基于支持向量机的气液两相流流型识别新方法

A Novel Identification Method of Gas-liquid Two-phase Flow Regimes Based on Support Vector Machine

为准确识别两相流型,提出了基于小波包多尺度信息熵和支持向量机的流型识别方法.利用小波包变换对采集到的水平管空气-水两相流压差波动信号进行3层小波包分解,得到8个不同频带的信号,提取各频带信号的小波包多尺度信息熵作为流型的特征向量,运用支持向量机进行训练并识别流型.结果表明:与BP神经网络相比,采用支持向量机进行流型识别可以获得更高的识别率,表明该方法是有效、可行的.

In order to identify the two-phase flow regime, a novel method of flow regime identification based on support vector machine and wavelet packet multi-scale information entropy was proposed. The collected differential pressure fluctuation signals were decomposed by the three- layer wavelet packets, and the eight signals of different frequency bands were obtained. The wavelet packet multi-scale information entropy of different frequency bands signals were taken as feature vectors of flow regimes. The feature vector were put into support vector machine and trained to realize the flow regime identification. The result showed that the SVM had higher identification accuracy than BP neural network. The results proved that the method is efficient and feasible.