Saumil Patel has a background in Computational Fluid Dynamics (CFD). His research interests include CFD, meshing techniques, high-order numerical methods, and high-performance computing. As a graduate student, Saumil worked with scientists in the Mathematics and Computer Science (MCS) division at ANL to develop NekLBM [1, 2] - a high-order CFD application that is derived from Nek5000. After graduating, he joined the Argonne Leadership Computing Facility (ALCF) as a post-doctoral appointee where much of his research was focused on developing numerical algorithms and workflows in Nek5000 to reduce time-to-solution for simulations which employ the Arbitrary Lagrangian-Eulerian (ALE) method [3].  His work has been leveraged by domain scientists in academia and DOE labs to perform internal combustion engine (ICE) simulations.        

Within CPS, Saumil is busy with multiple projects including: performing ICE simulations on Theta, enabling in-situ visualization and analysis for the Exascale Computing Project, and optimizing CFD applications for next-generation computing architectures.  Saumil is also working with data scientists at ALCF to investigate ways CFD applications can leverage machine learning software (like TensorFlow) to reduce time-to-solution and improve existing physical models of the near-wall behavior.  

Saumil received his Ph.D. in mechanical engineering from the Grove School of Engineering at the City College of New York (CCNY).  Before that, he worked as private high school tutor - helping students with their physics and mathematics homework. In a previous life, he was an investment banker in New York City, NY.   

[1] Patel, S., Min, M., Uga, K. C., & Lee, T. (2014). A spectral-element discontinuous Galerkin lattice Boltzmann method for simulating natural convection heat transfer in a horizontal concentric annulus.Computers & Fluids, 95, 197-209.

[2] M. Min and T. Lee, ​“A Spectral Element Discontinuous Galerkin Lattice Boltzmann Method for Nearly Incompressible Flows,” J. Comput. Phys. 230: 245-259 (2011)

[3] Patel, S., Fischer, P., Min, M., & Tomboulides, A. (2019). A characteristic-based spectral element method for moving-domain problems. Journal of Scientific Computing, 79(1), 564-592.