基于偏二叉树支持向量机的ERT系统流型识别研究  

Study on Flow Regime Identification of Electrical Resistance Tomography System Based on Binary Tree Support Vector Machine

作　　者：张华[1] 曲海旭[1] ZHANG Hua QU Hai - xu(The City College of Jilin Architectural And Civil Engineering Institute, Changchun, Jilin 130011, China)

机构地区：[1]吉林建筑大学城建学院,吉林长春130011

出　　处：《通化师范学院学报》2016年第12期63-65,共3页Journal of Tonghua Normal University

摘　　要：两相流在工业中应用广泛,它具有非常复杂的流动性和随机性,准确识别流型是两相流参数准确测量的基础.该文首先用小波包分解方法提取ERT系统测量的压差波动信号特征,然后构建偏二叉树支持向量机多类分类模型,最后向分类模型中输入特征数据进行流型识别.实验结果表明偏二叉树支持向量机多类分类算法较大提高了流型识别的准确度,是一种有效的流型识别方法.Two - phase fluid, which has complex flow characteristic and randomness, has been applied widely in the industrial production. The accurate identification of flow regime is the foundation of accurate measurement on two -phase flow＇s parameter. Firstly, the feature of differential pressure fluctuation signal, which is measured by electrical resistance tomography system, is extracted by wavelet packet analysis. Then, the multiclass model of binary SVM is constructed. Finally, the data about extracted feature is inputted into the multi - class SVM of binary. The experimental results show that the accuracy of two - phase flow regime identification has been improved remarkably. Binary SVM is an effective method for regime identification.

关 键 词：电阻层析成像 流型识别 小波包 偏二叉树支持向量机 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]