航空发动机直纹叶片的线形特征点云构造算法

针对航空领域形状复杂、高精度的直纹叶片曲面的逆向设计问题,借助该类叶片曲面的几何特征,提出一种线形特征点云的构造算法.在对叶片散乱点云边界进行快速识别的基础上,采用反投影法提取直纹母线矢量,利用退火算法求解边界数据点配准的多目标优化问题,以实现边界点云数据的非刚性配准.根据配准结果构造等分点,进而得到规则有序的线形特征点云,达到快速生成高质量叶片曲面的目的.本算法在处理直纹叶片点云数据方面具有较好的效果.对直纹叶片的曲面重建具有一定的理论意义和应用价值,通过实验验证了该算法的可行性.

Algorithm for constructing the line-shape feature point clouds of aero-engine ruled blades

By means of the surfaces' geometry features, an algorithm for constructing the line-shape feature point cloud was proposed for reverse design of ruled blade surface requiring high accuracy and complex shape in aviation field. On the basis of the quick recognition for boundary of blades' scattered point clouds, the ruled generatrix vector was extracted with the aid of reverse projection method. The annealing algorithm was applied to match boundary point clouds non-rigidly by means of solving multi-objective optimization problem of boundary point clouds. The equal diversion points were constructed to get inerratic-orderly line-shape feature point clouds which were beneficial to generate high quality blade surfaces quickly. The proposed algorithm has great effect in dealing with point clouds of ruled blades and has a certain theoretical significance and application value on surface reconstruction of point cloud of ruled blades surface. The feasibility of proposed algorithm is verified by experiments.