Sensorless control of direct-driven permanent magnet wind power generation system based on improved MRAS

Permanent magnet motors are coming more and more popular in wind generation applications. In the high efficient wind power generation system, the accurate information about rotor position is very important. MRAS schemes offer simpler implementation and require less computational effort. Because the voltage model involves pure integrator in the classical model reference adaptive system (MRAS). This will cause initial value and drift problems. These problems affect the precision of rotor position. This paper describes a improved sensorless vector control strategy of PMSG based on MRAS. It uses high-pass filter in the voltage model to avoid these problems. To avoid phase and amplitude errors brought by the high-pass filter, it has to insert high-pass filter in the current model simultaneously. To improve dynamic performance the system uses fuzzy controller to adapt the cut-off frequency on-line and replace the traditional PI controller. We implement a 7.5kW direct-drive generation system centered about the DSP, and through some laboratory experiments, we will show that the improved method is capable of precise estimating the rotor position under the condition of high or low speed.