CEAS EuroGNC 2026
|
Impact of Model Fidelity on Load and State Estimation in Hybrid Observers Using Wind Tunnel Data
|
| O. Luderer |
Research Engineer, Institute of Aircraft Systems Engineering, Hamburg University of Technology, Nesspriel 5, 21129 Hamburg, Germany. | | L. Rieck |
Research Engineer, Institute of Aircraft Systems Engineering, Hamburg University of Technology, Nesspriel 5, 21129 Hamburg, Germany. | | B. Herrmann |
Research Engineer, Institute of Aircraft Systems Engineering, Hamburg University of Technology, Nesspriel 5, 21129 Hamburg, Germany. | | F. Thielecke |
Professor, Institute of Aircraft Systems Engineering, Hamburg University of Technology, Nesspriel 5, 21129 Hamburg, Germany. |
|
| Abstract:
The hybrid load observer used for structural load monitoring combines a physical load observer based on a flight-dynamics model with a data-driven correction model. This correction model enables precise load estimation even under physical model uncertainties and neglected physical effects. As such, it is effective in cases where the physical model structure is not immediately apparent and complements purely physics-based modeling approaches, without neglecting the underlying physics. In this sense, the hybrid load observer may serve as a model-based source of essential load information required for modern load alleviation concepts. Besides load estimates, other quantities, such as aeroelastic states, are often relevant for flexible aircraft control. However, these have hardly been considered in previous applications of the hybrid load observer, as it has so far only been applied for load monitoring. Thus, the impact of underlying model uncertainties on crucial states for load control (e.g. aeroelastic states) remains pending. This open aspect is examined in the present work in more detail using wind tunnel data of a flexible wing and by explicitly accounting for physical model uncertainties. To this end, the hybrid load observer, based on an unvalidated design model, is compared with a conventional observer without a correction model that relies on a refined physical model derived from test data. Both physical load observers are based on an LPV extended Kalman filter. The advantages of using a correction model for load estimation are demonstrated in a direct comparison of both observers. Specifically, the hybrid load observer achieves a better load estimation result than the refined LPV extended Kalman filter despite its reduced physical model fidelity. However, the comparison also reveals the shortcomings of the hybrid load observer's current formulation, as model uncertainties affect the quality of the estimated structural dynamic modes and are not yet sufficiently corrected by the correction model.
|
| Keywords: Load Estimation; State Estimation; Hybrid Observer; Linear Parameter-Varying System |
 View PDF CEAS-GNC-2026-057 doi: 10.82124/CEAS-GNC-2026-057 |
| O. Luderer, L. Rieck, B. Herrmann, F. Thielecke: Impact of Model Fidelity on Load and State Estimation in Hybrid Observers Using Wind Tunnel Data. Proceedings of the 2026 CEAS EuroGNC conference. Madrid, Spain. May 2026. doi: 10.82124/CEAS-GNC-2026-057.
|
| BibTeX entry (UTF-8):
@Incollection{CEAS-GNC-2026-057,
author = {Luderer, O. and Rieck, L. and Herrmann, B. and Thielecke, F.},
title = {Impact of Model Fidelity on Load and State Estimation in Hybrid Observers Using Wind Tunnel Data},
booktitle = {Proceedings of the 2026 {CEAS EuroGNC} conference},
address = {Madrid, Spain},
month = may,
year = {2026},
doi = {10.82124/CEAS-GNC-2026-057}
}
|