Improved accuracy for finite element modeling in virtual air gap lenght computation

Virtual airgap offers a promising approach for continuous control of an inductance using embedded DC auxiliary windings that locally saturate the core. Of interest to designers, it can be thought to be equal to a physical airgap length. Compared to experimental data, a previous attempt to calculate equivalent airgap length using FEM provided inaccurate results. In this paper, the factors contributing to the discrepancy are evaluated and it is shown that although neglecting power loss do not cause serious error in the method, considering magnetic anisotropy in the core steel can noticeably improve the accuracy of the approach.