Design of a torque observer for detecting abnormal load

Abstract In manufacturing processes, human error and abnormal operation often result in excessive loads on the servomechanisms. To prevent the abnormal load from damaging the work piece or tool, it is important to find a means to detect the abnormal load so that an appropriate response can be performed. To estimate the load on the servomechanism, two types of torque observer structures are presented in this study. Based on the information obtained from the torque observer, criteria are proposed to determine whether the servomechanism is under the abnormal load, while the Wavelet analysis is employed to enhance the ability of abnormal load detection. Experimental results indicate that the presented observer structures are able to estimate the external torque accurately. In addition, the structure combined with Wavelet analysis is capable of identifying abnormal load events in servomechanisms.