| Stator winding insulation breakdown is one of the major causes of electric machine failures. Therefore, prevention and detection of such a failure are crucial for increasing the machine reliability and minimizing the financial loss due to motor failure.; This work focuses on developing a reliable and practical R S-based stator winding thermal monitoring scheme for induction machines. The motivation for RS-based thermal monitoring is to provide a sensorless temperature monitoring scheme that provides an accurate temperature estimate that is capable of responding to the changes in the motor thermal characteristics. Based on the evaluation of positive, negative, and zero sequence model-based RS estimation, it is concluded that RS estimation using the dc model provides the most reliable results. A new on-line stator resistance estimation scheme for line-connected machines using dc signal injection is proposed in this work. The new dc injection circuitry consists of a power MOSFET and a resistor, and provides controllable dc injection with negligible power loss. Experimental results verify the validity of using the proposed scheme for RS-based temperature monitoring.; In addition to temperature estimation, the estimate of R S is further processed to extract the effective stator thermal time constant information for monitoring the condition of the motor cooling system. Detection of defects or faults in the cooling system prevents the motor from operating at an increased temperature, and results in extended insulation life. The experimental results of the proposed scheme show that abnormal cooling situations can be detected from the RS estimates without additional hardware or parameter information.; In addition to the new RS estimation scheme, a simple, on-line stator winding turn fault detection method based on monitoring an off-diagonal term of the sequence component impedance matrix is developed. It is shown that the proposed method is immune to the motor non-idealities such as supply voltage unbalance variation, inherent motor asymmetry, and measurement errors. It is verified by experimental studies that the proposed method is capable of detecting small levels of turn faults reliably in an early stage. |
