A Fault Diagnosis Method of Wind Turbine Based on Convolutional Neural Network

Abstract

The invention discloses a method for diagnosing faults of wind power generators based on convolutional neural networks. Step 1 acquires the bearing vibration signals during operation of the generator from the data acquisition and monitoring control system of the wind farm at a certain time interval, and uses them as the Processed data; step 2 transforms the bearing vibration signal to obtain a cepstrum image, which is used as the sample data of the wind turbine generator fault diagnosis model; step 3 calculates the bearing characteristics of the wind turbine based on the cepstrum image obtained in step 2 Frequency, the natural frequency of the radial bending vibration of the bearing ring made of steel in a free state and the natural frequency of the rolling element made of steel balls, and then determine the fault defect type of the wind turbine; step 4 corresponds to the fault defect type of the wind turbine The cepstrum image is encoded and labeled as the data set of the model; step 5 builds a random forest-based convolutional neural network generator fault diagnosis model for wind turbine generators.

Description

technical field [0001] The invention relates to a generator fault diagnosis method, in particular to a wind turbine generator fault diagnosis method based on a convolutional neural network. Background technique [0002] At present, the vibration signal of the wind turbine bearing is collected by the vibration sensor. The vibration sensor is installed at the position of the bearing housing of the driving end and the non-driving end of the generator to obtain the vibration acceleration signal or speed signal of the bearing. After data processing, the bearing can be obtained. The characteristic frequency parameters of the operation have different characteristics due to the vibrations excited by faults of different types, properties, causes and locations, which are specifically manifested in the differences in frequency components, amplitudes, phase differences, waveforms, and energy distributions. According to this, People can identify faults from vibration signals. [0003] T...

AI Technical Summary

Problems solved by technology

Since the fault cepstrum of the wind turbine generator has a larger feature difference than the cepstrum during normal operation, and has certain regularity, the deep learning neural network visual recognition method can be applied to the fault diagnosis of the wind turbine generator to replace the cumbersome Boring manual diagnosis solves the problem that ordinary quality inspectors lack prior knowledge, thus greatly reducing labor costs

Images