Convergence Determination Method for Uncertainty Analysis Surrogate Models Based on MEAM

In recent years, uncertainty analysis has become a hot topic in the field of Electromagnetic Compatibility (EMC), and nonintrusive uncertainty analysis methods have been widely applied due to their advantage of obtaining results without modifying the original solver. Among them, the Surrogate Model Method has attracted widespread attention from researchers in the field of EMC due to its high computational efficiency and resistance to the curse of dimensionality. However, the issue of determining the convergence of the surrogate models seriously affects the computational efficiency and convenience of this method in practical applications. To address this issue, a convergence determination method for uncertainty analysis surrogate models based on Mean Equivalent Area Method (MEAM) is proposed in this paper. The complete convergence time of the Surrogate Model Method can be accurately determined through iterative calculation by this method, and the effectiveness of the proposed method is verified by calculating parallel cable crosstalk prediction examples from published literature. Finally, based on the proposed convergence determination method, the real-time convergence determination problem of the Surrogate Model Method is also preliminarily discussed in this paper, and by establishing a polynomial relationship, the real-time convergence of the Surrogate Model Method can be roughly determined.