Detection of inter-turn short-circuit in induction motors using Park–Hilbert method

Abstract

In this paper, a new method, called Park–Hilbert (P–H), is presented and used to detect inter-turn short circuit in three-phase induction motors. This technique is based on spectral analysis, via the FFT, of a new generated signal called PVMP–H which represents the Park’s vector modulus computed starting from the analytical signals obtained from the three-phase current by the Hilbert transform. A theoretical analysis is performed in order to show how the proposed method works. Various experimental tests carried out on a 3 kW three-phase induction motor, illustrate the efficiency of this new method and show that some spectral components of the PVMP–H are extremely sensitive to the occurrence of shorted turns. However, it is shown that the information about the motor state is distributed on many spectral components and in order to gather these useful information, the present paper suggests the monitoring of a new fault indicators computed starting from the most sensitive harmonics (basing on the multiple signature processing). This technique is applied on the proposed signal PVMP–H; and the experimental tests confirm that a small number of shorted turns can be effectively detected using these new fault indicators even under low load. A comparative study with the conventional MCSA and EPVA methods is also performed. The obtained results show the superiority of the P–H method and prove the effectiveness of the proposed fault indicators for extracting useful information from any motor signal.