双谱核主元分析在气液两相流流型识别中的应用

Application of bispectrum KPCA in identification of gas-liquid two-phase flow regime

针对压差波动信号的非线性和非高斯特性,提出了一种基于高阶谱和核主元分析相结合的流型识别方法。通过对气液两相流压差波动信号的双谱分析,提取了不同流型下信号的非高斯特征,以双谱分析核主元数字特征提取流型的特征,最后利用最小二乘支持向量机对流型进行智能识别。实验结果表明,提取的核主元特征反映了两相流的流动状态,最小二乘支持向量机可以有效地识别水平管道内的4种典型流型,整体识别率达到95%,为流型识别提供了一种有效的方法。

Aimed at the nonlinear and non-Gaussian characteristics of differential pressure fluctuation signals of gas-liquid two-phase flow, a flow regime identification method based on bispectrum and kernel principal component analysis （KPCA） was proposed. The non-Gaussian characteristics of differential pressure fluctuation signals of different flow regimes were extracted by using the bispectrum method. Then the kernel principal component numerical characteristics technology based on bispectrum analysis was used to extract flow regime characteristics. At last, least square support vector machine （LS-SVM） was introduced to intelligently identify flow regimes. The test results showed that bispectrum analysis was an extremely powerful tool for the analysis of nonlinear and non-Gaussian signals, and the extracted KPCA characteristics could reflect the flow state of two-phase flow. The LS-SVM could accurately identify four typical flow regimes of gas-water two-phase flow in the horizontal pipe. The whole identification accuracy was 95% for flow regime identification by using the new effective method.