Research On Wind Turbine Condition Monitoring Based On The Combination Of SCADA Data And Vibration Signals

Relevant data shows that common faults are mainly distributed in the three marjor areas: main bearing,gearbox and generator.So it's necessary to carry out condition monitoring to reduce downtime of the wind turbine.Now,the SCADA system is widely used to record the state information to identify the faults.But it is more easily influenced by outside factors such as environment,which will reduce the reliability of the monitoring results.Meanwhile,vibration signal as an important patameter reflecting the rotating parts makes vibration analysis became a reliable and effective method to condition monitoring and fault diagnosis of large rotating machines.The paper collect vibration information of the high incidence of failure parts and combine with the SCADA monitoring data.Then use nonlinear state estimation technique(NSET)fault diagnosis method improved by mahalanobis distance to make condition prediction model to implement condition monitoring and anomaly identify.Comparing with the traditional method based on fixed interval to constructure process memory matrix,the new algorithm is simple and effective.At the same time use the gearbox condition monitoring as an example to give evidence.Then from the whole state monitor into the components which is closely related with the gearbox.And use the forecast of these variables to find the cause of abnormal in order to achieve abnormal auxiliary judgment.Then establish an evaluation model using fuzzy comprehensive evaluation method to achieve condition assessment for wind turbine.For purpose of the engineering application a man-machine interaction interface of the wind turbine abnormal identify system has been designed,which based on the mixed programming technique of LabVIEW and MATLAB.The system can realize the remote control and has positive effect on the safety production and economic operation...