| Summary: | 碩士 === 國立交通大學 === 機械工程系所 === 103 === Brushless DC motors (BLDCM) has the advantage of high efficiency, high reliability, less noise and maintenance free nature as compared with brushed DC motors. In general, BLDCM relies on the Hall Effect sensors to detect the rotor position for switching of inverter. However, in some extremely working environments such as electromagnetic interference (EMI) and high temperature surrounding, the Hall sensors are not suitable; the operating region is restricted. For high reliability required system, sensor failure may cause instability in the control system. Not only that, the space and costs reduction make sensorless drive more popular. Sensorless technique of permanent magnet motor can be classified into three categories: using motor back-EMF, using inductance and using flux linkage. For BLDCM, which always use motor back-EMF can further divided into three categories: direct back-EMF detection, in direct back-EMF detection and the techniques using model-based estimation. However, sensored and sensorless drive techniques both suffer from several error sources during implementation, resulting in commutation error. Commutation error is unavoidable unfortunately.
This thesis analyzed in depth with current waveform under commutation error, non-ideal back-EMF, armature reaction and realized that current waveform is function of commutation error. A current integral method is proposed for commutation error compensation. This method compared two 30 degree current integration in one 60 degree conduction. The difference is caused by commutation error, and can feedback eliminated by PI controller. Let the motor having less noise, vibration and torque ripple. Experiment results showed that the proposed method is capable of running in different speed, load torque and speed variation operating with good performance.
|
|---|
