面向涡轮叶片变形测量的旋转适应性图像匹配方法研究

Research on Image Matching Method of Rotation Adaptability for Turbine Blade Deformation Measurement

研究了超临界二氧化碳涡轮发电系统叶片在工作情况下的变形测量方法。对于叶片变形测量可采用三维数字图像相关法技术路线(3D-DIC),但是需要解决由于叶片切向变形和相机触发误差引起的区域旋转图像的匹配问题(2D-DIC部分)。针对旋转图像的匹配问题,采用了一种旋转适应性图像匹配方法,对于检测点的初值搜索采用基于ORB特征的初值匹配方法,对于检测点的亚像素匹配采用基于反向组合高斯牛顿法(IC-GN)亚像素匹配方法。最终利用旋转-应变图像进行仿真,应变测量精度达到2%,验证了图像匹配方法对于涡轮叶片变形测量的适应性。

The method for measuring the blade deformation in supercritical carbon dioxide turbine power generation system under working condition was studied.Regarding the blade deformation measurement,a 3D digital image correlation method(3D-DIC)was employed and matching problem of region rotation image caused by blade tangential deformation and camera trigger error(2D-DIC part)was solved.Aiming at the rotation image matching,a rotation adaptive image matching method was adopted.The initial value matching method based on ORB feature was used for the initial value search of the detection point,including having the(inverse composition Gauss-Newton)IC-GN-based sub-pixel matching method used for the sub-pixel matching of the detection point.Finally,simulation with the rotation-strain image shows that,the strain measurement accuracy can reach 2%,which verifies the adaptability of image matching method in measuring the turbine blade deformation.