The Applied Research of Rotor Position Sensorless Detection of Switched Reluctance Motor Based on Genetic RBF Neural Network

Jun Hang,You-rui Huang,Lei Shen
DOI: https://doi.org/10.1109/ihmsc.2010.42
2010-01-01
Abstract:Due to the rotor position of switched reluctance motor (SRM) is a highly nonlinear function of stator windings current and flux linkage, so general linear and analytical methods are difficult to achieve precision results, in the paper, a method is presented that a genetic RBF neural network (RBFNN) is used to rotor position sensorless detection of SRM. Hence, extensive mapping ability of neural network and rapid global convergence of genetic algorithm (GA) are fully developed. The simulation is carried out based on the Matlab7.1. The neural network model is simulated for finding the rotor position at different currents from a suitable measured data for a given SRM. In order to testify the validity and accuracy of the model, a lot of simulation is carried out. Results of experiment show that the scheme not only can acquire the rotor position timely and exactly, but also has great robustness and adaptive ability.
What problem does this paper attempt to address?