Application of Least Squares Support Vector Machine to the Void-fraction Measurement of Two-Phase Flow

摘要

Based on least squares support vector machines (LS-SVM) techniques, a new method for the void-fraction measurement of two-phase flow was proposed. The sensor of Electrical capacitance tomography (ECT) system was used for obtaining the capacitance measurement values. These data reflect the void-fraction change of two-phase flow. The raw measurement capacitance values are given to regression LS-SVM to measure the void-fraction of two-phase flow. The new method implements direct void-fraction measurement of oil/gas two-phase flow without the need of time-consuming image reconstruction. The experimental results show that for 94.8% of the total testing data, the errors of void-fraction measurement are less than 5% and for the other 5.2%, the errors of void fraction are between 5% and 9%.