关键词: 多级同步感应线圈炮, 非参数模型, 循环神经网络, 粒子群优化, 出口速度预测

Abstract: An non-parametric model of multi-stage synchronous induction coil gun (MSSICG) based on the recurrent neural network (RNN) and particle swarm optimization algorithm (PSO) was proposed to solve the problems such as multi-physics field coupling and long iteration time of existing optimization methods. And the ejection velocity of the armature was also predicted by the model. Combing the method of orthogonal and random experiments, the sample set was obtained with the coil turns, trigger time and trigger position as inputs and the ejection velocity as output. The RNN algorithm was used to train the sample set and the non-parametric model was established. By the PSO algorithm, the parameters of the RNN model were further optimized to improve the prediction performance of the model. Using the established PSO-RNN model, the ejection velocity of the armature was predicted and compared with the experimental result. The MSPE, MAPE, and RMSE of the non-parametric model were 0.0028, 0.036, and 2.18 respectively, which reduced by 39%, 38%, and 46% correspondingly after the optimization of PSO. The difference between the predicted and experimental velocity was 1.2m/s with the error percentage of 1.8%, which was less than 5%. The study provided a novel idea for the modelling and engineering design of the MSSICG.

Key words: multi-stage synchronized induction coil gun, non-parametric model, recurrent neural network, particle swarm optimization, prediction of the export velocity