Proceedings Article | 11 October 2023
KEYWORDS: James Webb Space Telescope, Integrated modeling, Online learning, Process modeling, Control systems, Wavefront sensors, Thermography, Thermal modeling, Precision optics, Physics
Because of its size and complexity, the Observatory performance of the James Webb Space Telescope (JWST) in its operating condition was primarily verified by analysis prior to launch, breaking the standard NASA paradigm of test-as-you-fly. Most of the requisite analytical predictions were achieved through Integrated Modeling which combines models of the various physics and subsystem behaviors affecting system-level performance: thermal, jitter, structures thermal distortion, gravity release and alignments, optics, wavefront sensing and control, attitude control, straylight and launch loads analysis. We will briefly describe the process by which models were managed, constructed, verified, validated, and used to support the analyses. We will address how uncertainties were accounted for given the many unknowns of the nanometric performance of the first-of-its-kind large deployable cryogenic precision optics in space. Finally, we will compare the pre-flight predictions to actual performance on-orbit for several key analysis use cases and conclude with lessons learned for future missions. The term “jitter” is often synonymous with high-frequency, uncompensated contributions to line-of-sight (LOS) stability or image motion. For the James Webb Space Telescope (JWST) mission, this working definition was extended to include dynamic contributions to wavefront error (WFE). Pre-launch verification of on-orbit jitter performance was not possible via testing, therefore high-fidelity end-to-end modeling was needed for credible verification-by-analysis. The jitter analysis was a main thrust of the overall JWST Integrated Modeling effort and was supported by a robust model verification and validation activities. Following launch, during the latter stages of the 6-month commissioning phase, LOS measurements were obtained and compared to the pre-launch predictions, then to a set of revised predictions. Here we present the modeling methodology supporting the jitter analysis, the pre- and post-launch analysis results, the comparison with the flight data, and conclusions drawn from this comparison.
