双目立体摄影测量的风力机叶片振动频率分析
Vibration Frequency of Wind Turbine Blades Based on Binocular Stereo Photography
买买提明·艾尼 1刘慧洁 1古丽巴哈尔·托乎提 1吴志强 1亚库普江·艾合麦提2
作者信息
- 1. 新疆大学机械工程学院 乌鲁木齐,830047
- 2. 佰博机电科技有限公司 乌鲁木齐,830011
- 折叠
摘要
为了准确识别物体模态频率,降低测试成本及环境要求,提出了一种双目摄影测量辨识结构模态频率的方法,并对某型风力机叶片进行了测试.首先,通过两台摄像机从不同方位同步并连续采集振动结构及人工布设的多个标记点的图像;其次,采用高斯滤波进行图像去噪,经过Canny算子分割特征点边缘信息,用最小二乘法和迭代法拟合圆曲线,对标记点进行高精度的亚像素定位,计算摄像机交会范围内每一个测试点的时间响应函数与三维空间坐标信息;然后,使用KLT(Kanade-Lucas-Tomasi)算法实现特征点跟踪匹配,利用改进的快速傅里叶变换算法,将时域信号转换为频域信号,辨识了叶片的前4阶模态频率;最后,与加速传感器、激光多普勒测振仪(laser Doppler vibration,简称LDV)的对比结果表明,双目摄影测量获取的模态频率误差仅在5％范围内,验证了该方法的可靠性和有效性.
Abstract
A method for identifying structural modal frequencies based on binocular photogrammetry is pro-posed.Gaussian filtered image is used to denoise,the edge information is segmented by the Canny operator,and the least squares method is used to fit the circular curve to the sub-pixel with high precision.Positioning,ob-taining the coordinate information of the three-dimensional space of the measuring point,using the KLT(Kanade-Lucas-Tomasi)algorithm to achieve feature point tracking and matching,and using the improved fast Fourier transform algorithm to convert the time domain signal into the frequency domain signal,successfully identifying the blade's first 4th order modal frequency,and compared it with the results of the acceleration sensor and laser Doppler vibrometer(LDV).The modal frequency error obtained by binocular photogrammetry is only within 5％,verifying the dual reliability and effectiveness of stereoscopic photogrammetry methods.
关键词
双目摄影/模态频率/激光多普勒/加速度传感器/振动测量Key words
binocular photography/modal frequency/laser Doppler/accelerometer/vibration measurement引用本文复制引用
出版年
2024
NETL
NSTL
万方数据