Engineering Design of Gas-Condensate Pipelines With a Compositional Hydrodynamic Model

Abstract

Summary. Gas condensation in pipelines designed to transport natural gas is common. The radical difference in the engineering design required for this system, compared with that required of "dry gas" pipelines, makes the problem of significant interest to the gas industry. Because the point and quantity of condensation is not usually known a priori, any attempt to develop a predictive capability for such a system must have an inherent means of providing this information. This requires a good coupling of the gas- phase-behavior model with the appropriate hydrodynamic model. This work attempts to develop such a model. With a two-parameter equation of state (EOS) to describe the phase behavior of the natural-gas system, a multiphase hydrodynamic model developed from fundamental fluid dynamics is used to describe the hydrodynamic behavior of the resulting two phases. The model which consists of a system of nonlinear algebraic and ordinary differential equations (ODE's), was solved numerically. Output from the model solution includes quantity of condensate at any point in the pipeline, pressure drop, and other hydrodynamic variables. The model can predict the various engineering parameters of interest in the design of such pipelines and could be used for feasibility studies and for optimal location of fluid-handling equipment. Introduction Natural gas supplies more than 50% of the residential and commercial heating needs of the U.S. Industrially, natural gas is used both as fuel and feed stock, but before being put to its end use, natural gas must be found, produced, gathered, processed, and transported to the consumer. Although producing and gathering methods are relatively standard in any field operation, the degree of processing may vary widely. Depending on the scale of available processing, the transported natural gas may be dry (if the necessary processing equipment-e.g., separators and scrubbers-are available) or wet (if little or no processing is done). The problem of adequate but optimal design of gas pipelines, with the inherent probability of condensation, becomes accentuated for offshore gas fields, for instance, where space is limited and processing sometimes is kept to a minimum. Such a pipeline design must allow for multiphase flow arising from condensation. The importance of a good transportation system cannot be overstated. It is the link between the producer and the consumer, and its cost may dictate the final price paid by the consumer for gas. Typically, the transportation system may consist of pipelines a few hundred meters to hundreds of kilometers in length and may range in diameter from about 7 cm [2.76 in.] to ~ 1 m [ ~ 3 ft]. Along the system, slug catchers may remove any liquids that may accumulate in the line and compressor stations may boost the gas pressure. Because of the distances between producers and consumers, some pipelines traverse both undulating terrain and regions of varying temperatures. The combined effect of the imposed temperature profile and the pressure drop in the pipeline can even cause a dry gas to experience some condensation within the pipeline. This condensation is possible because of the phase behavior of the gas. Retrograde condensation is quite common in such systems. Reservoir conditions at which natural gas is found typically are above the upper dewpoint curve. It is reasonable to say that if a gas is allowed to flow and is subjected to a decreasing temperature profile, the dewpoint can be crossed and retrograde condensation can occur. The extent of retrograde condensation depends on the gas composition. While lean gases may exhibit a very small or nonexistent retrograde region, the region may be quite significant when heavier hydrocarbons (C5+) are present in the gas. Thus, a gas-condensate system is composition-dependent and must be subjected to compositional modeling to predict the distribution and composition of the phases. Recent works emphasized this need. The basic engineering parameters in the design of a multiphase pipeline system are the pressure drop, liquid loading, and phase velocities. Various models are available to predict pressure drop, liquid holdup, and phase velocities in pipelines, but most rely on correlations developed from data obtained under limited conditions. The application of these models outside the operating range of the data bases used in their development invariably leads to erroneous results. Gregory and Aziz, Fayed and Otten, and Cawkwell and Charles compared the predictions of these models and found that they varied widely and did not match the measured data. This investigation uses a fundamental hydrodynamic model developed from the basic laws of mass and momentum conservation. This model is coupled with a phase-behavior package, resulting in a compositional hydrodynamic model. Within the phase-behavior model, phase properties are predicted with the Peng-Robinson EOS. This study is restricted to horizontal pipelines operating under high pressure. Survey of Available Methods Adewumi and Mucharam characterized the available methods into two categories: the single-phase-safety-factor (SPSF) and the steady-empirical-two-phase (SETP) approaches. In the early years of wet gas pipeline design, SPSF approaches, which basically used single-phase gas equations with the application of an "efficiency" factor to account for the liquid phase, were common. This efficiency factor was actually an empirical number whose value varied because of the inadequacy of the equations to describe two-phase flow. The single-phase gas equations used most were the Panhandle A and the Modified Panhandle equations. Several researchers pointed out that this approach should be used only for first estimates. SETP approaches were first applied in the late 1950's and are still in use today. These approaches use empirical correlations relating flow pattern, liquid holdup, and two-phase pressure drop in pipelines. Many of these correlations consider the flow as homogeneous and treat properties as representing a mixture rather than the individual phases present. The Lockhart-Martinelli, Flanigan, Dukler, and Beggs-Brill correlations are the most popular (1973). Various investigators compared the results of several of these correlations with measured data. The comparisons prove that the correlations were inaccurate in describing multiphase flow in pipelines, especially those with condensation. Calculated pressure drops varied greatly from the measured data and from each other. Hydrodynamic Approach Problem Definition. This paper presents a compositional hydrodynamic model that describes the hydrodynamic behavior of gas/ gas-condensate flow in pipelines. The flow is treated as a steady-state two-phase system in which gas is condensing and/or vaporizing at some section of the pipeline. One of the most difficult features of this system is that the condensing/vaporizing section must be identified as part of the solution to the problem. SPEPE P. 381⁁