| Research on electromagnetic scattering from electrical large target, randomly rough surface, and the composite model of target and rough surface, has been attracting more and more attention in recent years, due to its extensive applications to airborne radar observation, target identification and acquisition, ground-penetrating radar, etc. However, for practical engineering application, it is urgently desired to develop fast and effective numerical computation methods for these complex electromagnetic models. The main contribution of this dissertation is the development of a series of novel methods for computerized scattering calculation of targets and randomly rough surfaces.In numerical simulation of scattering from randomly rough surface, the tapered wave is often used to eliminate the "edge effects" caused by the finite truncation of rough surface. The taper width g and surface length L are very important for computation efficiency and effectivity. However there is no explicit and feasible criterion for choosing these parameters. In chapter 2, a new feasible criterion is proposed, which relates the taper width g and incident angleθ_i, in base of theHelmholtz wave equation. Then the surface length L is determined, according to multiple considerations such as the correlation characteristics of rough surface and the reasonable energy truncation. The criterion of parameters for numerical simulation of rough surface scattering is validated numerically.In numerical simulation of individual target scattering, the Fast Multipole Method (FMM) and the Multi-Level Fast Multipole Algorithm (MLFMA) have been developed to accelerate the multiplication operation of impedance matrix and current vector for scattering computation of electrical large size target. The numerical quadrature for plane wave expansion of sphere wave function is very important for computation precision of FMM and MLFMA. Chapter 3 analyzes the integrand spectrum in the expansion equation, and presents a new numerical quadrature standard, which satisfies the sampling theorem. Compared with previous quadrature scheme, the new integration standard is testified numerically to be appropriate for any cluster size.In chapter 4, the composite target and rough surface scattering computation is discussed, and a fast and effective coupling iterative algorithm is developed. Based on the Green's function of half-space above an infinite rough surface, new coupling integral equations are derived for difference scattering field computation, which describes the interactions between the target and rough surface. Compared with Johnson's numerical model, the new derived equations compute differene scattering field, need not to compute two cases of with and without target above the rough surface respectively. The difference radar cross section (d-RCS) takes account the scattering from the target and its multiple interactions with the underlying rough surface. Since the direct scattering from the rough surface is excluded, the d-RCS removes the effect of different illuminated surface length under the tapered wave incidence. Moreover an iterative approach, combining CG (Conjugate Gradient) for target and FBM (Forward Backward Method) for rough surface, is developed to solve the coupling equations. Because the scattering contribution from the rough surface is dominated in the specular direction, only a small section of rough surface in the specular direction is needed to compute the rough surface scattering contribution on the target, which makes the computation much faster. Moreover, how to choose the parameters g and L is discussed. Using Monte Carlo method to realize the ocean-like rough surface with P-M (Pierson-Morkowitz) spectrum, the bistatic scattering from the 2-D perfect electrical conduct (PEC) target above an oceanic rough surface is numerically simulated.Moreover, the coupling FBM-CG iterative approach is generalized for dielectric target and multiple targets scattering above a PEC rough surface in chapter 5．For each target, different surface segments in the specular directions are truncated to compute the rough surface scattering contribution on each target. In iteration process, appropriate surface sections are truncated for each target computation respectively.Because the numerical solution of rough surface's integral equation occupies much memory and CPU time, which is a bottleneck for electrical large size target scattering and rough surface scattering under low grazing incidence, chapter 6 further develops a hybrid iterative algorithm, which effectively combine the analytic KA (Kirchhoff Approximation) for rough surface scattering computation, and the numerical MoM (method of moment) for target scattering. The KA and MoM coupling iteration takes account the interactions between the target and the underlying rough surface. Since is just one numerical integral of induced current on the target should be performed in KA computation, much memory and computation time is reduced. With Monte-Carlo method, this hybrid KA-MoM iterative algorithm is used for scattering computation from a 2D and 3D target above a randomly rough surface.The works of this dissertation provides a new set of feasible numerical methods with high efficiency and accuracy for scattering computation of electrically large target, randomly rough surface, and the composite model of target and rough surface.
|
PDF Full Text Request