Online Fault Detection and Diagnosis in Three-Phase Induction Machines using the αβ-Vector Approach: A Practical Implementation

This paper focuses on the problem of online-based fault detection and diagnosis of induction motors. The overall idea is not restricted to develop a new on-line algorithm but to extend it towards it's implement on a protective Intelligent Electronic Device (IED). This paper proposes a new approach for on-line fault detection and diagnosis, making the fault diagnosis process fully automatic (without expert inspection) and co-allocated with the protective algorithms implemented at the IED. Since the algorithm is IED built-in, it uses the stator currents acquired by the IED itself. These currents are converted into an alpha-beta space state and the correspondent Principal Components are computed (inside the IED). Based on this computation, the fault diagnosis algorithm decides upon the fault (or no fault) conditions and reports them via the IED's HMI (Human-Man Interface). In this paper the foundations of the fault detection and diagnosis algorithm are presented, and shown how it was implemented in the IED. Field experiments confirm this approach as feasible and it is being considered for commercial use.