1. An improved void fraction prediction model for gas-liquid two-phase flows in pipeline-riser systems.
- Author
-
Chang, Yingjie, Xu, Qiang, Zou, Suifeng, Zhao, Xiangyuan, Wu, Quanhong, Wang, Yechun, Thévenin, Dominique, and Guo, Liejin
- Subjects
- *
TWO-phase flow , *POROSITY , *ELECTRICAL capacitance tomography , *PREDICTION models , *STANDARD deviations , *STRATIFIED flow - Abstract
• The phase interface in the upstream horizontal section was reconstructed using ECVT. • Upstream interface drop and slugging are induced by the flow in the riser. • Lockhart-Martinelli model is found to be best and is used for improved correlation. • An improved void fraction prediction model is proposed for pipeline-riser systems. • This improved model leads to considerably reduced prediction errors. This paper introduces the application of void fraction prediction models (25 models total) from a simple horizontal pipe to a complex, real pipeline. The flow mechanism in the pipeline-riser system is explained based on the upstream phase interface and differential pressure in the downstream riser. The transition boundaries of "stratified to intermittent flow" and "stratified smooth to stratified wavy flow" in the horizontal section were predicted well by the models based on simple pipes. Electrical capacitance volume tomography was used to obtain the upstream void fraction just before the downstream gas-liquid eruption, α 1 , or during multiple cycles, α 2. The difference between α 1 and α 2 are ranging between −10 % and 5 %. Finally, an improved correlation is proposed based on the model found to be best (Lockhart and Martinelli, with error values (root mean square error (RMSE) and mean absolute percentage error (MAPE)) of about 12 % and 18 %); the corresponding error levels (RMSE and MAPE) amount roughly to 5 % and 10 %, respectively. The improved model also delivers a good prediction for α 1 when separating in two flow regimes. Overall, the finally proposed correlation reads and is found superior in terms of accuracy to all previously published correlations. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF