Study On Efficient Full-wave Electromagnetic Algorithm For RF Circuit And Antenna Simulations

With the development of high-frequency wireless communication technology,RF devices and circuit design must be analyzed by means of electromagnetic fields.The moment method is an effective method to solve the electromagnetic scattering problem.The moment method has the advantages of high precision and good stability,but it needs to solve a dense matrix equation.The computational space complexity is O(N~2)and the time complexity is O(k·N~2).Where k is the number of iterations.With the increasing scale of RF circuits,the traditional method of moments can no longer meet the simulation requirements of large-scale complex circuits.In order to meet the simulation requirements of large-scale circuits,a variety of fast algorithms based on the moment method have been proposed and used.Both radio frequency integrated circuits and printed circuit board-based RF circuits are planar hierarchical structures that use a planar layered Green's function and a fast algorithm independent of the integral kernel to greatly improve computational efficiency.The work of this paper mainly involves a multi-layer Green's function interpolation method(MLGFIM)which has been proposed in recent years and is independent of the integral kernel.The main content of this paper is as follows:Firstly,we implement a multi-layer Green's function interpolation method with space complexity O(N log N)and time complexity O(k·N log N),which is independent of the integral kernel.This method is applied for the electromagnetic simulation of radio-frequency circuits.Next,it is considered that the planar circuit is structurally similar to the near two-dimensional,even near one-dimensional structure.Based on this structure,we propose a new adaptive multi-layer grouping technology to improve the computational efficiency of MLGFIM,thereby improving the simulation efficiency of RF circuits.Then,we used pre-conditional techniques for MLGFIM to further reduce computation time overhead.Using the multi-level grouping of MLGFIM,the interaction between the underlying near-field packets is calculated,and a sparse matrix is constructed.Then the approximate inverse matrix of the matrix is calculated as the precondition matrix,which reduces the k in the solution time complexity.