Two-phase flow feature extraction and regime identification in horizontal pipe

Abstract

The correct identification of two-phase flow regime is the basis for the accuracy measurement of other flow parameters in two-phase flow measurement. Electrical capacitance tomography (ECT) is a new measurement technology. It is often used to identify two-phase/multi-phase flow regime and investigate the distribution of solids. The support vector machine (SVM) is a machine-learning algorithm based on the statistical learning theory (SLT), which has desirable classification ability with fewer training samples. So, it provides a new approach for flow regime identification. The capacitance measurement data obtained from ECT system contain flow regime information. Feature parameters, which can reflect flow regime, were extracted. Using these feature parameters and SVM method, simulation experiments were done for 4 typical flow regimes. The results showed that this method is fast in speed and can identify these 4 flow regimes correctly, and prove this method is efficient and feasible.