Reliable prediction of spatially developing hydrodynamic slug flow is critical for various purposes, such as separator design and pipeline integrity management. Slug tracking is a state-of-the-art model for this application. Its ability to run on coarse grids results in manageable computation time for industrial scale uses.
For years, the predictive capability of slug tracking had been limited by the arbitrary slug initiation process. Recently, we proposed a mechanistic slug initiation model to overcome this limitation. However, this model was still incomplete, as it considered only slug unit cell initiation in local separated (stratified or annular) flow. There are also cases where the unit cells can be formed by the initiation of Taylor bubbles; henceforth called bubbles, in local bubbly flow. Examples of these are gas injection or vaporization in a liquid stream with little initial gas.
In the research referred in this paper, we extend the mechanistic slug initiation model to account for bubble initiation. We apply a minimum slip analysis to determine which initiation mechanism, from separated or from bubbly flow regime, is taking place at any given time and location.
The bubble initiation rarely occurs in the laboratory because gas injection and/or appreciable vaporization are typically absent from the experimental design. When the bubble initiation model is activated, it practically has no effect on laboratory-scale simulation results. In some field cases, it is found that slug length is vastly overpredicted if the slug formation caused by the transition from bubbly to slug flow regime is not accounted for. For these cases, the bubble initiation model improves slug tracking performance. In field cases where the slug tracking performance is already good, the bubble initiation model has no significant effect. We conclude that activating the bubble initiation model can improve the slug tracking performance when physically needed and does not produce detrimental effects otherwise.