Research On Condition Monitoring And Fault Diagnosis Of MW Wind Turbine

With the amount of rapidly growing wind generator turbine systems in recent years, the wind turbine failure rate is also rising and the wind turbine reliability problems become more and more serious. The gear box installed at the top of the wind turbine tower has a short lifetime, and would need to be replaced in less than five years. Common faults are caused by other phenomena such as the base break down of the generators, short circuits, poor contact slip rings, rotor overheating and causing fire, rupture and resulting in the collapse of the tower, etc. Because the transmission chain of a large-scale wind turbine system is located very high up in the nacelle, which is hard to approach and works around the clock, the maintenance of wind generator turbine system will cause the reduction of energy production and a high cost of replacement parts. Condition monitoring equipment should, therefore, be installed in megawatt-class wind turbines by monitoring the operational status of machine parts to achieve effective fault diagnosis, and work out reasonable maintenanceplanning, so as to ensure wind turbine operation safety.The following tasks have been set out in this paper during the process of establishing the megawatt-class wind turbine condition monitoring and fault diagnosis system:Firstly, the vibration characteristics of the tower will influence wind turbine stability. The natural frequency of the tower should avoid the external vibration frequency, to eliminate wind turbine resonance. Therefore, coupling vibration models of wind turbines have been established, including the blade flap and the tower "fore-aft bending". Based on the deformation consistency of the rotor, the tower and the nacelle, The simultaneous equations is found by found the blade/rotor dynamics equation and nacelle/tower dynamics equation. The vibration modal of the system, have been deduced by the equation using decoupling analysis. The scope of the system resonance frequency and the influence to tower natural frequency caused by physical dimension of tower are theorically. This research can lay the foundation for condition monitoring of the wind turbine and optimization designs of the towers. According to the structural characteristics of wind turbine and operation mechanism, the condition monitoring projects of transmission chain and tower systems are determined. Sixteen vibration-shock sensors are used to collect the information of faults in the transmission chain. Two two-way vibration acceleration sensors are used to collect the fault information of the tower system. Generalized resonance theory and the resonant demodulation technology are used to pick up fault signals. The active diagnosis is adopted during fault diagnosis which is useful to draw the structural parameters of mechanical into the mathematical model reflecting fault rule, to get characteristic spectral lines of mechanical extract from mechanical parts of the spectrum characteristics, and distinguish fault by measured the amplitude of the characteristics spectrum.Using FEM models consist of hexahedral element to analyze dynamic characteristic of the tower, main shaft and coupler and obtains each component modal, stress distribution and displacement. They provide the basis for fault diagnosis of the condition monitoring system and alarm threshold setting.To achieve stable operation, a notch filter is added to the controller to filter out external frequency that equate with the tower and transmission chain natural frequency based on the calculation results in blade software and the result of condition monitoring. The results the synthetic judgment of the wind turbine operating state, will determine the wind turbine operating control and help to establish a reasonable maintenance plan.