基于双目摄影测量的风力机叶片振动测量方法

VIBRATION MEASUREMENT METHOD OF WIND TURBINE BLADE BASED ON BINOCULAR PHOTOGRAM METRY

提出了一种带径向畸变和切向畸变的成像模型,利用Levenberg-Marquardt算法进行非线性优化参数标定的双目摄影振动测量方法,并对某型风力机叶片进行了模态频率辨识。两台摄像机同步采集振动图像序列并用自适应动态阈值对图像进行目标分割,以编码点区域灰度梯度跳跃最快的方向作为边缘,进一步用最小二乘法拟合边缘,用灰度平方加权质心法获取动态连续图像序列中编码点的精确亚像素中心坐标。每一个编码点类似于微小的"位移传感器",能够反映叶片时域振动响应。最终,从时域响应数据中获取叶片固有模态频率,与有限元法和激光多普勒测振仪分析结果比对,相对误差都在5%以下,验证了双目摄影测量方法的可靠性和有效性。

binocular photogrammetry vibration measurement method for considering the radial and tangential distortion parameters calibration by using the Levenberg-Marquardt algorithm based nonlinear optimization was proposed.As an example,the modal frequency identified of a wind turbine blade.The vibration images sequence of the blade was acquired synchronously by two cameras and the target was segmented by adaptive binary dynamic threshold method,the direction with fastest grayscale gradient of the coding point as the edge,the circle curve was fitted by the least squares and iterative method,and the precise suppixel central coordinates of the coding point in the dynamic consecutive images were obtained by the gray square centroid method.Each coding point is similar to a small"displacement transducer"that reflects the blade’s time domain vibration response.Finally,the blade’s natural modal frequency was obtained from the time domains response,and compared with the results of the finite element and Laser Doppler Vibrometry,the relative error of vibration frequencies less than 5%,which verified the reliability and effectiveness of binocular photogrammetry method.