A Distance-Adaptive Integration Strategy for Accelerating Electromagnetic Modelling of Deep Cavity Problems

The computation of electromagnetic scattering from electrically large deep cavities poses significant challenges to existing full-wave numerical methods. In addition to the well-known convergence issues, the accuracy of such problems is also found to be sensitive during the numerical implementation process. To solve the cavity problems, based on the integral equation (IE) method, the method of moments (MoM) can be used to obtain a linear equation system. In this process, a significant factor affecting accuracy is the rule of Gaussian quadrature. Conventionally, to obtain good accuracy, high-order Gaussian quadrature rules are necessary, which in turn increase the computational time for filling the impedance matrix. To maintain accuracy while improving numerical efficiency, a distance-adaptive integration strategy (DAIS) for Gaussian quadrature is proposed in computing the matrix elements. Numerical experiments show that IE methods with DAIS can efficiently solve various types of deep cavity targets. Moreover, it can be flexibly embedded into the classical multilevel fast multipole algorithms (MLFMA) to obtain good performance for targets involving electrically large deep cavities.