With a solid background in Math & Physics, I currently working on the research, development and application of various topics in petroleum reservoir simulation, including Reduced-Order Modeling (ROM) methods, data-driven models, and applications using data mining techniques. I have a strong interest in machine learning to mine large datasets as well as data visualization to present ideas.
I am an INTJ according to Myers-Briggs Type Indicator, one of the rarest personality types, often being described as "the Counselor". I enjoy the process of being helpful to the situation/organization from within.
Honors & Awards
First prize (17th nation-wide) in 20th Chinese Physics Olympiad, Chinese training team for International Physics Olympiad (IPhO) (2003)
Frank G. Miller Fellowship Arard for Outstanding Academic Performance, The Department of Energy Resources Engineering, Stanford University (Autumn 2011-12)
Professional Affiliations and Activities
Memeber, Society of Petroleum Engineers (SPE) (2011 - Present)
Member, Stanford Chinese Association of Petroleum Engineers (2011 - Present)
Member, International Association for Mathematical Geosciences (IAMG) (2012 - Present)
Member / mountaineering team cameraman, Mountaineering Association at Peking University (MAPKU) (2004 - 2008)
Education & Certifications
M. S., Stanford University, Petroleum Engineering (2014)
M. S., Peking University, History of Sciences (2011)
B. S., Peking University, Physics (2008)
Hiking, rock climbing, meditation, scientific, philosophical and religious reflection
Current Research and Scholarly Interests
My research for the PhD period is about the analysis and development of several Reduced-Order Modeling (ROM) methods. POD-TPWL and POD-DEIM are two types of ROM (Reduced-Order Modeling) methods. It enables one to run many simulation scenarios with small perturbations VERY FAST.
My current research is about another type of ROM method, POD-DEIM method.
For previous work on TPWL, one important issue for TPWL is to characterize how far this perturbation can go to ensure the accuracy and applicability of the method. My previous research was to delve into this issue and analyze the theory behind the engineering practice.
This topic is interesting and important, also because it can shed light on the reservoir simulator architecture and fast simulation methods through many aspects. It involves previous work done on efficient nonlinear solver (e.g. Rami Younis, Xiaochen Wang, Boxiao Li), proper orthogonal decomposition, and linear solver. It can also be used in the context of reservoir history matching, production optimization (Marco Cardoso, Jingcong He), etc.
My research during the Master's period is an attempt to use Proper Orthogonal Decomposition (POD) based Reduced-Order Modeling (ROM) method as a linear-solver preconditioner for pressure systems in the reservoir simulator.
The aim of ROM method is to achieve faster performance in reservoir simulation. This is a favorable method particularly for optimization or history matching, where one needs to run a large number of similar models with similar control schemes.
Innovation Group Intern 2013, Quantum Reservoir Impact LLC (6/1/2013 - 9/1/2013)
•Based on the deep understanding of the ECLIPSE model, developed the “Reservoir Simulation Diagnostic Tool” as a part of the “Quantum Analysis Toolbox” featuring fast reservoir diagnostics.
•Contributed to the development of fast reservoir simulation methods.
- Fluid flow effects of compartmentalized distribution of compaction bands in an aeolian sandstone in 3D Petroleum Geoscience 2016
- Scenario Discovery Workflow for Robust Petroleum Reservoir Development under Uncertainty International Journal for Uncertainty Quantification 2016
SCENARIO DISCOVERY WORKFLOW FOR ROBUST PETROLEUM RESERVOIR DEVELOPMENT UNDER UNCERTAINTY
INTERNATIONAL JOURNAL FOR UNCERTAINTY QUANTIFICATION
2016; 6 (6): 533-559
View details for DOI 10.1615/Int.J.UncertaintyQuantification.2016018932
View details for Web of Science ID 000391888200005