Fault Warning and Reliability Analysis of Wind Turbine Failure Based on Data-driven

The global wind power industry is growing rapidly and wind energy has become an integral part of the power system. In order to improve the stability of wind turbine generators (WTGs), it is essential to accurately predict and warn of their operational status. Today, most WTGs are equipped with monitoring and data acquisition systems that can collect and record comprehensive operational status information, providing a reliable source of data for assessing the unit's operational status. Effectively analysing the data and providing early warnings of faults is a key issue in the study of the stable operation of wind farms. In this study, historical and real-time wind turbine data, wind farm operation and maintenance data and meteorological data are used as samples. Firstly, a Weibull linear model is constructed to analyse the performance of the wind turbine and to map the power bands for reliable operation of the wind turbine. A wind turbine wind power prediction model was developed using a BP neural network prediction model, and then the predicted wind speed was used to obtain the power of the wind turbine, and the predicted power was compared with the normal power operating band. If the predicted power is within the power range, the wind turbine is highly reliable. If the predicted power is beyond the power range, the reliability is low, at this time, a warning signal should be issued and the monitoring of the turbine operation status should be strengthened.