Accurate position estimation is an integral part of any motion control system. Besides, instantaneous position of poles plays an important role in electronic drive system of modern PM machines. Resolvers, due to their structure, are the most reliable position sensors in harsh environments. However, resolvers are not as accurate as optical encoders in normal conditions. To increase the accuracy of these sensors, various solutions are used, such as optimization in winding arrangement, rotor and stator contour, slot–pole combinations, and improvements in resolver-to-digital converter. In this paper, fractional slot concentrated winding is used in an axial flux resolver. By the aid of winding function approach, it is shown that increasing the coil pitch improves the estimated position accuracy. The theory is tested for double-layer and four-layer windings with different displacement angles. A new double-stage resolver is also introduced. The proposed structure, in which a meta-heuristic algorithm determines the slot dimensions, the number of turns, and the displacement angle between the stages, results in lower total harmonic distortion of output voltage and, consequently, lower estimated position error. The theoretical and analytical expressions are verified by a three-dimensional time stepping finite element analysis and experimental results. Good agreement between simulation and experimental results validate the advantages of the optimized resolver.
Read full abstract