面向火箭构型论证的运载能力快速分析方法

Aiming at the problem of fast calculation of the carrying capacity under different launch vehicle configurations and task conditions, a speed loss calculation method based on the Gaussian function and combined neural network is proposed. Firstly, based on the analytic solutions of state variables, the Gaussian function is used to fit the core velocity loss from gravity. Meanwhile, in order to improve the sampling density of multi configuration and multi task samples, simplify the data modeling process and enhance the adaptability of the method, a combination form of radial basis networks (RBF) and deep neural networks (DNN) is used for the extraction and regression analysis of state variables. Then the task constraints are transformed into the required velocity increments, and the carrying capacity is obtained by numerical iterations. The simulation results show that the accuracy deviation of the proposed calculation method is about 0.35%, and the calculation time is less than 2 seconds, which can provide theoretical support for the rapid demonstration of the overall parameters of the launch vehicle and the research of task planning.