Bayesian Uncertainty Propagation Using Multi-Fidelity Subsystem Models in Design of Precision-Pointed Space Telescopes

The primary outcome of this project includes a fully integrated structural-thermal-optical (STOP) analysis of a space-based precision-pointed telescope design, such as WFIRST. The STOP analysis will provide rigorous uncertainty quantification and sensitivity analysis for the system. This work will address the need for uncertainty propagation throughout the spacecraft modeling process. The current process is to model each subsystem and add an estimated uncertainty downstream of the model’s worst-case outputs. While this is usually sufficient to cover all potential mission scenarios, it also leads to overly conservative designs. Additionally, it is very difficult to pinpoint specific portions of the spacecraft that must be tested more thoroughly during the verification and validation process.

Working to create a Bayesian-based model validation (BMV) system for this STOP analysis will help to not only reduce the time it takes for model validation, but it can also provide engineers with more information about the system earlier in the design process. By providing uncertainty propagation throughout the models, engineers can pinpoint at any stage of the design process which design parameters have the most uncertainty in value. Monte Carlo methods will be used to perform this uncertainty quantification, and techniques such as two-stage Markov Chain Monte Carlo and the control variate framework will be used to reduce the computational cost of propagating uncertainty. This can be further analyzed with a distributional sensitivity analysis, which will identify the model parameters with the most variance. Laboratory tests can then be designed to provide more information about those specific parameters. A multi-fidelity model will be implemented to help reduce the computation time for the distributional sensitivity analysis (DSA), as the DSA can be iterated in a low-fidelity model if that model can estimate the system response with a similar accuracy to the high-fidelity models. With the proposed methodology of uncertainty and sensitivity analysis for the integrated multi-fidelity structural-thermal-optical model, more can be learned about the system, even in early stages of design, and this model validation can help to determine the system response in scenarios that cannot be easily tested on the ground for large, space-based telescopes.

This is early/mid-stage (TRL 2) — the most realistic path in is NASA SBIR/STTR, which funds small businesses and research institutions to develop technology aligned with NASA's needs (equity-free, phased funding). Check whether a current SBIR/STTR solicitation topic overlaps with this project's technology area, or contact the project directly (above) to ask.

None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.