Artificial neural network based prediction of store separation trajectories

Date

2024-9-06

Author

Yaldız, Kübra

Metadata

Show full item record

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

Item Usage Stats

111
views

199
downloads

Cite This

In this study, trajectory estimation for a store is performed quickly and reliably using artificial neural networks. The well-known wing/pylon/store configuration named "EGLIN" in the literature, which has experimental data, is used. Parametric studies are conducted to understand the effects of the number of neurons, the number of hidden layers, and the learning rate parameters in artificial neural networks. Time-dependent computational fluid dynamics (CFD) analyses required for the aerodynamic database are performed using STAR-CCM+, a well-known commercial CFD solver, while the six degrees of freedom mathematical model required for trajectory prediction is created using MATLAB. The artificial neural network models are also developed in the MATLAB environment, allowing the neural network model to be integrated with the mathematical model. The results obtained for flight conditions not included in the neural network training set are found to be quite close to the time-dependent CFD and experimental results. Therefore, it can be said that this approach is a good alternative for the store separation problems.

Subject Keywords

Artificial neural network, Store separation, Unsteady CFD, EGLIN, 6DOF

URI

https://hdl.handle.net/11511/111487

Collections

Graduate School of Natural and Applied Sciences, Thesis