On the Analysis of High Viscous Horizontal Slug Flow Characteristics by Changing Pipe Diameter

Abstract

The analysis of slug characteristics for high viscous horizontal flow was experimentally carried out by changing the pipe diameter. Slug flow is the general pattern of oil-gas two-phase flow in a horizontal pipe. Each slug characteristic plays an important role in predicting the pressure gradient and average liquid holdup, contributing to designing production systems. Most of the existing prediction models were developed to evaluate low and medium viscous flow, so an experimental verification is needed to predict the flow parameters of high viscous flow. Furthermore, previous studies for high viscous flow have only been conducted in 50.8-mm-ID (2-in.) pipes. The scalability of the observed behavior in larger diameters is still under consideration. 628 experimental tests were conducted in a 76.2-mm-ID (3-in.) horizontal pipe. Six different oil viscosities were considered. Among them, data sets with 587 cP, 181 cP, and 155 cP were used to analyze the pipe diameter effects compared with the 50.8-mm-ID (2-in.) pipe experimental data. Superficial liquid velocity varied from 0.02 m/s to 0.35 m/s and superficial gas velocity varied from 0.1 m/s to 3.6 m/s to match the experimental matrices of previous studies. Statistically calibrated two-wire type capacitance sensors were used to measure the liquid holdup. The pressure and pressure gradient were obtained using different transducers. Flow pattern, pressure gradient, average liquid holdup, slug liquid holdup, film liquid holdup, slug length, slug length distribution, and slug frequency were measured and analyzed. Not only data obtained from this study, but also previous data from 2-in. ID pipes were used to compare with the 3-in. ID pipes results. The pipe diameter and oil viscosity affect the flow transition boundary, while the pressure drop decreases and the average liquid holdup increases for larger pipe diameters. Moreover, the slug liquid holdup and slug frequency increase, and the film liquid holdup and slug length decrease as the pipe diameter increases. The experimental results were also used to evaluate different flow pattern maps, existing models and correlations for two-phase slug flow. Some degree of discrepancy was observed between experimental and predicted results, especially for predicting the pressure gradient. This indicates that the coupled effect of high viscosity with pipe diameter is yet to be completely understood and needs to be modified. The simplified Lockhart and Martinellis separated pressure gradient prediction model was suggested in this study by subtracting the accelerational pressure gradient from the original equation. This model presents lower absolute average relative errors than the original model and relatively acceptable average relative errors in the entire range of datasets, emphasizing the importance of separating liquid and gas Reynolds numbers in high viscous experimental data.