Investigation of different system identification methods for launch vehicles

Date

2024-1

Author

Demirel Bayrı, Hazel

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

139
views

271
downloads

Cite This

This thesis focuses on identifying the aerodynamic stability and control derivatives of vehicles, specifically launch vehicles. Such vehicles’ dynamics and control exhibit unusual characteristics, including as quick and large changes in inertial and aerodynamic parameters, as well as interactions between low and high-frequency modes. Additionally, accurate vehicle modeling is difficult since, for example, dedicated flight testing with specific maneuvers are disallowed due to stringent safety and financial constraints. The Pegasus launch vehicle's aerodynamic database and first flight test data were used as a case study to evaluate the effectiveness of system identification methods. First, outcomes of a three degree-of-freedom model were used in the least square estimation method to predict aerodynamic derivatives. Furthermore, feed-forward neural network approach was used to predict total aerodynamic force and moment coefficients, and the least square method was used to estimate aerodynamic derivatives from the outputs of the neural model. On the other hand, the output error approach has been developed as a different method to obtain aerodynamic derivative coefficients. Lastly, the predictability of the parameters for these three techniques was compared by using mean-squared errors.

Subject Keywords

Aerodynamic parameters, System identification, Least-squares, Neural network, Output-error, Pegasus launch vehicle

URI

https://hdl.handle.net/11511/108370

Collections

Graduate School of Natural and Applied Sciences, Thesis