A 5‐equation, transient, hyperbolic, 1‐dimensional model for slug capturing in pipes. (1st June 2017)

• Record Type: Journal Article
• Authors: Ferrari, Marco; Bonzanini, Arianna; Poesio, Pietro
• Is Part Of: International journal for numerical methods in fluids. Volume 85:Number 6(2017)
• Page Start: 327
• Page End: 362
• Publication Date: 2017-06-01
• Subjects: 2‐phase flow -- 2‐fluid model -- slug capturing -- finite volume -- partial differential equations -- hyperbolic; Fluid dynamics -- Mathematics -- Periodicals; 532
• DOI: 10.1002/fld.4387
• Languages: English
• ISSNs: 0271-2091

Abstract

Summary: A novel numerical scheme for slug capturing in pipes using a 1‐dimensional transient hyperbolic 5‐equation 2‐fluid model is presented. Previous work has shown that 1‐dimensional 2‐fluid models are able to capture slug flow automatically. In this work, a similar approach is further developed using a new numerical scheme, applied to a hyperbolic 5‐equation 2‐fluid model. Starting from a finite volume discretisation of a 5‐equation 2‐fluid hyperbolic model and adding appropriate closure relations, a second‐order code is implemented and applied to air‐water flows in horizontal pipes, simulating the 2‐phase to 1‐phase flow process. The code is evaluated in some common standard test cases. A slug capturing application is also discussed. We show, in an air/water horizontal pipe, slug initiation, growth, and development. Moreover, a grid refinement analysis is performed showing that the method is grid independent and we show the code capability to take into account eventual surface tension effects, through the instantaneous pressure relaxation process. Finally, a prediction of flow regime transitions is shown and compared with a well‐known theoretical flow pattern map in addition to a preliminary comparison of computed slug characteristics against well‐known empirical correlations. Abstract : A novel numerical scheme for slug capturing in pipes, adopting a 1D transient, hyperbolic five‐equation two‐fluid model is presented. Through a finite volume method with second order accuracy in space, the code describes well slug initiation, growth, and development. It is able to predict well flow regime transitions; a grid refinement analysis shows the code grid independency. Computed slug flow velocities, frequencies, and lengths are well predicted if compared against empirical relations.