RESEARCH
Interests: AI/ML, Cloud/HPC/Exascale computing, AI-HPC convergence, High-fidelity (Multiscale/multiphysics) simulations, Fluid & aero- dynamics, Advanced M&S - CFD, FSI, High Order-Scheme, High speed flows (Hypersonic/Supersonic); Applications: Extreme temperature material processing, Porous media flow dynamics; Pore scale network modeling; Nanofluids; CSP-TES/HTF; Viscoelastic flow through deep veins; Airdrop Systems and Parachute dynamics;
Machine Learning:
Proof of concepts: ML Basic Validation |
High Performance Computing (HPC) & AI
AI-HPC Convergence has potential to provide disruptive solutions for many Engineering applications. We are working towards bringing AI enabled technologies at-scale. Our focus is AI embedded in physics based M&S for computationally intensive (exascale class HPC) multiscale and multiphysics problems. Exascale-class HPC with evolving AI, blockchain, IoT and 5G technologies could become realities in advance M&S such as CFD, FEA, and other computationally intensive simulations.
Biomedical Application
CFD Analysis of Deep Vein Thrombosis (DVT) & Pulmonary Embolism (PE)
DVT is a chronic condition with high risk of occurrence among Elderly and pregnant women as well as in many others due to medical or lifestyle related conditions such as Cancer, Diabetes, Obesity, long-haul flight. DVT can cause Pulmonary Embolism (PE) which often results into death. The computational investigation could help doctors and primary care physicians by better understanding the physics involved in the clot formation and its movement through the veins. (Collaborators: Dr. H. Jansen, Texas Tech & Dr. V. Udeowa, USAID)
Numerical investigation of the collagen & cardiomyocytes (with Dr. Joddar, Biomedical Engineering)
Hypersonic & Supersonic flows:
Air-breathing hypersonic vehicles are seen as an affordable access to space. Such vehicles typically have integrated airframe and scramjet propulsions systems such that tip of the vehicle at hypersonic flying conditions create a bow shock that provides compressed air to the scramjet engine. We developed a time-dependent Compressible Fluid Structure Interaction (FSI) computational methodology for solving coupled aero- and structural-dynamics in supersonic and hypersonic regimes. We modeled the scramjet-forebody interactions to understand the unsteadiness in the shock wave triggered by deformation in the structure of the hypersonic air-vehicles.
High order schemes:
We implemented WENO type HO schemes for advection-dominated flow with a theme in climate simulations. Assessing the impact of long-term anthropogenic and natural climate changes and its mitigation at regional level requires significant increase in computational resources and new methods. The research is aimed at developing a novel minimum-norm basis HO scheme by analyzing and developing: high-order single time-step integration for explicit methods and fully implicit methods that can provide coherent nonlinear solution for higher Courant numbers.
