Event Detail

Event Type: 
Applied Mathematics and Computation Seminar
Date/Time: 
Friday, October 28, 2016 -
12:00 to 13:00
Location: 
GLK 115

Speaker Info

Institution: 
OSU Mechanical Engineering
Abstract: 

Design and development of new combustion technology for higher efficiency, reduced emissions, and fuel flexibility is increasingly driven by computational modeling. Unfortunately, designing the next-generation of clean, efficient combustion systems requires a new level of accuracy unavailable from existing modeling tools. Therefore, we need to develop tools and strategies to enable the use of accurate, predictive models in practical engineering simulations. In this talk I will first discuss the challenges of using accurate, detailed models for chemistry in realistic simulations of reacting flows. In particular, the large sizes of chemical reaction models for the oxidation of liquid transportation fuels present significant computational difficulties. Next, I will describe a multi-stage approach to reducing the size of these large models while retaining a high level of accuracy and predictability, demonstrating its usefulness through case studies on surrogate fuels for gasoline. Finally, I will present our efforts in developing new computational strategies to exploit the massive parallelism of graphics processing units in reactive-flow simulations.