Research On Extremely Low Speed Estimation And Measurement Methord Of Indunction Motor Drives

The high-performance induction motor frequency control system plays a vital role in the national economy since field-orientation vector control is proposed, and is widely used in many industries such as mining, spinning and transportation. As the encoder measuring the rotor speed is very expensive and easily damaged, the speed sensorless techniques are rapidly developed these years to allow the system operate in severe industry circumstance and reduce the overall cost. Among these techniques, the adaptive full-order observer(AFO) receives widespread attention for its accuracy, parameter sensitivity and practicability. In this paper, the AFO is studied focusing on its stability and low speed estimation ability. Besides, a virtual speed oscilloscope program is developed based on HEIDENHAIN IK220 counter card to measure the very low speed.For the stability of full order state observer, the redesign of observer’s feedback gain and rotor resistance identification method are studied. The AFO algorithm stability includes regenerating speed estimation and parameter robustness. The estimation transfer function is established to derive the algorithm stable condition in regenerating mode. The influence of parameters deviation on flux observation is analyzed according to error equation between the actual and observed flux. Based on these work, the observer’s feedback gain is redesigned, eliminating the estimation unstable zone and enhancing the parameter robustness at the same time.The rotor resistance, which has poor robustness, is estimated on line through Kalman filtering method.Aiming at the inaccurate speed identification problem at very low frequency, the adaptation law is improved combined with low-frequency signal injection method.The limitation of AFO’S observability and its adaptation law is studied at zero stator current frequency. Through the frequency signal injection method, flux error is added to speed adaptive law. Both the estimation accuracy and sensorless low speed taking load operating performance are improved. A virtual speed oscilloscope software using IK220 to measure and observe low speed waveform is designed,which enhance the measurement accuracy significantly.Finally, this paper introduces the 2.2k W induction motor experiment platformwith STM32 ARM as the control unit. The IM control system’s hardware configuration, overall framework and software design processes are presented. The improved algorithm and speed oscilloscope program are verified on that platform.