MMAE Seminar - High-Performance Computing Enabling Paradigm Shift in Engine CFD
Armour College of Engineering's Mechanical, Materials & Aerospace Engineering Department will welcome Dr. Sibendu Som, Principal Research Engineer at Argonne National Laboratory, on Wednesday, September 16th to present his lecture, High-Performance Computing Enabling Paradigm Shift in Engine CFD.
Abstract
Vehicles powered by Internal Combustion Engines (ICEs), operating on a variety of fuels, are expected to remain as the primary mode of transportation in the foreseeable future worldwide. Traditionally, ICEs have been designed and optimized mainly based on extensive experimental testing, which is not only expensive, but also time-consuming. With the advent of High Performance Computing (HPC), simulation-based tools are becoming extremely important for providing unique insights into ICE operation. Computational modeling of engines is an arduous task, primarily because the length and time scales of the multiple processes governing their performance and emission characteristics are diverse in nature. In the past decade, engine simulations have been performed with phenomenological models and coarse mesh sizes and hence small clusters suffice to provide reasonable wall-clock times. As the multiphase, turbulence, and chemical kinetic models become more robust and finer spatial and temporal meshes are implemented, the computational power of the small clusters becomes insufficient. The primary motivation of this presentation will be to highlight recent sub-model developments in two-phase flow, combustion, and emission modeling at Argonne National Laboratory. Extensive validation of these models against x-ray radiography data from Argonne and other published optical engine and spray-flame datasets in literature will be mentioned. Our approach towards capturing the influence of injection transients on cycle-to-cycle will be highlighted. We have also been pioneering the use of Uncertainty Analysis tools for ICE simulations and recent applications of this technique towards combustion applications will be demonstrated. The presentation will highlight latest developments in scaling and load-balancing of engine simulations which enabled the largest engine calculation ever, at the Mira supercomputer at Argonne. Applications of the above mentioned tools towards predictive simulations of different low-temperature combustion concepts such as Gasoline Compression Ignition and Reactivity Controlled Compression Ignition will be shown.
Biography
Dr. Sibendu Som is a Principal Mechanical Engineer at Argonne National Laboratory and Computational Fellow at University of Chicago. His research focus is on the development of nozzle-flow, spray, and combustion models for drop-in biofuels, high-performance computing for internal combustion engine applications, and combustion chemistry. In this role, Dr. Som is responsible for developing predictive simulation capabilities to enable OEMs to develop high-efficiency engines for transportation applications. Several of the sub-models developed by Dr. Som’s group are part of commercial CFD engine modeling packages. Dr. Som is a founding member and technical lead on Argonne’s Virtual Engine Research Institute and Fuels Initiative (VERIFI) program. Dr. Som is a recipient of the prestigious High-Performance Computing Innovation Excellence Award by “International Data Corporation”, and winner of the Federal Laboratory Consortium Award for Excellence in Technology transfer by DOE. Dr. Som serves as track-chair for ASME-Internal combustion engine division and session-organizer for The Combustion Institute. He has also been an invited guest speaker at “Indo-US Science and Technology” and “The American Association for the Advancement of Science” forums. Dr. Som serves as guest editor for “Atomization and Sprays” and is on the editorial board for “Journal of Fuels”. Dr. Som has authored more than 100 papers with more than 40 each in journals and peer-reviewed conferences, and two book chapters.