Four-quadrant position sensorless control of switched reluctance motors based on on-line modeling

Xulong Zhang,Guojun Tan,Songyan Kuai,Songyong LIU,Qihu Wang

2012-01-01

Abstract:A position sensorless control method based on on-line modeling is presented for switched reluctance motors(SRM). Radial basis function(RBF) neural network of rotor position estimation for sensorless SRM control is established with two input variables: flux linkage and phase current. Real-time rotor position angle obtained from shaft encoder is adopted as learning sample data, and on-line modeling of SRM is realized. Excitation pulse method combined with sliding mode observer(SMO) strategy is proposed and four-quadrant SRM sensorless operation is achieved. Experimental results indicate that rotor position estimation error less than 2° and switching between quadrants is reliable, and the system has good dynamic and static performance. The proposed method is helpful for high-power sensorless control of SRM.