Reduced order model approaches for predicting the magnetic polarizability tensor for multiple parameters of interest

Abstract The magnetic polarizability tensor (MPT) is an economical characterisation of a conducting object, which can assist with identifying hidden targets in metal detection. MPT’s coefficients depend on multiple parameters interest including the object shape, size, electrical conductivity, permeability, and frequency excitation. computation follow from post-processing eddy current transmission problem solved numerically using high-order finite elements. To reduce computational cost constructing these characterisations for different parameters, we compare three methods by MPT be efficiently calculated two-dimensional parameter sets, levels code invasiveness. We compare, numerical examples, neural network regression eigenvalues projection-based reduced order model (ROM) enhanced ROM (POD–NN) predicting coefficients.