July 24, 2013
This week’s afternoon speech was oriented towards the computational fluid dynamics students in the Academy with Dr. Liou’s talk on engineering design and his work in CFD and optimization. As a senior technologist at NASA Glenn, Dr. Liou has worked in everything from industry to academia to government where he has received countless awards and honors. The talk began as a great respect for Theodore Van Kaman’s work and the students that studied under him was established. Then, Dr. Liou began his talk by defining how one goes about designing and eventually engineering design. This entry portion of the talk transitioned perfectly into an exploration of aircraft design and the methods in which engineers pull from nature for their design inspirations. For example, whale fins contain bumps or seemingly inconsequential anomalies on the leading edge, however, they actually help reduce drag and increase lift.
The talk then shifted towards optimization and CFD generation codes. Highlighting two main methods: Gradient and Genetic Optimization, Dr. Liou explained the need for both tools in the aerodynamic engineer’s toolbox. Gradient optimization refers to the slope recognition and assertion along a path. This method is both fast and has low operating costs. However, gradient optimization has trouble with situations with local minimum and discontinuities. In contrast, genetic optimization uses multi objective evolutionary algorithms to reach an optimal design. Genetic optimization is more robust than gradient in that it doesn’t matter if there are discontinuities or local minimum; however, it is very CPU intensive and takes far longer. The talk closed with many questions from the students, such as what brought Dr. Liou from this undergraduate and graduate work in mechanical engineering to computational work.