Advances in sediment disaster simulations using clustering techniques

Reference No.	20210015
Type/Category	Grant for Project Research-Short-term Joint Research
Title of Research Project	Advances in sediment disaster simulations using clustering techniques
Principal Investigator	Kazuo Matsuura（Graduate School of Science and Engineering・Associate Professor）
Research Period	August 30,2021. ～ August 31,2021.
Keyword(s) of Research Fields	Disaster Prevention Science, Machine Learning, Fluid Mechanics, Computation
Abstract for Research Report	In Japan, mountains and hills occupy about 70% of the land. Because the land area is small, the slopes of the mountains are steep and easily collapse. Rivers flow rapidly near the water source of the mountain and likely to overflow. In addition, heavy rainfalls associated with climate change augment the risk of sediment disasters. The purpose of this study is to carry out sediment disaster simulations and study methods for controlling sediment disaster damage. Although many simulations of sediment disasters have been reported, objects swept away by sediment disasters were not necessarily realistic, such as assuming simple shapes like spheres and cylinders. We will consider conditions for landslides and their downstream expansion by constructing three-dimensional mountain model data that take into account more realistic soil composition, tree distribution, and water content. In particular, real environmental components such as branch-structured trees, roots, and stones are considered in a unified manner as breakable and decomposable point cloud models with realistic shapes, mass, and strength as detailed as possible. They are incorporated into the flow computation based on Euler, Lagrange, or both standpoints, and the coupled flow simulations based on a two-way coupling of flow and the point cloud models are realized. By realizing such computations, it is expected that disaster occurrence conditions and risk assessments can be refined, and also disaster predictions that are more suited to individual situations can be made.
Organizing Committee Members (Workshop) Participants (Short-term Joint Usage)	Yasuhide Fukumoto（Kyushu University・Professor）