Method for early warning and diagnosis of temperature of main bearing of wind turbine generator

Abstract

The invention relates to the field of new energy wind power generation systems, in particular to a method for early warning and diagnosis of the temperature of a main bearing of a wind turbine generator. The method comprises the following steps that firstly, wind power plant monitoring data is acquired; secondly, parameters related to the temperature of the main bearing of the wind turbine generator are obtained; thirdly, a normal temperature model of the main bearing of the wind turbine generator is established; fourthly, the theoretical value of the real-time normal temperature of the main bearing of the wind turbine generator is calculated, wherein the real-time values of the parameters related to the temperature of the main bearing of the wind turbine generator in the second step are selected from the real-time data collected in the first step, the real-time values of the related parameters are input into a neural network trained in the third step, and the normal temperature valueof the main bearing of the wind turbine generator is generated; fifthly, whether the real-time temperature of the main bearing of the wind turbine generator is normal or not is judged. By means of themethod, faults of the main bearing of the wind turbine generator are effectively judged in advance, an extra sensor does not need to be installed, and the diagnosis precision and the diagnosis time advance are remarkably improved.

Description

technical field [0001] The invention relates to the field of new energy wind power generation systems, in particular to a temperature early warning and diagnosis method for a main bearing of a wind turbine. Background technique [0002] The wind turbine main bearing fault early warning method is mainly applied to a fault diagnosis based on data mining. It is an important fault early warning method. Due to the complexity of the wind turbine installation environment, once the wind turbine fails, it will lead to maintenance difficulties. . Therefore, the fault diagnosis technology for fans has become a difficult problem at this stage. In practice, the main bearing fault diagnosis methods for wind turbines can be divided into data-based, signal-based and model-based methods. Usually, model-based fault diagnosis methods require a complex physical or mathematical model to test the fault. Signal-based Fault diagnosis technology needs to install additional various sensors on the f...

AI Technical Summary

Problems solved by technology

Due to the correlation between the SCADA parameters of wind farms, the method of using parameter correlation to select the input parameters of the neural network has the problem of repeated use of parameters and data redundancy when the input parameters are highly correlated.

However, the input parameters of the neural network are selected through the subjective experience method. Since there are many parameters affecting the fan components, the selected parameters are inaccurate, resulting in low efficiency of the neural network, too few selected parameters, and insufficient accuracy.

[0004] In the actual application system, the data collected by the SCADA system is only used to judge the operating status of the fan and whether the fan has failed, resulting in a large amount of collected data that cannot be used

Moreover, the data obtained by many current SCADA systems may not be able to reflect the current state of the fan. Because these problems still exist, most SCADA technologies at this stage only pursue the control of the fan by the SCADA system, but tend to ignore the data collected by the SCADA system. use of data

Images