Add to Chrome
Add to Firefox
Add to EdgeSupport Vector Machine Application for Multiphase Flow Pattern Prediction
Jun 12, 2018
In this paper a data analytical approach featuring support vector machines (SVM) is employed to train a predictive model over an experimentaldataset, which consists of the most relevant studies for two-phase flow pattern prediction. The database for this study consists of flow patterns or flow regimes in gas-liquid two-phase flow. The term flow pattern refers to the geometrical configuration of the gas and liquid phases in the pipe. When gas and liquid flow simultaneously in a pipe, the two phases can distribute themselves in a variety of flow configurations. Gas-liquid two-phase flow occurs ubiquitously in various major industrial fields: petroleum, chemical, nuclear, and geothermal industries. The flow configurations differ from each other in the spatial distribution of the interface, resulting in different flow characteristics. Experimental results obtained by applying the presented methodology to different combinations of flow patterns demonstrate that the proposed approach is state-of-the-art alternatives by achieving 97% correct classification. The results suggest machine learning could be used as an effective tool for automatic detection and classification of gas-liquid flow patterns.
