摘要
提出一种深度学习检测方法CNN-BiGRU.采用CNN自适应地学习变量之间存在的关联特征,利用BiGRU对时间序列的敏感性,对风机叶片故障分类.对某风电场的SCADA数据进行增强、切片、标准化等预处理,实验结果表明:CNN-BiGRU分类模型能有效对叶片结冰故障进行准确检测,在时间效率和检测准确率方面较其他深度学习模型效果更好.
A detection method of deep learning named CNN-BiGRU was proposed.It used both the correlation characteristics between the adaptive learning variables of the CNN and the sensitivity to time-series of the BiGRU to classify the wind turbine blade faults.The SCADA data of a wind farm were pre-processed by enhancing,slicing and standardized processing.The results show that the CNN-BiGRU classification model can effectively detect the icing failure of blade,and it is better than other deep learning models in time efficiency and detection rate.
作者
王永平
张蕾
张晓琳
徐立
韩朋
张飞
WANG Yongping;ZHANG Lei;ZHANG Xiaolin;XU Li;HAN Peng;ZHANG Fei(Information Engineering School,Inner Mongolia University of Science and Technology,Baotou 014010,China;Department of Computer Science and Technology,Baotou Medical College of Inner Mongolia University of Science and Technology,Baotou 014010,China;School of Renewable and Clean Energy,North China Electric Power University,Beijing 102206,China)
出处
《内蒙古科技大学学报》
CAS
2022年第2期173-179,共7页
Journal of Inner Mongolia University of Science and Technology
基金
国家自然科学基金资助项目(61562065)
国家重点研发计划基金资助项目(2017YFE0109000)
内蒙古自治区自然科学基金资助项目(2019MS06036)
内蒙古科技大学创新基金项目(2014QDL046)。
关键词
风机叶片
故障检测
卷积神经网络
双向门控循环单元
wind turbine blades
fault diagnosis
convolutional neural network
bidirectional gated recurrent unit
