Optimization Methodology for a 2-D Course Correction of a 155 mm Spin-Stabilized Projectile

Overview & Synopsis

An optimization methodology applied to the design of a control device on a 155 mm spinstabilized projectile is presented. The contribution of a spoiler to the aerodynamic coefficients of a controlled projectile is estimated from numerical simulations using surrogate models, more specifically kriging and artificial neural network. Interpolations from the surrogate models are coupled with a 6 degrees of freedom flight mechanics code to compute the trajectories of the corrected projectiles. A 2-D correction of the trajectory is obtained through the maximization of the Expected Improvement computed fromthe kriging model of the objective function during the optimization process. Penalty functions are modeled by the calculation of the Probability of Feasibility relying on the kriging estimations of the performances of the projectiles. An analytic optimization problem is firstly resolved by using this methodology which is then applied to the course correction of an artillery projectile. The geometric characteristics of a spoiler are optimized in order to produce a correction in range and lateral deviation during the projectile flight. A sequential enrichment of the databases is achieved to increase the accuracy of the response surfaces and determine an optimum to the optimization problem.