Permutation Entropy applied to Fault Identification in Soft-started Induction Motors

This work presents a novel methodology to detect and differentiate faults in soft-started induction motors, based on the analysis of the stray-flux and current transient signals using the permutation entropy. The experimental results prove that even in this case, when soft-starters are used to run the motor and therefore the fault related patterns can be more difficult to identify, this methodology allows to detect and differentiate between faults with lower computational requirements than other methodologies. Moreover, this methodology can be used online.