Journal article 20 views 6 downloads
Experimental and numerical gust identification using deep learning models
,
Michael Friswell,
Andrea Castrichini
Applied Mathematical Modelling, Volume: 132, Pages: 41 - 56
Swansea University Authors:
Kayal Lakshmanan, Davide Balatti, Hamed Haddad Khodaparast , Michael Friswell
-
PDF | Version of Record
© 2024 The Author(s). This is an open access article under the CC BY license.
Download (1.58MB)
DOI (Published version): 10.1016/j.apm.2024.04.034
Abstract
Identifying gusts and turbulence events is of primary importance for designing future gust load alleviation systems, calculating airframe load, and analysing incidents. Due to the impossibility of their direct measurement, indirect methods are used and ad hoc experiments are necessary to validate th...
| Published in: | Applied Mathematical Modelling |
|---|---|
| ISSN: | 0307-904X |
| Published: |
Elsevier BV
2024
|
| Online Access: |
Check full text
|
| URI: | https://cronfa.swan.ac.uk/Record/cronfa66091 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract: |
Identifying gusts and turbulence events is of primary importance for designing future gust load alleviation systems, calculating airframe load, and analysing incidents. Due to the impossibility of their direct measurement, indirect methods are used and ad hoc experiments are necessary to validate the methodology. This paper employs Convolutional Neural Network and Long Short Term Memory (CNN-LSTM) as well as CNN models for in-flight gust identification. Two aeroelastic models, with different levels of fidelity, representative of a civil and commercial aircraft, are used to generate gust responses to train and test the Deep Learning (DL) models. The results highlight the capability of both LSTM-CNN and CNN models in reconstructing gusts across the entire flight envelope of a civil commercial aircraft. The CNN model demonstrated its ability to identify gusts and turbulence when they occur concurrently, similar to real-world scenarios, in a significantly shorter amount of time. Furthermore, its application to wind tunnel gust response measurements, where the inflow has previously been characterised, demonstrated the effectiveness of the proposed methodology for experimental measurements. |
|---|---|
| Keywords: |
Gust Identification; Inverse Method; Aeroelasticity; Deep Learning |
| College: |
Faculty of Science and Engineering |
| Funders: |
The research leading to these results has received funding from the Engineering Physical Science Research Council (EPSRC) through a program grant EP/R006768/1. The authors also acknowledge the EPSRC Impact acceleration fund. |
| Start Page: |
41 |
| End Page: |
56 |