A simple and accurate model to predict pressure drop in vertical gas wells

Predicting pressure drop in gas wells is one of the most important issues for designing deliquification technologies and optimizing gas production. Current pressure-drop prediction methods can be divided into analytical and empirical models. Although various models have been proposed to predict pressure drop in wellbores, no model has offered stable performance over wide liquid and gas flowrate ranges. Especially for gas wells with low liquid - gas ratios, very few models have been developed specifically. In this study, a simple and accurate three-point model is developed for predicting liquid holdup in wellbores. The model is based on the liquid holdups and gas velocities at the annular - churn, churn - slug and slug - bubble boundaries and is comprehensive for calculating the liquid holdup in vertical wellbores at different liquid/gas flowrates and pressures. The proposed method was verified with 39 collected laboratory and 182 field measured data points from the literature, which have wide parameter ranges. For comparison, some widely used models in the petroleum industry were also evaluated. In validation with the laboratory data with pipe diameters of 50.8 and 101.6 mm, although the new model has an average percentage error (E3) of 220.1% in the churn flow, ranking fourth of all the evaluated models, the other five indicators in the churn flow and six indicators in the annular flow are lowest of all these evaluated models, resulting in relative performance factors (RPFs) of 0.2 and 0, respectively. In the field measured data, gas and liquid productions range between 0.34 x 104-77.6 x 104 m3/d and 1-347.8 m3/d, tubinghead pressure ranges from 0.7 to 84.7 MPa, which can cover most gas wells. The validation demonstrates that the new model still has good performance with five indicators being lowest. The RPF of the new pressure-drop model in the field measured data ranges is 0.29, still lowest of all these evaluated models, showing better performance than these widely used models in petroleum industry for vertical gas wells. This indicates that the new model can accurately predict pressure drop under different conditions in vertical gas wells.