您的位置>> 首页 > 新闻动态 > 学术活动 > 阅读正文
学术活动
地探学院学术报告 (2018047) ——地球物理•地质宫论坛
发表于: 2018-10-08 13:11  点击:

报告题目:3D Numerical Modeling and Large Scale Geophyscial Inversion using Python

人:Lutz Gross

报告时间:20181011日,上午9:00-11:00

报告地点:吉林大学朝阳校区地质宫449

报告人简介:

Lutz Gross, associate professor of school of earth sciences in the University of Queensland. From 2001 to 2003, he works as a computational scientist in the CSIRO Mathematical and Information Sciences Division, Melbourne, Australia.

报告内容:

Numerical modeling and data inversion in geophyics rely on the solution of partial differential equations (PDEs). The esys-escript package has been developed to give researchers an easy-to-use tool to quickly implement and test geophysical models using the programming language python which is now widely used in data analysis. It uses the finite element method (FEM) on structured or unstructured meshes and can be applied to strongly coupled systems of PDEs as well as to loosely coupled, non-linear and time dependent PDEs. A particular strength of esys-escript is that python scripts developed on a desktop computer are highly portable and can run without modifications on parallel supercomputers with thousands of cores using a hybrid mode of threading via OpenMP and distributed computing using MPI.

The talk will give a brief overview on esys-escript and present some application cases. The solution of geophysical inversion problems as a PDE constraint optimization problems using esys-escript will be presented in more details. A particular focus will be how computational scalability across large numbers of cores can be achieved. As an application virtual monitoring of CO2 sequestration will be discussed.

Acknowledgment: This work has been funded by the Australian National Collaborative Research Infrastructure Strategy (NCRIS) and the Carbon Capture and Storage Research Development and Demonstration Fund.



吉林大学地球探测科学与技术学院

                                          2018108


版权所有:吉林大学地球探测科学与技术学院2017 ©  

地址:吉林省长春市西民主大街938号  邮编:130026 电话:0431-88502426  E-mail:xxx@jlu.edu.cn