Parametric Geometric Metamodel of Nonlinear Magnetostatic Problem Based on POD and RBF Approaches
Allaa Eddine Boumesbah, Thomas Henneron
Abstract
A parametric geometric metamodel is built for a nonlinear magnetostatic problem, using proper orthogonal decomposition (POD) approach combined with radial basis functions (RBFs) interpolation. Furthermore, the geometrical variation of the problem is modeled using an RBF interpolation for smooth mesh deformation. The metamodel is applied for a single-phase EI inductance, and the aim is to create precise flux cartographies based on few solutions of the original finite element (FE) model. The results show that the POD-RBF approach can reduce efficiently the evaluation time of a parametric nonlinear magnetostatic problem.
Topics & Concepts
Interpolation (computer graphics)MetamodelingParametric statisticsRadial basis functionNonlinear systemFinite element methodComputer scienceParametric modelAlgorithmApplied mathematicsMathematicsArtificial intelligencePhysicsStructural engineeringEngineeringArtificial neural networkProgramming languageMotion (physics)StatisticsQuantum mechanicsModel Reduction and Neural NetworksProbabilistic and Robust Engineering DesignNumerical methods in engineering