A new mechanistic model for two-phase flow in vertical and inclined pipes
was proposed on the basis of the drift-flux approach. The proposed model,
unlike the other mechanistic models [Ansari et al. (1994); Xiao et al. (1990);
Zhang et al.(2003)] that incorporate a system of nonlinear equations to solve,
uses an explicit equation for liquid-holdup prediction, thus reducing
computation time significantly. Coupled with some simplified assumptions on
pressure/volume/temperature (PVT), such a simple form of a
liquid-holdup-prediction formula enables an analytical integration of the
pressure gradient in two-phase flow along the pipe. This procedure is used
usually to speed up the calculation of bottomhole pressure (BHP) for a large
number of wells for oil-production-optimization purposes (Khasanov et al.
The drift-flux approach can predict liquid holdup for bubbly flow quite
accurately. However, for slug flow, it usually underestimates the void
fraction. Because slug flow is the most common flow in producing wells, this
leads to the pressure drop being overestimated significantly; this can be
proved by comparing computational results to the experimental data and
mechanistic models. Small gas bubbles in liquid slugs should be accounted for
when predicting liquid holdup for slug flow more accurately. Gas in the slug
body is considered by adding a proper term to the void-fraction expression.
This term is based on the correlation for liquid holdup in the slug body. The
model was evaluated with Rosneft's field data and the Tulsa University Fluid
Flow Projects (TUFFP) databank. The model was evaluated by comparing it with
three mechanistic models for multiphase flow.
The major tasks for every oil company are oil-production maximization and
operational-costs reduction. This requires permanent well-production monitoring
by selecting the most promising wells and performing operations on those wells
to increase production (well-enhancement routines). The key objective of
well-production monitoring is to control the productivity index and well
potential for every well during the well lifecycle. This requires the well BHP
to be determined. In some cases, direct measurement of BHP is either difficult
or uneconomical; that is why BHP calculation is still a relevant problem.
The complexity of the pressure-gradient prediction grows out of the
multiphase character of the mixture flowing through oil wells. The multiphase
mechanistic models (those in Ansari et al. 1994; Xiao et al. 1990; Zhang et al.
2003) allow prediction of the pressure gradient with high accuracy. Such models
together cover the whole range of pipe-inclination angles and input parameters.
They are applicable for a detailed analysis, but because they usually
incorporate a system of nonlinear equations to solve, the computation time can
be quite long.
When an oil company operates thousands of wells, it is important for them to
use their regular analysis to choose those wells the optimization of which
would be most beneficial (Khasanov et al. 2006). For such cases, the use of
mechanistic models can be rather difficult because their iterative procedures
require lengthy computation times.
The purpose of this paper is to develop a simple mechanistic model for
pressure-gradient prediction that (1) is applicable for the whole range of
input data and (2) has a simple unified form of void-fraction expression for
all considered flow patterns.
© 2009. Society of Petroleum Engineers
View full textPDF
- Original manuscript received:
13 April 2007
- Meeting paper published:
27 June 2007
- Revised manuscript received:
15 August 2008
- Manuscript approved:
10 September 2008
- Published online:
2 March 2009
- Version of record:
26 February 2009