Fault Identification in Electrical Actuators by Autoregressive Models

Considering that several kinds of electrical actuators are found in real physical systems, in this paper is proposed a fault identification strategy for electrical actuators with autoregressive models. As it is known, autoregressive models are those kind which their parameters are tuned in order to adjust the parameters of a transfer function in order to model the nonlinear dynamics of a physical system. So in this paper is proposed an auto- regressive model in order to identify the failures in two kinds of electrical actuators; that in this case is a permanent magnet synchronous motor. The kinds of failures which are treated in this research study are basically electrical faults, which are related to the stator failures depending in the characteristics of the motor. The auto-regressive model is used in a configuration in which this failures are detected, and then by the stability characteristics of the dynamic system which represents the motor, according to the difference of the estimated and measured outputs a threshold is selected in order to identify a failure and to determine which kind of failure occurs in the motor. Some simulation experiments are shown in this paper in order to validate the theoretical results obtained, taking into consideration that many mechanical systems such as robots use this kind of actuators.