Optimization of Solid Rocket Grain Geometries

Abstract

Solid Rocket Motors (SRM) are employed in many space launch applications from the booster rockets on the now retired Space Shuttle to the new European launch vehicle Vega. Preliminary design of these complex three dimensional SRM, given certain requirements and limitations can be considered as an optimization process where a best geometrical solution is sought resulting in a desirable thrust profile. In this project, a derivate free direct search package titled NOMAD is employed together with the internally developed numerical burnback analysis tool SRP-GEO and ballistic solver SRP. An analytical model for the burnback analysis tool is also developed to take advantage of the support for surrogate functions within NOMAD. Due to the local nature of the optimizer, the results for complex geometries are shown to converge toward configurations different from the globally optimal geometries. Yet in most instances, the resulting calculated thrust profiles are shown to correlate well with the desired counterpart. This highlights the importance of carefully chosen initial values and boundaries while also emphasize the many possible solutions to a single problem.