CAD模型引导的涡轮叶片密集测量数据分割  

A CAD Model Directing Method for Turbine Blade Density Measurement Data Segmentation

作　　者：程云勇[1] 张定华[1] 张顺利[1] 卜昆[1] 

机构地区：[1]西北工业大学现代设计与集成制造技术教育部重点实验室,西安710072

出　　处：《中国机械工程》2009年第18期2214-2218,共5页China Mechanical Engineering

基　　金：国家863高技术研究发展计划资助项目(2006AA04Z144);国家科技支撑计划资助项目(2006BAF04B02)

摘　　要：使用光学照相设备与锥束工业CT等数字化测量手段可以为涡轮叶片的几何形状检测提供密集的测量数据。为了有效进行涡轮叶片各结构特征的几何尺寸分析,需要对涡轮叶片的测量数据进行分割,把具有复杂拓扑结构的空心涡轮叶片的测量数据分割成简单拓扑结构对应的测量数据子集。针对已有测量数据分割方法的不足,提出一种CAD模型引导的涡轮叶片测量数据分割方法。该方法以叶片CAD模型的线框表示为先验知识,与基于边的测量数据分割方法相结合,实现涡轮叶片密集测量数据的可靠分割。使用仿真与实测的叶片数据对分割方法进行了验证。Density measurement data of turbine blade could be obtained based on advanced digital measurement devices, such as ATOS（advanced topometric sensor） and cone beam industrial CT. In order to do the geometric shape analysis for each feature of the turbine blade effectively, the density measurement data of the turbine blade with complex topological structures should be segmented into relatively simple sub datasets according to blade structure. As the existing methods for data segmentation were still unable to get the required results reliably and automatically, a CAD model directing method for turbine blade density measurement data segmentation was presented. Combining the edge based data segmentation method, a line frame representation of the turbine blade CAD model was adopted as the preknowledge for directing the data segmentation process. The discussed method was verified by blade simulation data and real measurement data.

关 键 词：涡轮叶片 锥束CT 先验知识 数据分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]