Virtual Testing Research on the Erection Process of Large-Scale Launch Systems

Weight reduction optimization of large-scale launch systems is crucial for enhancing combat efficiency. To address the challenges of prolonged cycles and high costs in physical experiments, a virtual testing-based approach was proposed. Taking a hypersonic vehicle launch system as research subject, a rigid-flexible coupling model was developed by treating the hypersonic vehicle, launch canister, and erection cradle as flexible bodies while modeling the erection mechanism as a rigid body. Modal characteristics of flexible components were obtained through modal analysis. Erection dynamics simulations under two control schemes were done and the influence of four wall thickness parameters on the erection load-bearing capacity was investigated. Results demonstrated that the feedforward-augmented PD control reduced lateral acceleration overload on the hypersonic vehicle by 75.90% compared to open-loop control. Using the maximum Mises stress of the launch canister as the load-bearing index, parametric studies on four wall thickness parameters revealed that reducing thickness in the forward section (zone IV) improved load-bearing capacity, while thickness variations in the rear area of the front support base (zone II) and the cradle interface (zone III) significantly affected structural performance. A radial basis function (RBF) neural network surrogate model was established based on the load-bearing index (maximum Mises stress) and two critical thickness parameters (zones II and III). Multi-objective optimization for canister mass and load-bearing capacity was implemented using the NSGA-II algorithm, achieving a 12.79% weight reduction while maintaining structural integrity.