Sensorless Control Algorithm For PMSM Drives Using Auxiliary Signal Injection

PMSM is widely applied in many industry circumstances due to its high powerdensity and efficiency. Most of the general-ppurpose inverters for PMSM requirethe function of sensorless control and should be highly optimized for general use.Further, it is important to investigate parameter identification algorithm forcontroller parameter self-tuning and observer parameter design.First, an off-line parameter identification method is studied viavoltage-source-inverter (VSI) and no extra devices are required. This methodidentify the motor parameter after obtain the intial rotor position by injectingauxiliary signal. Stator resistance is identified through linear regression to suppressthe influence of inverter nonlinearities and the inductance identification is real izedby injecting high frequency signal, moreover, corresponding mathematical model isestablished to analyse the error which is verified through the experiments. Finally,model conpensation and adaptive compensation are put forward upon this model toreduce inductance identification error.In order to realize PMSM sensorless control during low speed region, a PMSMsensorless control strategy based on high frequency square-wave signal injection isstudied. This method obtains higher injection frequency under certain carrierfrequency which can increase the speed loop bandwidth and is suited for low carrierfrequency control system. A simplified signal processing method is proposedwithout to extract current information in static frame using coordinate rotatingtransition. Then quadrature phase-locked loop is combined to accomplish theelectric angle estimation. This method enjoys less calculated amount and morepractical application. A robust position observer is designed upon mechanical modelof the electric machine which increases disturbance rejection capacity of thealgorithm. Meanwhile, the relevant parameter design is given as well.Finally, the proposed algorithms are achieved on an experiment platform basedon ARM that test the parameter identification as standstill and performance of thesensorless control strategy. The identification results compared with the instrumenttests indicate great accuracy and the studied sensorless method enjoys highprecision and control performance.