I grew up in Hunan, China, and always wanted to be a fighter pilot. But since I cannot handle spinning, I settled on being an aerospace engineer. I obtained a B.S. degree from the National University of Defense Technology (NUDT, which produced the fastest supercomputer in the world three years ago) majoring in solid rockets. I discovered CFD during my senior year in college and decided to switch majors. In fact, my senior design project was to simulate supersonic flow over a sphere using a numerical method called artificial pressure developed by my advisor. CFD fascinated me so much that I am still doing it to this day.
I pursued a M.S. degree at NUDT, but did not finish the program. Instead, I decided to go for a Ph.D. degree directly at Glasgow University in Scotland thanks to the Sino-British Friendship Scholarship Scheme. My first journal paper was published in the Journal of Computational Physics on Total Variation Diminishing schemes. After obtaining my Ph.D., I did postdoc research in Glasgow and Oxford before joining CFD Research Corporation in Huntsville, Alabama. The opportunity to teach attracted me to academia, first Michigan State University, then Iowa State University, and presently the University of Kansas.
I have been a passionate advocate for high-order CFD methods. I strongly believe in their potential in revolutionizing industrial large eddy simulation. I’ve edited a book, published several review articles, and co-organized several International Workshops on High-Order CFD Methods. I also served as an Associate Editor for the AIAA Journal from 2006 to 2016. In 2008 I was awarded the degree of Doctor of Science in Engineering by Glasgow University, and more recently became an AIAA Fellow in 2015.
- Location: Lawrence, Kansas
- Current position: Spahr Professor of Aerospace Engineering, University of Kansas
- Current computer: All three types, Windows (office), Linux (for CFD simulations) and MacPro (travel)
- One word that best describes how you work: Focused
How do you know Pointwise?
I have known Pointwise for many years, but cannot recall when I came across Pointwise for the first time. I believe I first heard of Gridgen more than two decades ago. Later I met your President, John Chawner, at numerous AIAA conferences and workshops. I have particularly enjoyed reading his all-encompassing CFD blog “Another Fine Mesh.” I also got to know several Pointwise engineers, e.g., Steve Karman and Nick Wyman, because of our mutual interest in high-order mesh generation. Steve generated many high-order meshes for the 5th International Workshop on High-Order CFD Methods, and the AIAA High Lift Workshop. We have interacted a lot to test and debug some of those meshes.
What do you see are the biggest challenges facing CFD in the next 5 years?
This is a very tough question. You may get ten different answers if you ask ten people because the answer depends so much on each person’s perspective. If we take a look at the big picture, there are significant challenges in physical modeling, numerical algorithms, and pre- and post-processing. These challenges can be illustrated by a grand challenge problem, flow through a gas-turbine engine. Physical modeling challenges include turbulence models, combustion models, and fuel injection models et al. Numerical algorithm challenges comprise highly accurate and efficient solution methods capable of handling dynamic moving boundaries and taking advantage of extreme-scale supercomputers. Pre- and post-processing challenges consist of mesh generation including high-order mesh generation, adaptation and flow visualization of huge and dynamic data sets.
What are you currently working on?
Ever since I came to academia nearly 18 years ago, the main theme of our research has been adaptive high-order methods for large eddy simulations (LES) of vortex-dominated turbulent flow. It is mind-boggling that it took a decade and a half to mature our high-order CFD tool, hpMusic (hp-adaptive MUlti-physics SImulation Code), which is capable of handling industrial LES on unstructured meshes. Along the way, we developed/experimented with many numerical methods, such as spectral volume, spectral difference, discontinuous Galerkin, and the flux reconstruction/correction procedure via reconstruction (FR/CPR) methods. In addition, we had to overcome many challenges in shock-capturing, time integration, high-order mesh generation, and parallel implementations.
We are currently working on scalable time integration algorithms for extreme-scale computers, dynamic overset and sliding mesh approaches for helicopter and rotor/stator problems, hp-adaptations for LES, and computational aeroacoustics including jet noise predictions.
