Advances in Slug Flow Characterization for Horizontal and Slightly Inclined Pipelines

Abstract

SPE Members Abstract Current methods for characterizing slug flow behavior are examined in light of experience with their use in multiphase pipeline design. Isograms of slug flow properties, displayed on a flow pattern map, are used to examine slug flow characteristics and identify conditions under which current prediction methods perform well. Using this technique, the equilibrium film simplification is shown to not be justified for a large portion of the slug flow region. Slug length isograms on a flow pattern portion of the slug flow region. Slug length isograms on a flow pattern map are shown to be particularly useful for visualizing slug length behavior. New methods are developed for predicting the existence of slug flow, modeling slug length behavior. New methods are developed for predicting the existence of slug flow, modeling slug length behavior predicting the existence of slug flow, modeling slug length behavior in inclined flowlines, and for predicting pigged slug length under slug flow conditions. Methodology is developed to determine slug flow characteristics and is applied to an example pipeline connecting the Eldfisk and Ekofisk fields in the Norwegian North Sea. Introduction The ability to accurately characterize slug flow behavior for long, large diameter pipelines is essential for the successful operation of many satellite offshore fields. These are often connected to existing facilities through full well-stream multiphase flowlines and must be designed to not upset the end-line production facilities. While the slug flow pattern is complex and not fully understood, a number of recent advances have improved our ability to predict important parameters such as slug velocity, liquid holdup, and length. This work will review existing slug flow models and correlations and present new methods and techniques useful in the characterization of slug flow. Current methods of estimating slug velocity, liquid holdup, and length are presented first and their accuracy examined. Data collected by Kouba are used as the basis for many of the comparisons. Kouba conducted 53 tests covering a wide range of gas and liquid rates within the slug flow pattern. The test facility consisted of a 3 in. diameter, 1370 ft long, pattern. The test facility consisted of a 3 in. diameter, 1370 ft long, kerosine-air, horizontal pipeline. The liquid holdup of the slug body and bubble/film region were measured using dynamically calibrated capacitance sensors. Average slug and bubble lengths were also obtained. For this work these data were re-examined to obtain the film holdup one pipe diameter in front of the slugs, thus yielding the reading most closely approximating equilibrium conditions. Table 1 shows a summary of these test results. To better understand the wide variations observed in slug length, the concept of plotting slug length isograms on a flow pattern map is introduced. This analysis technique is shown to be useful in visualizing the behavior of others lug flow characteristics as well. The accuracy of existing methods for predicting the occurrence of slug flow is discussed and a method of defining the transition boundaries directly from the slug flow pattern model is developed. This method provides for greater consistency between the flow pattern model and the provides for greater consistency between the flow pattern model and the flow pattern transition model. The effect of terrain on slug flow characteristics is discussed and a model is developed for predicting slug length changes due to inclination, for the case of global steady-state flow. A model is developed to describe pigged slug growth for the special case when the pig is preceded by slug flow. Techniques for the characterization of slug flow are then applied to an example pipeline connecting the Eldfisk and Ekofisk fields in the Norwegian sector of the North Sea. SLUG FLOW CHARACTERIZATION The Hubbard and Dukler model serves as the basis for our current understanding of the slug flow pattern. Figure 1 illustrates the various interactions observed during slug flow. The slug front is moving at a translational velocity of VSF and the gas bubble front is moving at a velocity of VBF. Assuming homogeneous flow in the slug body, the fluid velocity in the slug, v,, is essentially the mixture velocity, vm = VSL + Vsg. The liquid holdup in the slug is denoted by Hs. As the slug moves, it scoops liquid from the preceding film zone into the slug body and accelerates it to vs liquid is also shed from the tail as it moves through the pipeline. P. 125