Modeling and design of wound-rotor synchronous machines using mesh-based magnetic equivalent circuits
Abstract
In this research, a magnetic equivalent circuit (MEC) model for population-based design (PBD) of salient-pole wound-rotor synchronous machines (WRSMs) is derived. The model builds upon those previously considered by the machines and drives community in several ways. First, a unique approach to represent fringing flux to the rotor poles is developed based upon observations of flux paths seen in finite element solutions. Second, the airgap flux tubes are simplified by including rotor fringing as separate static flux tubes. Finally, the flux tubes representing the rotor pole tip are selected to more effectively capture localized saturation. To enable rapid evaluation of design candidates, a mesh-based solution is applied to solve the MEC. A mesh approach is selected since it has been shown to have superior convergence properties compared to nodal-based models. However, a challenge of mesh-based models of machines is that the mesh configuration in the airgap changes with rotation. This has limited the use of the mesh approach in general. In this research, a relatively straightforward algorithm is derived to determine the mesh configuration based upon the so-called shapes that the airgap reluctances form. Straightforward algorithms are also derived to create the system matrix and the Jacobian matrix used in the corresponding Newton-Raphson algorithm. Finally, a method of coupling the MEC to models of external electric circuits is developed. The MEC model has been shown to match well with finite element models of several machines having a wide range of power levels and topologies. In addition, the MEC model has been applied to optimize a 2 kW portable power generator that was subsequently constructed. Strong correlation between designed and measured performance has been observed.
Degree
Ph.D.
Advisors
Pekarek, Purdue University.
Subject Area
Electrical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.