With Boeing paying a lot of attention to how it will improve the 737 to take on the Airbus A320NEO, researchers from Auburn University in Alabama have chosen an interesting time to show how naturally-inspired genetic adaptation could be used to help shape it in the future.
The work involves combining computational fluid dynamic solvers with something called genetic algorithms (GA). Originally developed in the 1960s and 1970s, GA’s attempt to take natural evolutionary processes such as adaptation and survival of the fittest, and incorporate them into a computer system. The theory is that evolutionary based algorithms can then be used to optimize a variety of real world systems.
Mimicking some of the process in Darwinian evolution theory, GAs are numerical optimization algorithms built around natural selection and natural genetics. However, unlike evolution, GAs are aimed more at improving solutions to a design problem rather than pure optimization. So, in this way GAs produce a set of guesses of the solution to a particular problem, and can calculate how good or bad these individual solutions are.
In addition, as the level of sophistication and computing power has improved, researchers have used parts of different GAs to form a new, better average solution. Researchers say one of the main benefits of using a GA is that unlike other methods which need an initial “guessed solution” to aim at, the algorithm can start without a single point to get the optimization running.
In addition, by using a whole set of possible solutions (called a population), the GA can use natural-like processes such as selection, crossover and mutation to help direct members of each population toward the desired goals.
However, just as in nature, there are flaws. GA is not guaranteed to find the very best selection. This is because the very same processes of crossover and mutation that generally help drive it in the right direction can also randomly lead to the selection of a non-perfect solution.
Also, apparently GA cannot verify the robustness of individual design solutions – instead it basically attempts to meet the desired goals and will adjust design parameters accordingly.
Auburn University/Guy Norris
Auburn University/Guy Norris
Describing the process at the recent AIAA Joint Propulsion Conference in San Diego, Calif., Auburn’s GA expert Vivek Ahuja showed how the process evolved several unusual new wing shapes for the 737 – including some with inward pointing winglets. Most of the more unusual wingtip responses and exaggerated dihedral were a result of adaptations to the restricted span permitted by the study to suit typical gate sizes.
As the 737 once again becomes the focus for Boeing’s Product Development team, it is interesting to note how the industry’s own process’ of adaptation and survival of the fittest have already pushed the evolution of the twinjet well beyond the wildest dreams of the original designers. Darwin would no doubt have felt suitably vindicated.