Broken rotor bar fault detection in line start-permanent magnet synchronous motor

High efficiency motors are being gradually exerted in many industrial applications because of their positive impacts on the environment by reducing energy consumption and CO2 emission. In this regard, Line Start Permanent Magnet Synchronous Motors (LS-PMSMs) have been introduced to the market recently. Due to the unique configuration, LS-PMSMs are allowed to reach Super Premium Efficiency levels accompanied with high torque and power factor. However, since the use of LSPMSMs in industry is in its infancy, no efficient scheme has been reported for faults detection in this type of motor. Online monitoring and setting of preventive maintenance programs in the industries is one of the important issues. Therefore, in order to classify different indices of motor under fault condition, the electrical behavior of LS-PMSMs motor under broken rotor bar should be considered and the electrical parameters should be characterized. The main aim of this research is to investigate the effects of broken rotor bar fault on LS-PMSMs performance, and also to find reliable fault-related feature for this fault. The proposed detection strategy for broken rotor bar in LS-PMSM is based on monitoring of startup current signal. In this regard, a simulation model and experimental setup for investigation of broken rotor bar in LSPMSM is obtained. The current signal is used to extract the fault-related features using three different signal processing method. Finally, the ability of these features is validated for detection of broken rotor bar in LS-PMSM through statistical analysis. This study can be beneficial for the industry by using the online monitoring systems where the motor fault can be detected during its operation. Therefore, the proposed method can be used in the preventive maintenance programs. This research indicates the importance of load effects on broken bar detection in LSPMSMs. The current signal is collected at different load levels of starting torque within four steps, which increases from 0% to 65%. The experimental and simulation results substantiate that increasing the load, will also increase the starting time duration. The time duration of machine with one broken rotor bar also increases compared to healthy condition. The value of starting torque drops in the presence of broken rotor bar fault. In the time domain analysis, three features, namely peak to peak, shape factor and impulse factor cannot distinguish faulty state of motor from healthy state based on upward or downward trend. Skewness also fails to detect broken bar when the starting torque is high. In time domain analysis using of envelop signal, four features, namely RMS, RSSQ, Energy and Variance cannot distinguish faulty state of motor from healthy state at low level load. The variance feature also fails to detect the fault based on upward or downward trend. When the starting torque is high, Kurtosis feature is not a suitable feature to detect broken rotor bar. In the time-frequency domain analysis, Log Energy Entropy feature has satisfactory performances for broken rotor bar detection compare to Shannon Entropy feature. The result also presents that the most effective sub-band frequency is Detail of level 7 that includes the frequency band ranges of [39.06-19.53]Hz. The simulation results were validated with an experimental work to confirm the effectiveness of proposed methods.

Preview Text FK 2016 162 IR.pdf Download (1MB) | Preview

Actions (login required)

View Item