Slug flow

Abstract

Introduction: When two phases flow concurrently in a pipe, they can distribute themselves in a number of different configurations. The gas could be uniformly dispersed throughout the liquid in the form of small bubbles. There could be large gas bubbles almost filling the tube. There could be an annulus of liquid and core of vapor with or without drops of liquid in it. The interface could be smooth or wavy. When one describes how the phases are distributed, one is specifying the flow regime. Such a description is necessary before any mathematical model can be constructed which will predict a quantity such as pressure drop It is naive to expect that a single mathematical model would adequately encompass all possible two-phase flow regimes, even for a single geometric configuration. Therefore, we shall begin by saying that for this work the results that have been obtained and the conclusions that have been drawn apply only to fully developed slug flow in a round vertical pipe. Slug flow is characterized by large bubblesalmost filling the tubewhich are separated by slugs of liquid. The nose of the bubble is rounded and the tail generally flat. One may or may not find small bubbles in the slug following the large bubble. A number of typical slug flow bubbles are pictured in Figures 4-10. Bubbles very similar to these have been studied by Dumistrescu (1), and Davis and Taylor (2). Both these references consider the same problem. How rapidly will a closed tube full of liquid empty when the bottom is suddenly opened to the atmosphere. The approach used by both authors is to assume that the asymptotic rise velocity (for large times) can be calculated from potential flow theory. The boundary condition at the pipe wall is that the velocity is axial. At the bubble boundary it is assumed that the pressure is constant, The problem is then to find the shape of the bobble that would satisfy the constant pressure boundary condition.

(cont.) This was done approximately and in both cases the comparison with experiment was satisfactory though the deviations became large for small tubes. The work of Davis and Taylor, and Dumitrescu served as the starting point for this investigation. The boundary condition at the bubble wall for large bubbles, constant pressure, was still valid to an excellent approximation and the finiteness of the slug flow bubbles did not appear to make much difference in their rise velocity. In the next section, the fluctuation period, the mean density, and the pressure drop will be expressed in terms of the pipe area, the Taylor bubble rise velocity and the flow rates of the two phases. In subsequent sections the observations rade of bubble shape, length and velocity will be described and then a comparison of computed and measured pressure drops given.