As one of the most important parameters in evaluating the status of a DC motor, commutation spark has been widely used in condition monitoring and fault diagnosis. A new approach using image processing techniques on the commutation spark has been proposed. Advantageous over other methods using partial spark information, the details about the motor condition can be comprehensively retained and extracted from the sparking images. The sparking images were obtained by the cameras, which were fixed at the certain sites in DC motor beforehand. The images were processed through the sparking image preprocessing, segmentation, enhancement, and feature extraction. Comparing the characteristic parameters of the sparking image with the result of the subjective grading, the relationship between the parameters and the sparking grades had been analyzed to monitor the DC motor. The effectiveness had been demonstrated by application examples.
Comparative Study Between Classical Controllers and Inverse Dead Zone Control for Position Control of a Permanent Magnet DC Motor with Dead zone