Scramjet Fuel Injection Array Optimization Utilizing Mixed Variable Pattern Search With Kriging Surrogates

摘要

Fuel-air mixing analysis of scramjet aircraft is often performed through experimental research or computational fluid dynamics (cfd) algorithms. Design optimization with these approaches is often impossible under a limited budget due to their high cost per run. This investigation uses jetpen, a known inexpensive analysis tool, to build upon a previous case study of scramjet design optimization. Mixed Variable Pattern Search (mvps) is compared to evolutionary algorithms in the optimization of two scramjet designs. The first revisits the previously studied approach and compares the quality of mvps to prior results. The second applies mvps to a new scramjet design in support of the Hypersonic International Flight Research Experimentation (hifire). The results demonstrate the superiority of mvps over evolutionary algorithms and paves the way for design optimization with more expensive approaches.