Beijing Institute of Mathematical Sciences and Applications

Zhen Li , Xin Liang , Zhi Ting Ma , Hamid Mofidi , Li Wang , Fan Sheng Xiong , Shuo Yang , Wu Yue Yang

Chensong Zhang

Thursday, April 17, 2025 2:00 PM - 3:00 PM

A3-3-201

Zoom 787 662 9899 (BIMSA)

Multiphase multicomponent flow processes are widely encountered in various energy and environmental applications, including oil and gas recovery, geothermal energy utilization, groundwater contamination remediation, carbon dioxide sequestration, and hydrogen storage. These processes play a crucial role in both resource development and ecological protection. To accurately describe such complex phenomena, researchers have developed various mathematical models, such as black oil models, thermal recovery models, fracture models, and dual/multiple porosity models. Among them, general compositional models are emerging as fundamental tools for simulating complex flow behavior, due to their strong capability in representing a wide range of physical and chemical mechanisms. However, the solution of compositional models poses significant numerical challenges, as they typically involve high-dimensional variables, strong coupling, and pronounced nonlinearity. To address these issues, we are developing efficient and scalable simulators for multiphase multicomponent flow, with a focus on fast numerical algorithms and parallel solution techniques. This talk will present our recent progress in the development of such simulators, including the application of CPU-based and GPU-based parallelism as well as neural network techniques, and will discuss the key challenges and future directions in this research area.

张晨松，2002年获南京大学计算数学硕士学位，2007年获得美国马里兰大学应用数学博士学位，此后在宾州州立大学数学系从事博士后研究工作，于2011年进入中国科学院数学与系统科学研究院计算数学所，任副研究员。主要研究兴趣：自适应有限元方法、多层迭代法及其在流固耦合、油藏模拟等问题中的应用。2012年在第二十一届国际区域分解法会议（法国）做大会特邀报告，2013年在第16届全国流体力学数值方法会议做大会特邀报告。曾在Numer Math、SINUM、SIAM MMS、M3AS、JCP、Handbook of Numerical Analysis等国际一流期刊发表了多篇学术论文。