Research on a Mechanical Sensorless Vector Control of Induction Motor Using Image Processor

2004年至2005年，长崎大学工程学院开展了“基于图像处理器的感应电机机械式无位置传感器矢量控制研究”。提出了一种新型的混合型感应电机矢量控制系统，将基于磁链观测器的无速度传感器矢量控制与图像传感器(CCD摄像机)的信息相结合。图像传感器的信息在稳态运行中是有用的，适用于识别转子电阻。对实验系统中的数字信号处理器(DSP)控制器、逆变器、摄像机、视觉信息处理板、感应电机和直流发电机(负载)进行了安装和测试。开发了包含磁链观测器、速度估计、参数辨识、电流控制器和速度控制器的软件。(3)仿真和实验结果利用数字仿真和实验设备对系统的稳态和暂态特性进行了测试。由于数字仿真得到的结果与实验结果非常接近。验证了利用视觉传感器获取的信息成功地完成了转子电阻的辨识。结果表明，提高了速度估计的精度，获得了良好的暂态响应。此外，通过控制磁化电流，使铜损耗和铁损耗之和最小，从而提高了效率。由于图像传感器计算出的转子转速具有较大的时滞，且在稳态运行时有效，因此采用该转速进行矢量控制是不可能的。

The research project titled "Research on a Mechanical Sensorless Vector Control of Induction Motor Using Image Processor" has been carried out from 2004 to 2005 in faculty of engineering, Nagasaki University. The results obtained from this project are summarized as follows ;(1)The proposed system and its designWe have proposed a new hybrid type vector control system of induction motor in which a flux observer based speed sensorless vector control and the information of image sensor (CCD camera) are combined. The information of image sensor is useful in steady-state operation and is adapted to identify the rotor resistance. A simulation program and a linear model is derived for the analysis of proposed system.(2)Experimental systemA digital signal processor (DSP) controller, inverter, CCD camera, visual information processing board, induction motor and DC generator (load) are equipped and tested. Software which contains flux observer, speed estimation, parameter identification, current controller and speed controller is developed.(3)Simulation and experimental resultsSteady-state and transient characteristics are tested by digital simulation and experimental equipment. Since the results obtained by digital simulation are very close to the experimental results. It is confirmed that the rotor resistance identification is accomplished successfully by using the information from visual sensor. As the result, the accuracy of estimated speed is improved and good transient responses are obtained. Furthermore, the efficiency is improved by controlling the magnetizing current which minimizes the sum of copper loss and iron loss. Since the rotor speed computed by image sensor has long time lag and it is valid under steady-state operation, the vector control using this speed is impossible.