Regression Modeling and Particle Swarm Optimization for CAMUI Hybrid Rocket Engines

Hybrid rocket engines employing CAMUI fuel grain configurations offer enhanced fuel regression rates and combustion efficiency. However, their design remains heavily reliant on costly and time-consuming experiments. This paper presents a novel reduced-order numerical framework for simulating CAMUI grain evolution based on semi-empirical fuel regression correlations, enabling rapid prediction of internal ballistics and overall engine performance. Validation against LOX/HDPE hot-fire tests shows the capability of the model to reproduce fuel residual mass within 10\% and to capture mixture ratio evolution throughout the burn. An enhanced particle swarm optimization algorithm is coupled with the simulation framework to identify CAMUI geometries that satisfy performance constraints while minimizing unburnt fuel. The proposed approach provides a computationally efficient tool for preliminary design and performance assessment of hybrid rocket propulsion systems.