I recently embarked upon a new challenge to start a company, hoCFD LLC, to market our high-order CFD tool for LES in fluid dynamics, aerothermal, and aeroacoustic applications. A collaboration with GE has provided me an opportunity to understand the challenges our industry partner faces. We reacted to meet the needs of GE, which has resulted in a more robust and relevant commercial product, hoMusic. I am excited that GE became hoCFD’s first customer.
What project are you most proud of and why?
It is quite tough to single out a project we are the most proud of since we had some exciting ones. If I had to pick one, it would be the project from GE Aviation, working with GE Global Research on the Evaluation of Adaptive High-Order Methods for Industrial LES. In this project, we were given several industry-relevant benchmark LES problems to test hpMusic. The collaboration with GE identified some challenges that we needed to overcome in hpMusic in order to meet the needs of our industry partner. Fortunately, this did not deter my students. Instead, they were super-motivated that GE was willing to take a look at our academic solver. Within a short period of time, hpMusic progressed from a research code to a semi-production-level tool capable of solving quite challenging problems on supercomputers with over 10,000 cores. We published the latest comparison with a leading commercial CFD tool using a turbo-machinery benchmark problem on the same family of meshes. HpMusic delivered much higher accuracy with much less cost.
I am very proud of my students for this achievement. As a professor, I am obviously very happy to publish research papers with my students, and watch them grow in ability and confidence. It is an extra-bonus that our tool can perhaps be used in designing real products, which adds more meaning to what we do.
Are you reading any interesting technical papers we should know about?
I am constantly browsing papers with new ideas, algorithms, or new applications. Recently, I have been reading papers on mesh-adaptations and wall-modeling for LES as I believe they can dramatically reduce simulation costs. I am also reading papers on memory-efficient implicit algorithms for high-order spatial discretizations to hopefully develop them into an efficient and scalable tool.
What software or tools do you use every day?
We use Qt Creator as our development environment, Git for version control, Matlab or Mathematica for symbolic derivations, Tecplot 360 and ParaView for visualization, and Gmsh and Pointwise for mesh generation. I use the Microsoft Office suite for emails, papers and presentations, and skype for web conferences.
What does your workspace look like?
I have a Windows workstation, mostly for emails and office-related functions. I have a dual monitor Linux workstation for software development and visualization, as shown below.
Do you plan on attending any conferences or workshops this year?
Yes, I plan to attend at least three conferences. Dr. HT Huynh (NASA Glenn), Dr. Chongam Kim (Seoul National) and I are organizing a mini-symposium entitled Recent Progress and Trends in High-Order CFD Methods, at this year’s ECCM-ECFD 2018 conference to be held in Glasgow on June 11-15, 2018. In addition, I plan to attend the 2nd Chinese International Conference on CFD, to take place in Mianyang, Sichuan, on 17-21, 2018. Finally, I will give a talk in the International Council of Aeronautical Sciences conference (ICAS 2018), to be held in the city of Belo Horizonte, Minas Gerais, Brazil, on September 09-14, 2018.
What do you do outside the world of CFD?
I enjoy listening to classical/light music a lot, and reading autobiographies of successful people to pick their brains. I play tennis, volleyball, badminton and table tennis regularly to de-stress and recently played a fierce tennis match against a former student in Orlando. When I can, I watch KU basketball, as well as Iowa State and Michigan State basketball. There was a week last year when these three schools were ranked in the top four of the AP NCAA Basketball poll. I deserve to brag about my skills in picking basketball schools to work for.
What is some of the best CFD advice you’ve ever received?
“Build physical intuition into your CFD algorithms.” When a CFD algorithm violates basic physical principles, it will produce nonsense. For example, a downwind scheme for the wave equation violates the physics of wave propagation. The mathematical consequence is that the scheme is unstable, and sure enough produces nonsense.
If you had to pick a place to have dinner, where would you go?
There are quite a few nice restaurants and bars in Lawrence. Free State Brewing Co. is the most popular bar in Lawrence, and the food is also very good. KU Aerospace faculty often goes to 23rd St. Brewery for various gatherings and celebrations. Oriental Bistro & Grill and Red Pepper are my two favorite Chinese restaurants.