Flow regime detection of mini-pipe gas-liquid two-phase flow based on PCA and SVM

A simple flow regime detection method of gas-liquid two-phase flow is presented, used in mini-pipes. First, the capacitance signals are obtained, and some simple statistics eigenvalues, such as mean value, standard deviation, total energy, average amplitude, zero-crossing rate and probability density function, are analyzed and used to identify the flow regimes. The principal component analysis (PCA) method is used to reduce the dimensionality of data sets and eliminate multi-collinearity of variables. PCA of four pairs of capacitance sensor can also reduce the noise influence. Then support vector machine (SVM) method is used to detect flow regime. In experiment, bubble flow, stratified flow, slug flow and annular flow are observed in the pipes with inner diameter of 3.1mm. Four electrical capacitance sensors are installed. Based SVM theory, a set of binary classifier is constructed, and flow regime can be detected successfully. The results show that the presented method is effective, and can improve the accuracy of flow regime identification.