Adaptation of the classical dq method of analysis applied in machines with non-sinusoidal distribution of terminal quantities

The classical dq method for analysis of electrical machines is based on assumption that terminal variables are sinusoidally distributed. In case of non-sinusoidal distribution of these variables, a computationally intensive time-stepping technique is required. This paper presents a methodology, which allows fast and accurate analysis of sinusoidal as well as non-sinusoidal machines. The given technique uses magnetic flux density distribution along the air gap, obtained from finite element solutions for various operating conditions. These distributions are then converted into family of mutually orthogonal functions which forms the basis for analyzing the machine. The influence of magnetic saturation is taken into account by interpolating the magnitudes of the known basic orthogonal functions at corresponding RMS phase currents values. Presented approach enable calculations of back EMF in time domain for arbitrarily chosen load conditions. Validation of the proposed theory was done on an experimental setup comprising of an 8 pole, 2000 rpm, 8 KW machine.