An Adjoint-based Shape Optimization Design System for the Radar Cross Section Reduction

An efficient shape optimization design system for the Radar Cross Section (RCS) reduction using the adjoint approach is discussed. The Method of Moments (MoM) and the Multilevel Fast Multipole Algorithm (MLFMA) are applied to solve the Combined Field Integral Equation (CFIE) or the Poggio-Miller-Chang-Harrington-Wu-Tsai (PMCHWT) equation to calculate the RCS. As for the high-frequency scattering problem, the Physical Optics (PO) method is selected to calculate the RCS due to its efficiency. The gradient of the RCS is obtained by solving the adjoint equations. The code that calculates the derivatives is developed with the help of the Automatic Differentiation (AD) technique. The Free-Form Deformation (FFD) approach is used to parameterize the shape and the Sequential Quadratic Programming (SQP) is chosen as the optimization algorithm. Without a large number of function evaluations, this design system could efficiently obtain the shape that meets the stealth requirements.