DT-PTV离散相伪矢量的概率择优式剔除算法

Elimination of Discrete Phase Outliers in DT-PTV by Preferential Probability

提出一种基于三角嵌铺的粒子追踪测速(DT-PTV)算法的剔除离散相伪矢量新方法——概率择优式剔除法,引入了剔除冗余匹配的计算模块和扩大匹配信息量的"反算"概念,基于DT-PTV的"DT三角-粒子双层匹配"特性以及上述模块的多种组合,总结出适用于离散相流场的去伪多途径运算.通过对若干自定义流场进行多途径去伪并比较匹配准确率后得出,DT三角层面的信息最大化积累至粒子层面的最优选择是适用于DT-PTV离散流场测速的最有效去伪途径,在随机粒子场较低空间率(187.215)时用"三角层面有反算但无模块运算+粒子层面择优"途径计算,仍然能保持92%左右的匹配准确率,故将此途径命名为概率择优式剔除法.利用DT-PTV结合概率择优式剔除进行大气边界层风洞中的跃移层测速以及单颗跃移轨迹的提取,结果良好.

A new method called elimination by preferential probability （EPP） to remove discrete phase outliers in Delaunay tessellation-particle tracking velocimetry（DT-PTV）was proposed. In the present method, the modules to eliminate redundant matches and the inverse calculation to enlarge matching information were adopted and a multiple-way in outliers elimination used in discrete phase flow field was developed on thebasis of the double-layer matching pattern in DT-PTV and the combination modes of modules and inverse calculation. From the application in self-defined flow fields and comparison of accuracy rate, matching information maximizing in DT triangle level followed by most-suitable-matches selection in particle level is an optimum way of outlier elimination in discrete phase with an accuracy rate of about 92% in a low spatial rate （187. 215）. In addition, the DT.PTV combined with EPP was used for velocity measurement of sand phase in saltation layer of wind-sand flow, and the results of single-grain paths abstracted can meet the requirement.