Partial least square method-based wind turbine generator gearbox state monitoring method and control device

Abstract

The invention provides a partial least square method-based wind turbine generator gearbox state monitoring method, and belongs to the field of wind turbine generator state monitoring. The wind turbine generator gearbox state monitoring method comprises the steps of obtaining historical operation data of a wind turbine generator, selecting SCADA parameters influencing the oil temperature of a wind turbine generator gearbox, and performing data preprocessing; using the selected SCADA parameters for constructing variables representing the nonlinear relation, taking the selected SCADA parameters and the constructed variables representing the nonlinear relation as model input variables, and building a gearbox oil temperature model established based on a partial least square method; according to the residual distribution of the predicted value and the actual value of the gearbox oil temperature, setting a threshold value of the residual, and establishing a gearbox state monitoring model; and monitoring the state of the gearbox of the wind turbine generator through the gearbox state monitoring model. According to the embodiment of the invention, the goodness of fit and the prediction precision of the gearbox oil temperature model are improved, so that the state of the gearbox can be better monitored.

Description

technical field [0001] The invention relates to the field of state monitoring of wind turbines, in particular to a method for monitoring the state of a gearbox of a wind turbine based on a partial least square method and a control device. Background technique [0002] In recent years, as wind power has entered the era of grid parity, more and more attention has been paid to improving the operation level and economic benefits of wind turbines. Monitoring the status of wind turbines and giving early warnings in the early stages of performance degradation can effectively reduce unit failures and power loss, and improve the operating efficiency of wind farms. Especially for wind farms in remote areas or offshore wind farms, due to the inconvenience of operation and maintenance, it is particularly important to improve the level of unit condition monitoring. As an important part of the wind turbine, the gearbox is operated under variable working conditions and loads for a long ti...

AI Technical Summary

Problems solved by technology

However, there are two problems in the application process of the commonly used multiple linear regression method: 1) due to the multiple linear correlations between SCADA data, there may be problems such as data redundancy and overfitting; 2) the multiple linear regression method cannot Simulate nonlinear relationships between variables

Images