Abstract:The prediction of torpedo trajectory is of great significance to evaluate the striking performance of weapon systems against targets, in which hydrodynamic parameters are the key link to accurately predict the trajectory. With the development of modern CFD technology, numerical methods are used to obtain hydrodynamic parameters, which improve the forecast accuracy, but the computational efficiency is not high, which is not conducive to torpedo ballistic design and optimization. In this paper, we firstly establish a set of underwater equations of motion for torpedoes based on rigid body momentum and momentum moment theorems, and also numerically solve the established set of equations of motion by using the 4th order Runge-kuta method. Then an approximate model based on genetic algorithm (GA-BP) is proposed on the basis of the BP model, after which the model is optimized and an approximate model optimized by adaptive genetic algorithm (AGA-BP) is proposed. Finally, simulations and data comparison of the three models are conducted, and the results show that the AGA-BP prediction model is more stable and the GA-BP prediction model converges faster compared to the BP prediction model.
