Modeling of PMSM Full Power Converter Wind Turbine with Turn-To-Turn Fault

Licentiate thesis, 2014

Over the past several years the number of renewable energy installations has shown a strong increase rate. Due to the ambition to close down fossil and nuclear power plants, this rate of increase is not expected to decline during the coming decades. In particular, wind energy represents one of the most dominant renewable energy sources. Due to the fact that wind energy has become such an important source of electric power, stable operation as well as high availability is needed. With the increasing amount of installations and the increase in power ratings of the single turbines, the task of maximizing the operational time of the wind turbine is more and more important in order to reduce the cost of energy. Under this scenario, early detection of fault conditions can be important to minimize the damage and therefore reduce the downtime of the wind turbine. It is therefore important to understand the behavior of the faulty component, both in order to limit the damage, but also to be able to more efficiently detect the fault. This work presents a system model of the electrical parts of a full power converter wind turbine with a permanent magnet synchronous machine as generators. The system model is modular based, where each sub-system can be replaced by a model of a faulty component in order to study its impact on a system level. Statistics has shown that faults occur in all of the different sub-systems of the turbine, not only in the electrical parts. However, generator faults are among the most common wind turbine component to fail, and in addition it is a costly and time consuming component to replace. This work presents an analytical model of a permanent magnet synchronous machine with a turn-to-turn fault, which is the most common winding fault. There are other works available in this subject, however there is little information regarding how to detect a turn-to-turn fault. The derived analytical model of the faulty machine is verified using a finite element model. In order to simulate the transition of the generator from healthy to faulty conditions, a flexible model of the permanent magnet synchronous machine is developed. The flexible model can also be used to emulate the evolvement of the fault by using a variable resistance in the fault loop of the investigated winding. The developed model can also be used to understand which quantities that needs to be monitored to be able to early detect a faulty condition in the machine. It is shown that turn-to-turn faults in the permanent magnet synchronous generator can be detected through monitoring of the harmonic content of the stator current in the rotating reference frame as the harmonic content slightly change when a turn-to-turn-fault arise.