Prediction of Maximum Load in Aircraft Landing by using Artificial Neural Network

초록

Aircraft structures are exposed to high danger of structural failure in landing process compared with other flight situation. Therefore, it is important to guarantee structural safety in landing process. Structural damage detection in landing process might be realized by detailed landing simulation. However, in the actual operating conditions, structural damage should be detected within acceptable time after landing for appropriate maintenance, while reliable landing simulation generally requires unaffordable computing time. Therefore, in this paper, a new approach which can predict landing loads immediately after touch-down by using artificial neural network is developed. First of all, an aircraft model is constructed to generate sample datasets for the neural network training. Then, some landing conditions are selected as input variables for the training. Next, maximum loads from main landing gear to aircraft for given landing conditions are computed by landing simulation, and they are used as target variables of the training. The trained neural network in this work can estimate maximum loads with good accuracy and this means that the developed approach can be extended so that it can be applied to prediction of actual structural damage caused by landing instead of time consuming numerical analysis for each landing case.