Inverse design framework for optimizing solid propellant grains toward target performance profiles

This study presents a computational optimization framework for the inverse design of solid rocket motor (SRM) propellant grains to achieve target thrust-time performance profiles.
Conventional grain design approaches often rely on heuristic rules and iterative trial-and-error procedures, which are time-consuming and may lead to suboptimal solutions.

To overcome these limitations, the proposed framework integrates an artificial neural network (ANN) with a genetic algorithm (GA). A design of experiments (DOE) strategy is first employed to systematically explore the grain design space. Burnback simulations are then performed to capture the temporal evolution of the grain geometry and corresponding thrust characteristics, generating a dataset for ANN training. Once trained, the ANN serves as a fast surrogate model within the GA-based optimization loop, enabling efficient inverse design of grain geometries that satisfy prescribed performance requirements.

The results demonstrate that the proposed approach can accurately recover grain geometries corresponding to target thrust profiles, highlighting its potential as a robust and generalizable inverse design tool for solid rocket motor grain optimization.

Recommended citation: Kim, E., Joo, S., & Oh, S. (2025). "Inverse design framework for optimizing solid propellant grains toward target performance profiles." Acta Astronautica.
Download Paper