Nonlinear Analytical Model of Large-Scale Doubly-Fed Induction Generator-Motor Based on Coefficient Transfer

In the variable speed pumped storage unit, the doubly-fed induction generator-motors (DFIGMs) often operate in the high saturation region within power control due to the extensive speed range. To address the issue of motor saturation effect analysis and enhance the calculation accuracy for various operating conditions, this paper proposes a high-saturation analytical calculation method for DFIGMs based on model coefficient transfer learning to analyze the electromagnetic characteristics under extreme operating conditions. The unsaturated condition source domain data were obtained by constructing an accurate subdomain model. Fourier transform was applied to decompose the air gap magnetic density obtained from both the analytical model and finite element model respectively, to achieve precise correction of undetermined coefficients for each harmonic. Consequently, a highly accurate and efficient analytical calculation model for DFIGMs under high saturation was developed. The harmonic and torque characteristics of 10 MW DFIGMs under overload conditions are analyzed based on the model. Subsequently, an experimental test is conducted on a prototype to validate the accuracy of the analytical transfer model.