Permanent Magnet Synchronous Machines are subject to a variety of failures in various parts of their structure. These faults cause different and independent changes to the motor parameters and its response behavior. This requires different detection and mitigation methods based on the fault type, location, and severity. Therefore, an effective fault detection and identification method is required, not only to identify if the motor is healthy or faulted, but to detect the fault type, separate it from others, and estimate its severity.In this research, an algorithm is proposed to detect and separate between different faults in Permanent Magnet Synchronous Machines under different operating conditions. The incremental inductance approach is proposed when the motor it at standstill. This method uses the changes in the machine saturation, due to the presence of faults, as a fault indicator. Under steady state operation, the change in the machine commanded voltages is proposedas a fault indicator. However, if the motor is operating at steady state with high torque, the motor current or voltage signature analysis is proposed. The main advantage of the proposed method is that it doesn't require any additional hardware components. The same signals that are used for the controller can be used for fault detection, separation, and estimation. The proposed methods also does not require a complicated signal processing techniques. This makes the proposed methods fast, cost efficient and easy to implement.Three common faults in Permanent Magnet Synchronous Machines are discussed in this work: static eccentricity, partial demagnetization and turn-to-turn short circuit faults. Finite Element Analysis simulations and experimental testes were carried out for three Permanent Magnet Synchronous Machines under healthy and the faulted conditions. The differences between the motors are the winding topology, the input/output power, and the slot/pole combination. The first motor is a 12 poles, 72 slots with a distributed windings, the second motor is a 16 poles, 48 slots with a concentrated windings, the final machine is a 10 poles, 12 slots fractional slots concentrated winding machine. Both simulations and experimental results showed that the proposed methods were able to separate between the different faults with a high level of accuracy.
Read
