Abstract:The aerospace panel will be slightly deformed due to clamping during hole making, which will reduce the accuracy of blind hole making. Due to the cost of processing, the exact position of the rear panel cannot be determined by means of numerous laser sensors. In order to accurately predict the deformation of aeronautical panels, an improved neural network prediction algorithm was proposed. Firstly, the initial weights and thresholds of the BP neural network were optimized by the particle swarm optimization (PSO) algorithm, then the sparrow search algorithm (SSA) with fast convergence speed and strong global optimization ability was selected for secondary optimization on the weights and thresholds, thereby the SSA-PSO-BP neural network aviation siding deformation prediction model was established. The Abaqus software was used to obtain 50 sets of panel deformation data as the training and prediction data of the neural network (45 training sets and 5 test sets), and the neural network model was trained. In order to verify the accuracy of the built model, three models of BP, PSO-BP and SSA-PSO-BP were used to predict the test set, and MAPE and RMSE were used to evaluate the neural network model. The results show that the neural network model based on SSA-PSO-BP has less error in predicting the deformation of aeronautical panels, and the prediction result is more accurate.