Research on Constructing Surrogate Model of Rocket Aerodynamic Discipline Using Extreme Learning Machine

摘要

In order to decrease the long time consuming of the aerodynamic numerical calculation in the multi-discipline designing of the rocket, this paper proposed a surrogate model of rocket aerodynamic based on extreme learning machine, and described the process to construct the surrogate model in detail. Based on aerodynamic surrogate model of a rocket, this paper evaluated the precision of the surrogate model with a small learning sample, and it proved that this method is very feasible and effective.