Some Design Considerations For Two-Phase Flow In Pipes

Publisher Website
Google Scholar
Abstract
An engineer faced with designing a piping system for simultaneous gas and liquid flow is confronted with a multitude of correlations and flow models from the literature. Many of these methods have now been critically evaluated and compared by the authors in previous papers. using the approximately 15,000 observations contained in The University of Calgary Multiphase Pipe Flow Data Bank. In this paper, a strategy for designing gas-liquid pipelines, based on the results of these comparative studies, is proposed. The proposed method is examined in detail for a typical pipeline and is compared using data from an actual pipeline with some method commonly used by industry. Introduction: IN THE PAST 25 years, literally thousands of papers dealing with various aspects of the simultaneous flow of gases and liquids in pipes have been published. Many of these have contained methods of correlating and predicting holdup and pressure drop, generally along with some data that support the proposed method to a greater or lesser degree. Unfortunately, most of these methods are specific to a particular flow regime, gas-liquid system or flow apparatus, and have little application under a broader range of conditions. A number of methods have been proposed as having general applicability, but even these can usually be shown to be subject to large errors for certain flow regimes or fluid systems. The result is that the pipeline designer is confronted with an array of relatively untested or obscure prediction methods, usually with little or no rational basis for making a choice. The tendency has been for him to select a simple and familiar method, and use it under almost any circumstances. In an effort to provide such a rational basis, a study was undertaken at The University of Calgary to critically evaluate many of the known methods against as much data as possible. The objective was to produce an over-all design strategy which utilizes the correlation or model most appropriate under the given conditions specified by the design problem. The first step in this project was to establish a comprehensive file of experimental data. The result of this stage was The University of Calgary Multiphase Pipe Flow Data Bank(1,2). The data bank, which now contains over 15,000 observations (including those contained in the AGA/API Data Bank developed at the University of Houston), has been used as the basis for all of the subsequent critical evaluations. Over-all Approach to the Formulation of Design Strategy: It is generally recognized that no single correlation or mechanistic model can be expected to yield accurate predictions for holdup and pressure drop over the entire range of possible gas and liquid flow rates. Because many of the basic flow patterns which are observed also represent fundamentally different flow mechanisms, the flow patterns can be used as convenient subdivisions within which to test specific correlations. A design strategy that is consistent with the above is one in which a flow pattern is first predicted on the basis of the given design parameters (fluids, flow rates, pipe characteristics, etc.)

This publication has 0 references indexed in Scilit: