Generation of an optimal reluctance network model for fast 2-D and 3-D simulations of electromagnetic devices

The reluctance network method (RNM) is one of the numerical methods that allow computation of magnetic fields in different electromagnetic devices. In this method the analysed space is replaced by a network of magnetic reluctances, with magnetomotive forces as supply sources. It allows one to analyse magnetic fields in two-dimensional (2D) and three-dimensional (3D) domains, Turowski et al. (1990), and in comparison to other numerical methods (e.g. finite-element) the calculation time is much shorter, with accuracy sufficient in most cases of engineering CAD. The factors that influence the accuracy of the RNM modelling are discussed. The procedure of obtaining an optimal RNM model is described, using the example of a quasi-3D model of a linear oscillating motor, Mendrela and Pudlowski (1992). The generated RNM model allows one to calculate accurately magnetic fields, coil inductance and forces in the motor in a very short time, Worotynski et al. (1993), even on a personal computer. Computation results are compared with the measured ones