Accurate prediction of pressure gradient in gas wells is the theoretical basis of gas well performance analysis, production optimization and deliquification technologies design. Experiment is the best access to characterize the flow behavior of gas wells. For low-pressure experimental investigation and gas wells, the most difference is the pressure (gas density), which could lead to totally different flow behavior. Dimensionless numbers are often used in the flow pattern maps to account for the flow similarities at different conditions, which means liquid holdup in the high pressure can be also predicted at low pressure conditions if we choose proper dimensionless numbers for pressure scaling up. However, no studies have focused on this point before. Besides, gas wells have high GLR, most empirical models were intended to developed for oil wells, which have greater weight in low GLR, decreasing the accuracy in gas wells.
In order to predict the pressure gradient in horizontal gas wells, an experimental investigation of gas-water flow has been conducted. The experimental test matrix was designed to cover all the flow patterns. The experiment was conducted in a 5-m long pipe. The liquid holdup and pressure gradient were measured. Subsequently, the effect of gas velocity, liquid velocity, pipe diameter, and inclined angle on liquid holdup was analyzed. Then the dimensionless numbers proposed in the literature have been investigated and analyzed for pressure scaling up. Finally, a comprehensive model was established, which is developed for prediction pressure drop in gas wells. Some field and experimental data were provided to evaluate the new model. The results show that the Duns-Ros dimensionless number was not proper for pressure scaling up while the Hewitt-Robert Number performs best. Compared to widely used pressure gradient models with field data, the new model with Hewitt-Robert Number performed best, which shows that it is capable of dealing with prediction of pressure gradient in gas wells.