Position Sensorless Control System Of Swiched Reluctance Motor Based On Support Vector Machine

Switched reluctance motor dues to simple structure, low cost, operationalefficiency significantly, the wide speed range and simple flexible control such asexcellent features, has shown its competition in auto drive, household appliances,electric drive vehicles, textile machinery and other fields. Switched reluctancemotor as one kind of synchronous motors is dependent on the rotor sensor innomal motion. But the traditional position sensor limits its scope of application,so this paper proposed a new method of sensorless position detection.This paper describes the research status and principles of the switchedreluctance motor, after analyzing the flux analytical model, the dynamicsimulation model of the entire sysytem is proposed. Analysising of the principleand structure of support vector machine, on the basis of the nonlinear model ofthe motor, a switched reluctance motor rotor positon estimation method based onleast square support vector machine is proposed, which surmounted the problemof nonlinear of switched reluctance motor. Through method of learning offline,the nonlinear relation of current, rotor positon, flux linkage is given, thenswitched reluctance motor indirect rotor position estimation is realized. Theswitched reluctance motor drive and control circuits based on DSP are designed,and test on a real motor, verify the feasibility of scheme. The results show thatrotor position detection accuracy of the switched reluctance motor sensorlesscontrol system based on support vector machine is higher, the motor motion isstable.