Simulation research of muzzle disturbance based on artificial neural network

In order to reduce muzzle disturbance of gun, establish virtual prototype of a type of howitzer based on complex factors such as mass eccentricity of recoil parts, soil stiffness and damping, clearance of elevating mechanism, stiffness and damping of papilionaceous spring. Use velocity and angular velocity to reflect muzzle disturbance, then examine, simulate and analyze the virtual prototype. According to the simulation results, the virtual prototype including clearance and papilionaceous spring of elevating mechanism can reflect actual launch process and dynamics performance. Orthogonal Design Method is introduced to design simulation experimentation so as to obtain the sample couple. The back propagation algorithm is used to study and train the sample couple based on artificial neural network, then the nonlinearity relation was achieved between the design parameter and performance index. Establish the objective function, then the appropriate optimization arithmetic can be used to optimize vital parameters to reduce muzzle disturbance independent of virtual prototyping technology. Moreover, it can provide theoretical guide for optimization of muzzle disturbance.