随着海上风电的蓬勃发展,运维工作越来越成为突出问题。风电叶片作为风电大尺寸关键构件,其运维对机组至关重要。本文针对海上风机叶片人工运维检测存在的高风险、低效率和低精度等问题,提出了一种基于改进YOLOv5x(You Only Look Once v...随着海上风电的蓬勃发展,运维工作越来越成为突出问题。风电叶片作为风电大尺寸关键构件,其运维对机组至关重要。本文针对海上风机叶片人工运维检测存在的高风险、低效率和低精度等问题,提出了一种基于改进YOLOv5x(You Only Look Once version 5x)算法的海上风机叶片缺陷机器视觉检测系统。该方法引入了卷积块注意力机制(Convolutional Block Attention Module,CBAM),以增强神经网络对输入特征的感知能力,使用智慧交并比(Wise Intersection over Union,WIoU)作为损失函数,减少人工标注数据的误差,提高目标检测的准确性。基于海上风机叶片缺陷数据对模型进行训练,将训练好的模型封装成海上风机叶片机器视觉识别系统。试验结果显示,改进后的YOLOv5x算法,相比于原有的YOLOv5x,平均精度均值(mean Average Precision,mAP)提高了4.71%,准确率(Precision)提高了7.48%,且能满足实时性需求。展开更多
The topic of offshore wind energy is attracting more and more attention as the energy crisis heightens.The blades are the key components of offshore wind turbines,and their dynamic characteristics directly determine t...The topic of offshore wind energy is attracting more and more attention as the energy crisis heightens.The blades are the key components of offshore wind turbines,and their dynamic characteristics directly determine the effectiveness of offshore wind turbines.With different rotating speeds and blade length,the rotating blades generate various centrifugal stiffening effects.To directly analyze the centrifugal stiffening effect of blades,the Rayleigh energy method (REM) was used to derive the natural frequency equation of the blade,including the centrifugal stiffening effect and the axial force calculation formula.The axial force planes and the first to third order natural frequency planes which vary with the rotating speed and length were calculated in three-dimensional coordinates.The centrifugal stiffening coefficient was introduced to quantitatively study the relationship between the centrifugal stiffening degree and the rotating speed,and then the fundamental frequency correction formula was built based on the rotating speed and the blade length.The analysis results show that the calculation results of the fundamental frequency correction formula agree with the theoretical calculation results.The error of calculation results between them is less than 0.5%.展开更多
文摘随着海上风电的蓬勃发展,运维工作越来越成为突出问题。风电叶片作为风电大尺寸关键构件,其运维对机组至关重要。本文针对海上风机叶片人工运维检测存在的高风险、低效率和低精度等问题,提出了一种基于改进YOLOv5x(You Only Look Once version 5x)算法的海上风机叶片缺陷机器视觉检测系统。该方法引入了卷积块注意力机制(Convolutional Block Attention Module,CBAM),以增强神经网络对输入特征的感知能力,使用智慧交并比(Wise Intersection over Union,WIoU)作为损失函数,减少人工标注数据的误差,提高目标检测的准确性。基于海上风机叶片缺陷数据对模型进行训练,将训练好的模型封装成海上风机叶片机器视觉识别系统。试验结果显示,改进后的YOLOv5x算法,相比于原有的YOLOv5x,平均精度均值(mean Average Precision,mAP)提高了4.71%,准确率(Precision)提高了7.48%,且能满足实时性需求。
基金Supported by the National Natural Science Foundation of China under Grant No.50708015the foundation of State Key Laboratory of Coastal and Offshore Engineering, Dalian University of Technology
文摘The topic of offshore wind energy is attracting more and more attention as the energy crisis heightens.The blades are the key components of offshore wind turbines,and their dynamic characteristics directly determine the effectiveness of offshore wind turbines.With different rotating speeds and blade length,the rotating blades generate various centrifugal stiffening effects.To directly analyze the centrifugal stiffening effect of blades,the Rayleigh energy method (REM) was used to derive the natural frequency equation of the blade,including the centrifugal stiffening effect and the axial force calculation formula.The axial force planes and the first to third order natural frequency planes which vary with the rotating speed and length were calculated in three-dimensional coordinates.The centrifugal stiffening coefficient was introduced to quantitatively study the relationship between the centrifugal stiffening degree and the rotating speed,and then the fundamental frequency correction formula was built based on the rotating speed and the blade length.The analysis results show that the calculation results of the fundamental frequency correction formula agree with the theoretical calculation results.The error of calculation results between them is less than 0.5%.
