Title: Aeroengine state prediction based on generative adversarial networks and deep learning
Authors: Qiang Fu; Huawei Wang
Addresses: College of Civil Aviation, Nanjing University of Aeronautics and Astronautics, Nanjing, 211106, Jiangsu, China ' College of Civil Aviation, Nanjing University of Aeronautics and Astronautics, Nanjing, 211106, Jiangsu, China
Abstract: The artificial intelligence technology represented by deep learning provides the possibility to make a comprehensive characterisation of the aeroengine state from state monitoring information. The premise of these algorithms is based on big data. First, this paper applies the generative adversarial networks to generate aeroengine condition monitoring data to expand data volume. Experimental results confirm that the generated data can reflect the regularity of the original monitoring data after a large number of network training iterations. Second, the deep learning algorithm is employed to predict the aeroengine status of the monitoring data and its generated data. Prediction accuracy is compared with the traditional neural network prediction method, which demonstrates the effectiveness of the deep learning prediction and the combination of the generative adversarial networks and deep learning. This aspect can solve the problem of limited data volume.
Keywords: aeroengine; deep learning; generative adversarial networks; state prediction.
DOI: 10.1504/IJCAET.2021.118468
International Journal of Computer Aided Engineering and Technology, 2021 Vol.15 No.4, pp.486 - 500
Received: 13 Sep 2018
Accepted: 28 Jan 2019
Published online: 27 Oct 2021 *