基于高速摄影传感器的气液两相流型分层模糊识别  

Gas-liquid two-phase flow pattern hierarchical fuzzy identification based on high-speed photography sensor

作　　者：常佃康 薛婷[1,2] 张旖婷 

机构地区：[1]天津大学电气与自动化工程学院,天津300072 [2]天津市过程控制检测与控制重点实验室,天津300072

出　　处：《传感器与微系统》2016年第11期58-60,共3页Transducer and Microsystem Technologies

基　　金：天津市自然科学基金资助项目(12JCQNJC02200)

摘　　要：气液两相流流型的客观和智能识别对两相流其它参数测量具有重要意义。用高速摄像机作为检测传感器,获取内径为50 mm的绝热垂直管段内气液两相流流型图像,提取流型灰度图像的均值、标准差及二值图像的最大物体面积、宽度和高度五个特征向量,采用分层模糊推理方法实现泡状流、块状流、塞状流和乳沫状流的流型识别。对于131幅实验图像,该方法的识别正确率平均达到93.13%。Objective and intelligent identification of gas-liquid two phase flow pattern have important meaning for other parameters measurement. High-speed camera is used as detecting sensor to capture two-phase flow patterns image in adiabatic vertical section whose inner diameter is 50 mm,extract five eigenvectors of mean and standard deviation of gray image,and area,width,height of the max object of binary image; use hierarchical fuzzy reasoning method to identify bubble flow,slug flow,plug flow and churn flow patterns. For 131 experimental images,the average identification accuracy of this method is 93. 13 %.

关 键 词：气液两相流 流型识别 图像特征 模糊推理 

分 类 号：TN247[电子电信—物理电子学]