【摘要】 流型识别在工业工程中有着极其重要的意义,一直是两相流研究中的重要课题。本文以水平管道中的气/液两相流为研究对象,应用电阻层析成像系统获取不同流型条件下的测量数据,经预处理后提取小波尺度能量百分比与突起段时间作为特征值,并采用统计学习理论中支持向量机(SVM)方法对特征向量进行分析、识别流型。实验数据的分析结果表明,该方法具有较好的识别率。更多还原

【Abstract】 The flow regime recognition is of most important in industri ous engineering. Therefore, it is always a key subject of two-phase flow. While studying gas-liquid two-phase flow in horizontal pipe in this paper, both the distribution of energy in various regions of time and frequency using wavelet-based technique, and the duration of humps are extracted from the pre-processed measured data which is obtained in different flow regimes using electrical resistance tomography (ERT) system; finally feature vectors of the flow regime was recognized by employing support vector machine (SVM).更多还原