A Semi-Automated Framework for Launcher Remodelling that Handles Uncertain Data

Kurzfassung

This report is the master thesis A Semi-Automated Framework for Launcher Remodelling that Handles Uncertain Data, conducted at DLR’s Space Launcher Systems Analysis (SART) department within the S3D project. The objective of the thesis was to develop a scalable and robust framework capable of remodelling current launch vehicles (LVs) when key technical parameters are incomplete, inconsistent, or unavailable—an issue frequently encountered when reconstructing launcher configurations for emissions analysis. The framework integrates a structured Input Data Sheet (IDS), analytical rocketry relations, statistical uncertainty-handling methods, and automated trajectory optimisation using SART’s modelling ecosystem. Unknown parameters are treated probabilistically through Monte Carlo sampling, Latin Hypercube sampling, and Approximate Bayesian Computation (ABC). A fast Test Model was implemented to evaluate the performance of these statistical methods, while a high-fidelity Complex Model—combining mass modelling, aerodynamic modelling, and TOSCA-based trajectory optimisation—was used to validate final launcher configurations. Across multiple expendable launch vehicle test cases, Latin Hypercube sampling proved to offer the best balance between numerical convergence, parameter-space coverage, and computational efficiency. When applied to real launchers with up to five uncertain inputs, the framework generated validated configurations whose payload-to-orbit estimates deviated less than 2% from reference performance. The method also quantifies valid parameter ranges and identifies high-sensitivity characteristics, improving reliability when source data are limited. The resulting semi-automated workflow reduces manual effort and improves the scalability of launcher remodelling within SART. It provides a valid aid constructing the 2024 global launch emissions inventory under S3D.