The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Exibições de texto completo
127
O artigo empregou um Alexnet, que é uma estrutura de aprendizado profundo, para diagnosticar automaticamente os danos nas superfícies das pás dos geradores de energia eólica. As imagens originais das superfícies das pás dos geradores de energia eólica foram capturadas por visões de máquina de um UAV (veículo aéreo não tripulado) de 4 rotores. Primeiramente, uma Alexnet de 8 camadas, incluindo totalmente 21 subcamadas funcionais, é construída e parametrizada. Em segundo lugar, o Alexnet foi treinado com 10000 imagens e depois testado com 6 imagens em 350 voltas. Por fim, a estatística dos testes de rede mostra que a precisão média do diagnóstico de danos pelo Alexnet é de cerca de 99.001%. Também treinamos e testamos uma rede neural tradicional BP (Back Propagation), que possui camada de entrada de 20 neurônios, camada oculta de 5 neurônios e camada de saída de 1 neurônio, com os mesmos dados de imagem. A precisão média do diagnóstico de danos da rede neural BP é 19.424% menor que a do Alexnet. O ponto mostra que é viável aplicar a aquisição de imagens UAV e o classificador de aprendizagem profunda para diagnosticar automaticamente os danos às pás das turbinas eólicas em serviço.
Xiao-Yi ZHAO
Inner Mongolia University of Technology,Inner Mongolia Key Laboratory of Mechanical and Electrical Control
Chao-Yi DONG
Inner Mongolia University of Technology,Inner Mongolia Key Laboratory of Mechanical and Electrical Control
Peng ZHOU
Inner Mongolia University of Technology,Inner Mongolia Key Laboratory of Mechanical and Electrical Control
Mei-Jia ZHU
Inner Mongolia University of Technology,Inner Mongolia Key Laboratory of Mechanical and Electrical Control
Jing-Wen REN
Inner Mongolia University of Technology,Inner Mongolia Key Laboratory of Mechanical and Electrical Control
Xiao-Yan CHEN
Inner Mongolia University of Technology,Inner Mongolia Key Laboratory of Mechanical and Electrical Control
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Xiao-Yi ZHAO, Chao-Yi DONG, Peng ZHOU, Mei-Jia ZHU, Jing-Wen REN, Xiao-Yan CHEN, "Detecting Surface Defects of Wind Tubine Blades Using an Alexnet Deep Learning Algorithm" in IEICE TRANSACTIONS on Fundamentals,
vol. E102-A, no. 12, pp. 1817-1824, December 2019, doi: 10.1587/transfun.E102.A.1817.
Abstract: The paper employed an Alexnet, which is a deep learning framework, to automatically diagnose the damages of wind power generator blade surfaces. The original images of wind power generator blade surfaces were captured by machine visions of a 4-rotor UAV (unmanned aerial vehicle). Firstly, an 8-layer Alexnet, totally including 21 functional sub-layers, is constructed and parameterized. Secondly, the Alexnet was trained with 10000 images and then was tested by 6-turn 350 images. Finally, the statistic of network tests shows that the average accuracy of damage diagnosis by Alexnet is about 99.001%. We also trained and tested a traditional BP (Back Propagation) neural network, which have 20-neuron input layer, 5-neuron hidden layer, and 1-neuron output layer, with the same image data. The average accuracy of damage diagnosis of BP neural network is 19.424% lower than that of Alexnet. The point shows that it is feasible to apply the UAV image acquisition and the deep learning classifier to diagnose the damages of wind turbine blades in service automatically.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.E102.A.1817/_p
Copiar
@ARTICLE{e102-a_12_1817,
author={Xiao-Yi ZHAO, Chao-Yi DONG, Peng ZHOU, Mei-Jia ZHU, Jing-Wen REN, Xiao-Yan CHEN, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Detecting Surface Defects of Wind Tubine Blades Using an Alexnet Deep Learning Algorithm},
year={2019},
volume={E102-A},
number={12},
pages={1817-1824},
abstract={The paper employed an Alexnet, which is a deep learning framework, to automatically diagnose the damages of wind power generator blade surfaces. The original images of wind power generator blade surfaces were captured by machine visions of a 4-rotor UAV (unmanned aerial vehicle). Firstly, an 8-layer Alexnet, totally including 21 functional sub-layers, is constructed and parameterized. Secondly, the Alexnet was trained with 10000 images and then was tested by 6-turn 350 images. Finally, the statistic of network tests shows that the average accuracy of damage diagnosis by Alexnet is about 99.001%. We also trained and tested a traditional BP (Back Propagation) neural network, which have 20-neuron input layer, 5-neuron hidden layer, and 1-neuron output layer, with the same image data. The average accuracy of damage diagnosis of BP neural network is 19.424% lower than that of Alexnet. The point shows that it is feasible to apply the UAV image acquisition and the deep learning classifier to diagnose the damages of wind turbine blades in service automatically.},
keywords={},
doi={10.1587/transfun.E102.A.1817},
ISSN={1745-1337},
month={December},}
Copiar
TY - JOUR
TI - Detecting Surface Defects of Wind Tubine Blades Using an Alexnet Deep Learning Algorithm
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 1817
EP - 1824
AU - Xiao-Yi ZHAO
AU - Chao-Yi DONG
AU - Peng ZHOU
AU - Mei-Jia ZHU
AU - Jing-Wen REN
AU - Xiao-Yan CHEN
PY - 2019
DO - 10.1587/transfun.E102.A.1817
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E102-A
IS - 12
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - December 2019
AB - The paper employed an Alexnet, which is a deep learning framework, to automatically diagnose the damages of wind power generator blade surfaces. The original images of wind power generator blade surfaces were captured by machine visions of a 4-rotor UAV (unmanned aerial vehicle). Firstly, an 8-layer Alexnet, totally including 21 functional sub-layers, is constructed and parameterized. Secondly, the Alexnet was trained with 10000 images and then was tested by 6-turn 350 images. Finally, the statistic of network tests shows that the average accuracy of damage diagnosis by Alexnet is about 99.001%. We also trained and tested a traditional BP (Back Propagation) neural network, which have 20-neuron input layer, 5-neuron hidden layer, and 1-neuron output layer, with the same image data. The average accuracy of damage diagnosis of BP neural network is 19.424% lower than that of Alexnet. The point shows that it is feasible to apply the UAV image acquisition and the deep learning classifier to diagnose the damages of wind turbine blades in service automatically.
ER -