基于连续图像灰度序列混沌特性的油气水三相流型识别  

IDENTIFICATION METHOD OF OIL-AIR-WATER THREE-PHASE FLOW REGIME BASED ON CHAOTIC CHARACTERISTIC OF IMAGE GRAY SIGNALS

作　　者：周云龙[1] 李洪伟[1] 宋连状[1] 刘川[1] 

机构地区：[1]东北电力大学,吉林吉林132012

出　　处：《工程热物理学报》2009年第5期784-788,共5页Journal of Engineering Thermophysics

基　　金：吉林省科技发展项目(No.20040513)

摘　　要：以35号润滑油、空气和自来水为试验介质,应用高速摄像机对垂直上升管内的油气水三相流的六种典型流型进行了动态图像的拍摄。提取每一帧图像的灰度均值组成时间序列,对其进行混沌特性分析,提取序列的HURST指数,关联维以及分别以2和e为底的最大李亚普诺夫指数,组成特征向量,输入支持向量机进行流型分类。试验结果表明:连续图像的灰度时间序列的混沌特性能够对油气水三相流的典型流型进行很好的表征,结合支持向量机进行分类,识别率达到90%以上,为流型的在线识别提供了一种新方法。With 35# lubricants, air and water for the test medium, the application of high-speed cameras on the vertical up of the oil-gas-water three-phase flow in the six typical flow of the dynamic image photography. Extract the mean gray of each image and composition of the time series. Extract the chaotic characteristic, including the chaos associated, HURST index, the largest lyapunov index based on 2 and e. The characteristic vector were established. The SVM was trained using those vectors of flow regime intelligent identification was realized. The test results show that Vertical upward of the oil-gas-water three-phase flow in the dynamic gray-scale image signal fluctuations with the characteristic of chaos, it can make the three-phase flow typical of a good flow Characterization. Combined with SVM identification, the whole identification accuracy is above 90% for online flow regime identification of a new and effective method.

关 键 词：油气水三相流 图像灰度 李亚普诺夫指数 混沌关联维 HURST指数 支持向量机 

分 类 号：O359.1[理学—流体力学]