叶片自由曲面原位测量方法与实验研究

发布时间：2018-08-22 15:27

【摘要】：叶片是航空发动机、透平机械、大型螺旋桨等动力设备的重要构件,其设计制造质量决定了动力设备的整体工作性能。叶片面形多为变截面、扭曲度高的自由曲面,设计制造难度大。叶片进行机械加工后,通过磨削抛光工序修整其表面质量,而现有磨抛加工过程与测量过程分离,难以实现叶片高效高质量的光整加工。针对上述问题,研究叶片自由曲面测量理论,结合叶片磨抛加工设备,研发原位测量系统,对提升叶片精整加工的效率和精度具有重要应用价值。结合自主研发的叶片磨抛加工机床和已有检测技术,建立非接触式原位测量系统。该系统将测量过程带入叶片的加工过程,在不改变叶片装夹位置的情况下,完成对叶片型面数据的采集,减少定位误差。提出密集点法、特征点法和极坐标旋转法三种原位测量方案,以适应不同加工阶段对效率和精度的要求。开发控制软件,设计界面程序,实现对叶片零件的自动化测量。对叶片型面进行逐点测量涉及测点规划问题,测点分布对于测量速度、匹配及测量精度都有很重要的影响。研究了两种基于曲面的测点自适应规划算法—质心系法和矩形细分法。质心系法应用质心平衡原理,将各点的曲率指标类比于质量,通过不断迭代确定测点位置;矩形细分法通过设定阈值实现局部的以平面代替曲面。改进算法使其适用于无解析式的自由曲面,对叶片型面进行测点规划仿真分析,为兼顾精度和效率的特征点测量法奠定基础。分析原位测量系统的误差影响因素,结果表明,运动平台误差、工件表面颜色、PLS的非线性误差和倾角误差的影响较为显著。研究测量系统各运动部件的位姿关系,构建测量系统运动误差模型,根据散射光场在空间分布的朗伯定律推导PLS倾角误差模型,根据这两项误差模型提出误差补偿策略。开展原位测量实验,实测对象为某航天器发动机叶片的叶背和叶盆表面。应用高精度扫描仪获取叶片点云数据,据此对待测面形进行测点自适应规划。设计四组对比实验,每组实验保证测点数目相同,分别使用特征点法和等步长法进行测量。将实测数据进行插值并反求成面,匹配叶片理论模型,比较不同测量方式的精度。实验结果表明,建立的原位测量系统性能良好,可以满足叶片型面检测的要求。对比实验结果表明,同样测量方式下,实测点数目越多,反求曲面的精度越高;而在同等测点数目下,使用特征点法测量的数据反求曲面精度更高,证明了基于测点自适应规划的特征点测量法的优越性。
[Abstract]:Blade is an important component of aero-engine, turbine machinery, large propeller and other power equipment, its design and manufacture quality determines the overall performance of power equipment. The one-sided shape of the blade is mostly a free-form surface with variable cross section and high distortion, so it is difficult to design and manufacture. After machining, the surface quality of the blade is repaired by grinding and polishing process, but the existing grinding and polishing process is separated from the measuring process, so it is difficult to realize the high efficiency and high quality finishing of the blade. In view of the above problems, the theory of blade free-form surface measurement, combined with vane grinding and polishing equipment, and the development of in-situ measurement system have important application value in improving the efficiency and precision of blade finishing. A non-contact in situ measurement system was established by combining the developed blade grinding and polishing machine tool and existing testing technology. The system introduces the measurement process into the machining process of the blade, and completes the data collection of the blade profile without changing the clamping position of the blade, thus reducing the positioning error. In order to meet the requirements of efficiency and precision in different processing stages, three in-situ measurement schemes, i.e., dense point method, characteristic point method and polar coordinate rotation method, are proposed. Develop control software, design interface program, to achieve automatic measurement of blade parts. The point planning problem is involved in the point-by-point measurement of the blade profile, and the distribution of the measuring points has an important effect on the measuring speed, matching and measuring accuracy. Two adaptive programming algorithms for measuring points based on curved surface, the centroid system method and the rectangular subdivision method, are studied in this paper. The centroid system method uses the centroid balance principle to compare the curvature index of each point to the mass and determines the measuring point position by iteration. The rectangular subdivision method realizes the local replacement of the surface with a plane by setting a threshold. The improved algorithm makes it suitable for free-form surface without analytic expression. The point planning simulation analysis of blade profile is carried out, which lays a foundation for the feature point measurement method which takes account of both accuracy and efficiency. The error influencing factors of in-situ measurement system are analyzed. The results show that the error of moving platform, the nonlinear error of workpiece color pls and the error of inclination angle are more significant. In this paper, the position and pose relations of the moving components of the measurement system are studied, and the motion error model of the measurement system is constructed. According to the Lambert's law of the spatial distribution of scattered light field, the PLS inclination error model is deduced, and the error compensation strategy is proposed according to the two error models. An in situ measurement experiment was carried out to measure the blade back and the blade basin surface of a spacecraft engine blade. The point cloud data of the blade are obtained by using a high precision scanner, and the point adaptive planning is carried out according to the measured surface shape. Design four groups of comparative experiments, each group of experiments to ensure the same number of points, using the characteristic point method and equal step size method to measure. The measured data are interpolated and inversely calculated to match the blade theoretical model and compare the accuracy of different measurement methods. The experimental results show that the system has good performance and can meet the requirements of blade profile detection. The experimental results show that the higher the number of measured points, the higher the accuracy of the reverse surface, and the higher the accuracy of the reverse surface measured by the method of characteristic points is under the same number of measuring points. The advantages of feature point measurement based on adaptive programming are proved.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG580.6;TG806

【相似文献】