A Fault Diagnosis Method of Wind Turbine Based on Polarization Maintaining Optical Fiber Measurement and Neural Network Classification

Abstract

The invention relates to a fault diagnosis method for wind power generators based on polarization-maintaining optical fiber measurement and neural network classification. The polarization-maintaining optical fiber measurement system is introduced to obtain the current signal of the wind-driven generator. It has small volume, light weight, simple insulation structure, large dynamic range, No magnetic saturation, no explosion hazard, no influence on normal power transmission, can directly measure the diagnostic signal during the actual use of the wind turbine; and in the specific fault diagnosis process, through the convolutional neural network with recurrent neural and frequency domain calculations to obtain the three types of characteristics of the polarization angle time series, and use a single hidden layer neural network to realize state recognition. The polarization angle can reflect the current sequence, and then reflect the stator and rotor faults of the wind turbine. This method can provide more abundant fault information of the wind turbine.

Description

technical field [0001] The invention relates to a fault diagnosis method of a wind power generator based on polarization-maintaining optical fiber measurement and neural network classification, and belongs to the technical field of wind power generators. Background technique [0002] In order to solve the problems of energy depletion and air pollution, wind power, as a kind of clean energy, is being commercialized on a large scale. The long-term maintenance of a large number of wind turbines that operate around the clock has become the focus of attention. Traditional fault diagnosis technology has obtained signals The main methods include: using MEMS accelerometer sensor to pick up mechanical vibration signal, using microphone to pick up acoustic signal, and using ammeter to pick up current signal after disconnecting the load. Among them, mechanical vibration signal pickup requires additional equipment, which is often resisted by equipment manufacturers. The acoustic signal...

AI Technical Summary

Problems solved by technology

The acoustic signal pickup method lacks the diagnostic information of the motor stator and rotor, and the current signal needs to be disconnected from the load, which will affect the normal operation of the wind turbine, and it is difficult to apply it in the operation of large-scale wind power equipment.

On the other hand, the feature extraction of fault signals is mostly based on frequency domain information. Although the traditional signal processing method is convenient, it may ignore the fault information hidden in the time domain of the waveform, resulting in insufficient fault diagnosis or state identification information.

Images