Impacts are common damage events in aviation scenarios that can cause damage to the structural integrity ofan aircraft and pose a threat to its safe operation. Therefore, it is crucial to monitor impact events. A region-to-point monitoring method is proposed to address the challenges posed by the large area of monitored aircraft structures and the long distance between sensors. Firstly, to fully use the information in the original impact signal and reduce the aliasing effect caused by the reinforced structure, the original signal is decomposed into several modes with different frequency bands by Variational Mode Decomposition (VMD). The Multi-scale Permutation Entropy (MPE) value is then calculated to reflect the various characteristics of each mode, which is used as a basis for classification. Secondly, Transfer Component Analysis (TCA) is selected as a transfer learning method to reduce the difference between the features of the source domain and the target domains' features. Thirdly, the TCA-transformed source domain data are used to train the Probabilistic Neural Network model (PNN), and the unfamiliar target domain data are used to verify the impact area identification. Finally, based on regional location, the system identification technology and weighted centroid algorithm can be used to obtain the history of impact force and the precise coordinates of the impact location.
Keywords: Aircraft structure; Impact force history; Impact location; Structural health monitoring; Transfer learning.
Copyright © 2023 Elsevier B.V. All rights reserved.