Eventos Anais de eventos
ENCIT 2016
16th Brazilian Congress of Thermal Sciences and Engineering
DEVELOPMENT OF A NEW APPROACH FOR THE PREDICTION OF PATTERN TRANSITION OF TWO-PHASE STRATIFIED FLOW IN DUCTS AND PIPES
Submission Author:
Daniel Rodriguez , Madrid
Co-Authors:
Elmer Gennaro
Presenter: Daniel Rodriguez
doi://10.26678/ABCM.ENCIT2016.CIT2016-0070
Abstract
Present-day methods for the prediction of flow pattern transition in two-phase stratified flows, like those typical of oil extraction and chemical processing industries, rely mainly on the so-called one-dimensional two-fluids model. This mechanistic model simplifies the problem by averaging the flow over the cross-section and introducing heuristic correlations for the friction factors. More fundamental methods for understanding the underlying physics, based on variations of the Orr-Sommerfeld equation, have not been developed up to an applicability state due to their higher associated computing expenses. This contribution presents the most recent steps in the development of a new approach for the prediction of flow pattern transitions from steady, smooth stratified flow to wavy stratified, and towards the emergence of slugs or pistons. The true three-dimensional description of the flow field is considered into a multidimensional, linear instability eigenproblem, that eliminates the constrains associated with both the one-dimensional two-fluids model and the Orr-Sommerfeld equation, by introducing the real two-dimensional cross-section geometry. Next, the evolution of interfacial instability waves as they propagate downstream is computed using a novel formulation of the parabolised stability equations valid for two-phase flow. Use of modern computational and algorithmic techniques are instrumental in the successful implementation of the proposed approach. Optimized sparse linear algebra for high-order spatial discretisations and shared-memory parallelisation enable us to compute the three-dimensional, unsteady velocity fields in a laptop computer.
Keywords
