Efficient Algorithm Based On Characteristic Basis Function Method And Its Applications In Electromagnetic Scattering Problem

With the development of electronic technologies for aeronautics, astronautics and navigation, the techniques of radar target stealth and anti-stealth, RCS and electromagnetic compatibility are required for accurate analysis on 3-D objects with complex structures. Because of the request, this is becoming a hot topic in electromagnetic field. In this dissertation, the research on one branch of computational electromagnetic has been studied. Based on the surface integral equations expanded by moment method, a new method called adaptively modified characteristic basis function method (AMCBFM) is proposed. The whole study clue is based on this method in order to seek one effective approach to analyze the electromagnetic characteristic of some intricate targets. The universal programs and accelerative technologies are required to slove the complex electromagnetic characteristics of different oblects.This dissertation includes two aspects in principle. One part is the study on the new method, including its principium and related accelerative technologies, and the other part is the applications for analyzing electromagnetic scattering problems on homogeneous dielectric objects as well as mixed homogeneous dielectric and PEC objects. The parallel algorithm executed in windows environment is also discussed.In the first portion, the AMCBFM is proposed. This method divides the studied objects into many blocks and uses an adaptively modified technique to overcome the shortcoming which is not convenient to choose the levels of characteristic basis functions (CBFs) in the conventional characteristic basis function methods (CBFMs). And a simple way is applied to control the current error precision. Some different parameters including the number of blocks and the levels of CBFs are discussed to find out the effect on the computational efficiency, and the appropriate values are turned up. The results from the method agree well with that from the conventional methods. Subsequently, the QR factorization based on Dual-MGS is combined with AMCBFM to reduce the RAM storage and improve the computational speed. Moreover, it is successful to use the model based parameters estimation (MBPE) technology and AMCBFM in analyzing the monostatic RCS and wideband response of large electric objects. And the influence of several parameters on the results, such as the third sampling point, precision error and subsection sampling, is discussed and it shows that the number of sampling points reduces sharply.In the second portion, the studies on the further application of AMCBFM are carried out. The PMCHW equations for homogenous dielectric objects are reviewed, which are based on the equivalent principles and surface integral equations. After combined with AMCBFM, the new method, PMCHW-AMCBFM, is proposed. Moreover, the overlap domain between adjacent blocks and the levels of CBFs are also investigated to seek the suitable values. Subsequently, the surface integral equations, EFIE-PMCHW, for the electromagnetic scattering problem of mixed homogenous dielectric and PEC objects are introduced and combined with AMCBFM triumphantly called as EFIE-PMCHW-AMCBFM. And the electromagnetic scattering of microstrip antennas and microstrip antenna arrays are analyzed accurately, which validates the validity and veracity of this method. At last, the dynamic-static load balance parallel algorithm based on AMCBFM is executed in windows environment. This algorithm overcomes the disadvantages of static load balance parallel algorithm and has the advantages of dynamic load balance parallel algorithm. And the algorithm combined with QR-AMCBFM and EFIE-PMCHW-AMCBFM is implemented to solve the electromagnetic scattering problems aroused by various large electric objects effectively.