Wind-driven generator three-phase rotor current micro-fault diagnosis method

Abstract

The invention discloses a wind-driven generator three-phase rotor current micro-fault diagnosis method. The method comprises the first step of extracting fault information of three-phase rotor current to obtain a training data set, a confirmed data set and a testing data set; the second step of constructing a double-layer sparse bayes extreme learning machine model; the third step of creating a pairing multi-label classification method, and constructing a classifier set based on a micro-fault diagnosis algorithm model through the combination of a double-layer sparse bayes extreme learning machine; the fourth step of using the training data set to train the micro-fault diagnosis algorithm model, using the confirmed data set and a firefly algorithm to conduct dynamic optimum-seeking iteration, and finally determining an optimal parameter to complete the micro-fault diagnosis algorithm model; the fifth step of inputting the testing data set into the micro-fault diagnosis algorithm model to obtain a diagnosis result of a micro-fault. According to the wind-driven generator three-phase rotor current micro-fault diagnosis method, the wind-driven generator three-phase rotor current micro-fault can be diagnosed, and the wind-driven generator three-phase rotor current micro-fault has the advantages of being high in diagnosis efficiency, and high in diagnosis precision.

Description

technical field [0001] The invention relates to a current fault diagnosis method, in particular to a wind power generator three-phase rotor current micro fault diagnosis method. Background technique [0002] Wind turbines have been using planned maintenance for a long time. This maintenance method cannot fully and timely understand the operation status of the equipment; and after-the-fact maintenance is due to insufficient preparation in advance, resulting in long maintenance work and serious losses. In recent years, scholars at home and abroad have conducted extensive research on the significant fault diagnosis of wind turbines and achieved good results. It will be drowned in noise or large faults with obvious symptoms because of its small symptoms and weak signals. This makes it difficult for traditional diagnostic methods to deal with signals mixed with weak faults. Therefore, the existing diagnostic methods must be effectively improved or new methods established to achi...

AI Technical Summary

Problems solved by technology

[0002] Wind turbines have been using planned maintenance for a long time. This maintenance method cannot fully and timely understand the operation status of the equipment; and after-the-fact maintenance is due to insufficient preparation in advance, resulting in long maintenance work and serious losses.

In recent years, scholars at home and abroad have conducted extensive research on the significant fault diagnosis of wind turbines and achieved good results. Can be drowned in noise or large faults with obvious symptoms due to their small symptoms and weak signals

This makes it difficult for traditional diagnostic methods to deal with signals mixed with weak faults. Therefore, it is necessary to effectively improve existing diagnostic methods or establish new methods to achieve the diagnosis of micro faults.

Images