Fault diagnosis of induction motor rotor broken bars based on rectification technique and Goerztel algorithm

Abstract In the early diagnosis of broken rotor bars (BRBs) in induction motors (IMs), the fault sideband frequency under no-load or light-load conditions is weak and difficult to extract. Meanwhile, existing methods for obtaining speed information rely on speed sensors, which increases hardware costs. To address this issue, a speed estimation method has been adopted, and a novel diagnostic technique combination method has been investigated in this paper. This method utilizes the rectification technique as the demodulation technique, and the Goertzel algorithm is employed as the spectral analysis tool. The stator current is demodulated by the rectification technique, and the fundamental frequency component is converted into a DC component, which effectively suppresses the impact of the fundamental spectrum leakage on BRB fault diagnosis and speed estimation. To verify the investigated method, the experiments are carried out on two datasets obtained from the Sao Paulo University and the experimental platform, under different severity of BRBs and loads. The results show that the investigated method can obviously extract the fault characteristic frequency, and it is easy to implement with less computation, strong real-time performance, and low hardware cost.