Speeding up pre-processing time in the CBFM when using very large blocks

The Characteristic Basis Function Method (CBFM) is a numerically efficient technique that reduces the computational burden of Method of Moments when analysing electrically large problems . It works with a set of high-level basis functions, called CBFs, that are defined over relatively large domains, though they are expressed in terms of the conventional subdomain functions. The CBFs are tailored to take into account the geometrical properties of the surface upon which they are defined; hence, these bases are not limited to a predetermined and/or fixed shape, e.g., planar facets. Their use leads to manageable-size matrices, called reduced matrices, that are amenable to direct solution in many cases, without relying upon iteration or preconditioning. Furthermore, for very large problems, the reduction achieved in the number of unknowns translates into significant CPU-time saving if an iterative solution process is utilized afterwards to solve the reduced matrix. The procedure for generating the CBFs is very flexible, and it is possible to use different approaches depending on the required accuracy and the features of the geometry under analysis. The geometry must be divided into different blocks in a preprocessing stage, so that each one of them will support several CBFs.