Tuning a multi-fluid model for gas lift simulations in wells

A multi-fluid formulation based on the k–ɛ turbulence closure is used for modeling bubbly flow in vertical pipes. Each bubble-size group is considered as a separate dispersed phase. The current k–ɛ based multi-fluid models suffer from the problem of large overprediction of void fraction peak and lead to satisfactory results in limited ranges. In this study, first, we establish a model for gas lift simulations in wells. For this purpose, we propose new modified lift and wall force coefficients by tuning the model with many experimental databases. As shown here, the tuned multi-fluid model is able to predict flows in all the three bubbly flow subregimes, namely the wall-peak, core-peak, and transition subregimes, with reasonable accuracy. The predictions by the tuned model are compared with other numerical simulations, as well. Finally, the tuned model is used to simulate gas-lift problems in oil wells and the performance of eight empirical and semi-empirical correlations for predicting pressure drop is investigated, carefully. It is observed that, there is a correlation (i.e. Guet, S., Ooms, G., Oliemans, R.V.A., Mudde, R.F., 2004. Bubble size effect on low liquid input drift-flux parameters. Chem. Eng. Sci. 59, 3315–3329) which predicts gravitational pressure drop in very good agreement with the tuned multi-fluid model in all ranges of bubble diameters, including the wall-peak and core-peak regimes.

► A multi-fluid model is established by tuning with many experimental databases.
► The results of the tuned model are compared with other numerical simulations.
► Eight empirical pressure drop correlations are compared in oil well simulations.