Estimating temperature and gas concentrations in propellant flames using metamodeling and global optimization

We propose a methodology combining polynomial metamodeling and global optimization for estimating temperature and gas concentrations in propellant flames from moderate-resolution spectroradiometer measurements. We retrieve seven parameters: temperature, soot absorbance, concentrations of CO2 , H2O, CO, and HCl and the projected source area, and we estimate a Cramér-Rao Lower Bound, which provides a reliable estimate of the minimal achievable error. We propose two sampling-based methods for optimization that give parameter estimation close to this bound. Our methodology is validated on representative synthetic cases and one experimental measurement, showing promising results for most parameters. A more complex retrieval method is discussed in the outlook section.