Research On Fault Diagnosis And Fault-tolerant Control Of Detection System For DFIG Converter

At present, wind power technology is a promising and relatively maturetechnology among the renewable energy application. Global wind power industry stepsinto a stable growth phase after years of rapid development. Meanwhile, the scale ofoffshore wind farms has been expanding to be a new round of growth in wing powerfield. China is still in infancy in this field and it has a large potential for development.With the increase of grid-connection capacity of wind power and the development ofoffshore wind farms, requirements for the wind power technology are higher. The paperresearched on the fault diagnosis of current detection system and speed detection systemand the reconfiguration of control system aimed at improving the safety and reliabilityof the doubly fed induction generator (DFIG) which is used most widely in the windpower system.The control system of DFIG is generally carried out in a continuous time domain,while it is achieved by programming in the discrete time domain in practice. The paperderived the mathematical model of DFIG based on the T equivalent circuit and deducedthe continuous-time state equations and the discrete-time state equations with stator androtor currents as the state variable. And discretization error was analyzed. The systemstability was analyzed from the perspective of DFIG poles location and rotor currentcontroller. The observability of the system was also analyzed to provide theoreticalsupport for the current observer design.Reliable current detection system is essential for the stable operation of DFIGconverter. This paper established stator and rotor current reduced order state observerfor DFIG and proposed a fault diagnosis scheme for current detection system based oncurrent prediction. It enabled the observed current values track the actual one rapidlyand stably by pole-placement method. The observed currents realized redundancy ofcurrent signals and could be used to implement system reconfiguration after failure. Thefault diagnosis method proposed by the paper could quickly diagnose the fault signal,theoretically only taking one control cycle at least. And the fault diagnosis method wasnot influenced by power alteration and grid failure. Research results shown that thepower converter could maintain network operation by fault detection and fault-tolerantcontrol at partial detection signal condition. And the reliability of wind power systemcould be improved effectively.The reliable operation of speed detection system is also important. On the basis ofrotor current reduced order observer, the paper utilized model reference adaptivemethod to build speed adaptive observer. The speed observer could estimate the actual generator speed correctly and rapidly. Meanwhile a method of fault diagnosis for speeddetection system was proposed which could quickly diagnose the fault signal. Theconfiguration was realized with the observed values of speed and rotor position angle.Research results showed that it could achieve controllable operation after speeddetection system failure and improved the reliability of the wind power system.Since the current and speed observers needs large amount of computation,controller based on TMS320F28335for DFIG converter was designed and produced.Network operation of the double-fed wind turbine was realized and the current andspeed observers and fault diagnosis proposals were verified with the controller.