FEM based Methods for Uncertainty Quantification in Electromagnetics

Abstract

Many electromagnetic problems are subjected to variations in geometry and material properties, which have an impact on the performance. So quantifying these uncertainties helps in performance prediction and optimization in manufacturing of such electromagnetic systems. In addition, a statistical analysis of the response of such stochastic systems provide a better insight for experimental testing and validation. In this paper, the material uncertainty is modelled as a gaussian random process and is discretized using Karhunen Loeve expansion. The system response based on finite element modelling is analysed for its impact using perturbation method, intrusive and non-intrusive spectral stochastic finite element method (SSFEM) and stochastic collocation method. The stochastic responses from these methods are validated using Monte Carlo method, and the results indicate that SSFEM is the most computationally efficient approach for this problem.