Fault Diagnosis And Prediction Of Wind Turbine Based On Production Data

In the development of wind power matures today, the high maintenance costs of wind turbine equipment affect the economic efficiency of enterprises, thus reducing the unit downtime and equipment failure rate, improve the utilization of equipment has become a research hotspot. Based on the statistical analysis of the domestic wind turbine operating data, the key feature quantity is obtained, and the health status evaluation model of the wind turbine is established, which provides the basis for the real-time evaluation of the health status of the wind turbine. At the same time, through the operation of the data mining unit high frequency fault equipment, Fault diagnosis and prediction, in order to achieve the purpose of improving the availability of equipment. The main work is as follows:First, this paper analyzes the common faults and causes of wind turbine equipment,analyzes the common faults and causes of wind turbine equipment, analyzes the high frequency faults of wind turbines at home and abroad, excavates the running data of SCADA system of wind farms, and extracts the effective state features as wind turbine equipment health status evaluation, performance indicators; Using the association rule algorithm to analyze the alarm point of SCADA system, and to extract the effective leading alarm information.Secondly, the study of state feature quantity, combined with analytic hierarchy process and evaluation index system construction principle to rationally divide the feature quantity, get the wind turbine health status evaluation index system, and on this basis, based on the gray system theory variable weight fuzzy comprehensive evaluation algorithm, a healthy state evaluation model wind turbine. Through the verification, the results show that the wind turbine health status evaluation model averaged 3.273 hours in advance to monitor the unit in sub-health state.Thirdly, in this paper, the high frequency fault of wind turbine is diagnosed and predicted. The low temperature alarm of the inverter is about 76% of the alarm of the inverter, and the intrusion of the low temperature fault alarm information and the running data of the inverter is found. It is found that the low temperature fault alarm of the inverter only occurs in the unit shutdown state and is affected by the ambient temperature and wind speed Larger. Gearbox oil temperature abnormal alarm about 55%of the gear box alarm, through the production run data mining, found that gearbox oil temperature fault by the gearbox speed and its running time, the impact of ambient temperature. Based on the three influencing factors of gearbox oil temperature anomaly failure, this paper establishes the prediction model of abnormal oil temperature anomaly of gearbox. Compared with the existing prediction model which only considers the influencing factors of ambient temperature, this paper designs the prediction model Comprehensive, predictive results more accurate.