Simulation And Design of Rocket Propulsion Nozzles

Abstract

Nozzle geometry is one of the most important design decisions in rocket propulsion, yet it is also one of the most constrained. Traditional design methods usually determine the nozzle shape based on a single operating condition, accepting potential losses across the rest of the flight envelope as a necessary compromise. This project challenges this approach by treating the nozzle contour as an optimizable variable instead of a fixed one.

A simulation is developed based on one-dimensional isentropic flow theory, with a cosine divergence correction applied to account for off-axis exhaust momentum. The nozzle geometry is parameterized as a discrete radius profile of ten points along a fixed axial length, giving the optimizer direct control over the full contour shape rather than a small number of scalar parameters. Performance is evaluated across a range of ambient pressures to study how the optimal design changes with altitude.

The simulation was tested against 14 real engine configurations, including the NASA RS-25, SpaceX Merlin 1D in both sea-level and vacuum variants, Saturn F-1, SpaceX Raptor Vacuum, and several others. A Nelder-Mead optimizer is leveraged to search the ten-dimensional radius profile space to maximize thrust for each engine at its design condition. Results show consistent thrust improvements across all configurations, with sea-level optimized engines showing larger percentage gains than vacuum engines. The simulation produces important results that are consistent with compressible flow theory and provides a foundation for more in-depth multicondition optimization work.