基于激光器标定的涡轮叶片点云配准方法

Point Cloud Alignment Method for Turbine Blades Based on Laser Calibration

为解决由涡轮叶片铸造形面误差和装夹误差而导致的气膜冷却孔加工精度较低的问题,提出一种通过获取精确测量点云,以点云配准获得叶片位姿偏差的识别点激光传感器安装误差的标定方法。首先,基于机床运动学模型和标定平面几何约束条件建立方程组,求解激光传感器发射点位置偏移和激光束方向偏差;其次,通过测量磨砂标准球对标定结果的准确性进行验证,发现相比于标准直径的测量误差为0.36%;最后,经标定后的激光传感器测量得到叶片形面实测点云,点云数据归一化后,用Go-ICP算法与设计模型进行点云配准,获得从测量点云到模型点云的旋转平移矩阵,配准结果显示两组点云之间的欧式距离平方误差总和为0.0121 mm,经仿真验证平均误差率为1.97%。

In order to solve the problem of low machining accuracy of film cooling holes caused by turbine blade casting shaped surface error and clamping errors,a calibration method to identify installation errors of a point laser sensor is proposed to obtain an accurate measurement point cloud. A point cloud alignment is then carried out to obtain the blade posture deviation. Firstly,an equation group is established based on the kinematic model of a machine tool and the geometric constraints of a calibration plane to solve for the positional offset of the laser sensor emission point and the directional deviation of the laser beam. Secondly,the accuracy of the calibration results is verified by measuring a frosted standard ball with a measurement error of 0.36% as compared to the standard diameter. Finally,the laser sensor is calibrated to obtain an actual point cloud of a blade surface. After normalization of the point cloud data,a Go-ICP algorithm is used to align the point cloud with the design model to obtain a matrix describing the rotation and translation from the measured point cloud to the model point cloud. The alignment results showed that the sum of the squared errors of the Euclidean distance between the two-point clouds is 0.0121 mm,with an average error rate of 1.97% verified by simulation.