Graduate students Yousef Sardahi and Yousef Naranjani were recently awarded summer Bobcat Fellowships from the Mechanical Engineering graduate group. The fellowships will help to support their continuing research over the summer semester.
Yousef Sardahi, left and Yousef Naranjani, right, of the Mechanical Engineering group
Naranjani and Sardahi both work with Professor Jian-Qiao Sun on projects involving multi-objective optimization. In plain terms, this means that they work to design systems that have multiple, usually conflicting objectives.
“An extremely simple example would be that you need to design a very high-quality machine but keep your costs very low — those objectives conflict, so you would need to find the perfect balance of both quality and cost in order to be successful,” Sardahi said.
Sardahi, a third-year Ph.D. student, works primarily on multi-objective optimization for non-linear control systems.
Recently, he has partnered with a former UC Merced visiting scientist, Guomin Song, to optimize a pressure regulator for a common rail system in a diesel engine. Their goal is to make the pressure inside the system stable, but fixed at a certain reference point. Ideally, their successful optimization would lower the overall emissions of the engine and reduce fuel consumption.
Once the testing is complete for the common rail project, Sardahi will move on to working on optimizing a fuzzy control system — a much larger system that has a huge number of parameters to optimize. He expects that the fuzzy control system optimization will continue into the fall semester and beyond.
Naranjani, who is entering the final year of his doctoral studies, is currently working on the multi-objective optimization of subsonic airfoils — in other words, he’s trying to design the most efficient commercial airplane wings possible. To do that, he’ll use a combination of a simple cell mapping method and other evolutionary methods to design a complex algorithm.
This kind of multi-dimensional problem — which in this case has 10 discrete variables that need to be taken into consideration — requires the use of a supercomputer to find the most precise design specifications.
Naranjani is currently learning about the particular coding requirements for the supercomputer, and hopes to be able to run his algorithm within a few months.
