Abstract
Abstract Problems peculiar to the design of multiphase (gas-liquid) pipelines are discussed using a multiphase-flow pipeline design program- A number of design methods currently used by the petroleum industry are compared in detail, with partieu1u"r reference to pipelines transporting natural gas – condensate mixtures. It is demonstrated that large differences in "predicted pressure drop may be obtained with different methods. INTRODUCTION IN SPITE OF THE FACT that many principles of single phase pipe flow also govern the pipe flow of gas liquid mixtures, in general, the design problem for the latter is unique. Because of complex interactions between the fluid phases, neither a rigorous analytical approach nor conventional single-phase flow correlations can usually be applied with much success. In this paper, some of the problems peculiar to two-phase flow are identified and pressure-drop predictions by a number of design methods currently used in the petroleum industry are compared for a typical gas-condensate pipeline. Design Considerations For Two-Phase Flow A number of characteristics of two-phase flow pose special problems for the pipeline designer that are not encountered with single-phase flow. These are discussed in detail in the book by Govier and Aziz(1) and will only be dealt with briefly here. FLOW PATTERN Depending on the relative gas and liquid flow rates, two-phase flow may exist as anything from a continuous gas phase with finely dispersed liquid droplets to a continuous liquid phase with finely dispersed gas bubbles. A number of well-defined flow patterns exist which result, in some cases, from the domination of buoyancy forces and, in others, from the domination of shear forces. It has been demonstrated by Mandhane et al.(4) and Gregory et al.(2) that the accuracy of most of the available friction-pressure-drop or in-situ-liquid- volume-fraction correlations can depend on the flow regime. It is thus important that the designer be able to predict the expected flow pattern from readily available information before performing his pressure-drop calculations. A number of correlations for doing this are available in the literature, including one recently proposed by Mandhane et al(J) LIQUID HOLDUP Almost invariably, the gas and liquid phases travel through a pipe at different velocities. This gives rise to a liquid holdup effect, because the fraction of the pipe volume occupied by the liquid phase under flowing conditions will be significantly different from the volume fraction of the liquid in the two-phase mixture entering the pipe. The total pressure losses for a two- phase mixture generally include a hydrostatic-bead contribution, which is calculated using a mixture density. Thus, the designer must be able to predict the in-situ liquid volume fraction under flowing conditions, as this is directly related to the mixture density. Numerous correlations are available in the literature for this purpose and, as noted above, these correlations may be more accurate for some flow regimes than others.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have