An Evaluation of Recent "Mechanistic" Models of Multiphase Flow for Predicting Pressure Drops in Oil and Gas Wells

Abstract

ABSTRACT The reliable calculation of tubing pressure drops in oil and gas wells is important for the most cost effective design of well completions. None of the traditional multiphase flow correlations works well across the full range of conditions encountered in oil and gas fields. Consequently, two of the recently published "mechanistic" models, one by Ansari, the other by Hasan & Kabir, were evaluated. The performance of these methods was compared against traditional correlations in three ways: The predicted against measured pressure drops were compared for stable flow conditions using 246 data sets collected from 8 producing fields, including a gas and gas-condensate field. None of these data were available to the developers of any of the multiphase flow models evaluated. Suitable methods should reliably predict the "lift curve minima". This determines when a well may need to be "kicked off', artificially lifted or recompleted. The multiphase flow model must not contain discontinuities or be subject to convergence problems. No single traditional correlation method gives good results in both oil and gas wells. In fact, most of the traditional methods which work reasonably in oil wells give very poor predictions for gas wells. Hasan & Kabir's mechanistic method was generally found to be no better than the traditional correlation methods. However, the Ansari mechanistic model gave consistently reasonable performance. Although it did not give the most accurate results in every field, it gave reasonable results across the complete range of fields studied. The Ansari method also gives a reliable prediction of the lift curve minima. Areas in which it needs improvement were identified. By comparison the best of the traditional methods, the Hagedorn & Brown correlation, gave good results for stable flow conditions in oil wells, but it does not correctly predict the lift curve minima. A field example shows how this can lead to erroneous conclusions.