Estimation of Maximum Strains and Loads in Aircraft Landing Using Artificial Neural Network

Hard landings account for a large proportion of aircraft accidents and are generally judged by the intuition of pilots. For this reason, there are frequent false judgments that lead to unnecessary and costly ground inspections. False judgments can be reduced significantly if detailed load information is available, such as strains or loads on critical areas of aircraft structures. This study used an artificial neural network (ANN) to develop a numerical model that can estimate the maximum strains for areas of interest and landing loads from basic flight parameters. The results can be used to provide the required detailed load information. An efficient and accurate landing simulation model was constructed and used to build reliable datasets for training. Basic flight parameters from immediately after touchdown were used as input data for training, and the corresponding maximum values of strains and landing loads were obtained from the landing simulation model as target data for training. This information was used to train the ANNs with the Levenberg-Marquardt backpropagation algorithm. A performance evaluation using test data confirmed that the trained ANNs can successfully estimate strains and landing loads with sufficient accuracy.