Ali Davoudi with adviser Patrick L. Chapman

This project is investigating order reduction of highly accurate physics-based models to ease computationally inefficient modeling of magnetic devices. In prior work, we have successfully implemented model-order reduction techniques of linear and nonlinear stationary magnetic devices. The next step extends the techniques to electric machines, which are much more complicated. Techniques such as the finite element method or magnetic equivalent circuits can be used to set up an initial, highly detailed and accurate model. Such models are, however, omputationally very intensive. The core of this research is to take such high-order models and reduce them mathematically so that they remain accurate while becoming much more useful. The proposed techniques are systematic and facilitate total automation of modeling. Specific potential challenges include accommodating relative motion, nonlinear core losses, and laminations. Managing computer resources for limited memory and precision is also a challenge. These challenges represent the direction of our current work, which began earlier this year.