which have been constructed over years; for example, subway systems, domestic water and sewage pipes, and artificial river banks. We are developing cutting edge technology to represent these complex city components in numerical simulations.

We also aim to create large-scale numerical simulations of geotechnical hazards like liquefaction and slope failure because the underground conditions that produce such hazards are not well understood. We have developed efficient and manageable numerical methods to help us simulate such geotechnical hazards.

By combining these advanced simulations with seismic simulations and meteorological simulations, we aim to produce multi hazard simulations, in other words, we are developing “digital ensemble” for estimating impacts of possible disaster risk. We believe our research results will benefit citizens and governments alike, and will help create resilient societies.

Recent Achievements

Hazard map based on large-scale simulations
We have developed a next-generation hazard map depicting a possible scenario resulting from the predicted Nankai Trough Earthquake. Using large-scale numerical simulations, the map includes four cities in Hyogo prefecture.

Previously, hazard maps were made from the results of simpler numerical simulations and empirical equations based on distribution of ground motion index values. However, the latest simulations using the K computer for next-generation hazard maps have simulated propagation and amplification of seismic waves, and then calculated their physical impact on buildings. In other words, the latest simulations have explicitly calculated the fragility curve. This calculation is an essential component for estimating damage to cities caused by earthquakes because the simulation includes the elemental physical process of the propagation of waves and the response of buildings to these waves. We are applying it to the other disasters like floods, landslides and mud slides; and we will estimate the propagation of the disaster impact into traffic and economy.

Our numerical simulation developed to create next-generation hazard maps has two advanced features. The first is the comprehensive simulation of an entire city 100km2 in area. Comprehensive simulation means both geotechnical wave propagation and urban structures to be simulated simultaneously. This large-scale comprehensive calculation has been performed on the K computer.Consequently, this research achievement in the computational science had been chosen as a finalist for the Gordon Bell Prize. It is much more expected with Fugaku in the near future. The second feature enables the technology to represent existing buildings in numerical simulations as data and equations. By using newly developed data processing platform, hundreds of thousands of buildings can be represented in such simulations.

Numerical city structure model

Research team Leader Satoru Oishi

Team Leader
Satoru Oishi

Biography: Detail
Team Leader, Computational Disaster Mitigation and Reduction Research Team, AICS, RIKEN (-present)
Professor, Kobe University
Associate Professor, University of Yamanashi
Dr. Eng., Kyoto University
Master Eng., Graduate School of Engineering Kyoto University
Annual Report
FY2017 (PDF 1MB)
* FY2012~FY2016
Unit Leader: Muneo Hori

FY2016 (PDF 2.43MB)
FY2015 (PDF 5.65MB)
FY2014 (PDF 2.15MB)
FY2013 (PDF 6.50MB)
FY2012 (PDF 973KB)