вход по аккаунту


Holdup in Vertical Three Phase Flow.

код для вставкиСкачать
Dev. Chem. Eng. Mineral Process., 7(1/2),pp.57-67, 1999.
Holdup in Vertical Three Phase Flow
G.S. Woods
F. G. Wilson Engineering Ltd, Old Glenarm Road, Larne, Co Antrim
P.L. Spedding, J.K. Watterson and R.S. Raghunathan
Department of Aeronautical Engineering, The Queen 's University of
Bevast, David Keir Building, Stranmillis Road-, Beljast, BT9 JAG,
Northern Ireland
Vertical three phase air-water-oil data in a. 0.026 m i d pipe were presented. The
liquid holdup data presented as a complex pattern againsty, the oil liquid volumetric
ratiofor annular flow at one value of total superficial liquid velocity. Theform of the
pattern was explained in quafititive terms depending on the nature of thejlow pattern
present. Generallyfor the water dominated regime below f o = 0.5 the overlying oil
annular firm protected the water annular fi-om the momentum effects of the gas.
Above fo
0.5 the opposite took place and the water annulus $lm thinned until
inversion to oil dominatedjlow occurred. This was a region of intense miring of the
liquid phases and the hold-up rose dramatically.
Although three phase gas liquids vertical flow is used extensively in industry very
little work has been published on the subject. Early work on vertical flow of gas, oil
and water was conducted on flowing wells where changing conditions along the riser
meant that the work bore little relevance tothe fundamentals of fluid dynamics
G.S.Woods et al.
involved and was more concerned with the PVT behaviour of the well fluids and gas-
lift operation [l-31. Shean [4] studied the upward flow of air-oil-water mixtures in a
0.019 m i.d. pipe. Flow patterns, holdup and pressure loss data were presented for
V T from
1.22 to 6.10 m s-' and oil to liquid fractions from 0 to 1.0. Four flow
regimes were identified and the regime map and frictional pressure prediction method
of Govier et al. [5] were extended to the three phase situation. Pleshko and Sharma
[6] tested the flow regime transition model presented by Taitel et al. [7]for two phase
gas liquid flow against vertical three-phase flow data recorded in a 0.051 m. i.d. tube.
The conclusion reached was that two-phase flow models tested were unsuitable for
predicting three-phase flow transitions.
In summary existing three phase vertical data are meager and give little insight
into holdup of the various phases. What is clear that the data bear very little
resemblance to the two phase condition.
A rig was constructed and verified. It consisted of a vertical test section made of
0.026 m i.d. perspex pipe, and fitted with quick closing valves to measure holdup and
tapping points with gas liquid separators to measure pressure loss. The flow regimes
were determined by a combination of techniques, visual, video, pressure drop data,
holdup data etc. The oil used in the experiments was light liquid parafilm Finavestan
A 50 B density 829 kg m", viscosity 0.01 167 kg m" s-' and surface tension 0.03 Nm-l
at 20°C. Details of the apparatus are given elsewhere 181.
Figure 1 presents the liquid holdup data obtained for one total liquid rate of
V L=
~ 0.0625 m .s-l and one gas rate
V,G = 22.0 m S-l
with the oil/ratio
varying from 0 to 1.O. There were three liquid holdups, the total liquid holdup RLT,
the water holdup
EL,and the oil holdup Rw.
Holdup in vertical three phase flow
Figure 1.
Total, water and oil holdups against oillliquid ratio for vertical air-water
flow in a 0.026 m i.d. pipe. Superficial liquid velocity of 0.0628 ms-l
and superficial gas velocity of 20 ms-1.
G.S.Woods et al.
Figure 2.
Liquid film thickness versus-oil/liquid ratio for vertical air-water-oil
pow. Diameter = 0.026 m, VSL = 0.0628 m s-1, VSG= 20 ms-l.
Holdup in verrical three phaseflow
Since the flow was generally in the high gas velocity annular type regime it was
possible to calculate the film thickness assuming the liquids formed annulus or
annular rings around the inner pipe wall. It was also possible to calculate the average
liquid velocities. These values are presented in Figures 2 and 3. Figure 4 presents the
observed total pressure loss, the head pressure loss calculated from the holdup and the
frictional loss as the difference between the two.
Using the data in Figure 1 to 4 and other observations it was possible to ascertain the
mechanisms which occurred so as to present the complex set of data obtained in the
illustrated figures. The detailed explanation given below follows the effect of varying
the oil/liquid ratio f from zero to unity, ie. from left to right on Figure 1.
At f,
0 two phase aidwater annular flow occurred. Because of the relatively
low liquid and high gas rates chosen the water annular film had very little surface
disturbance. The initial addition of oil smoothed the water surface and since the oil
preferentially adhered to the water surface a slight fall occurred in the frictional
pressure drop and a +4-5 % increase took place in the total liquid holdup,
the oil rate was increased the oil moved as a series of fast moving discrete annular
rings that passed rapidly over the water annulus film and tended to partially shield the
water interface from the direct influence of the gas. The result was that a steady
increase in
RLTtook place. There was a corresponding fall in velocity of the water
phase, a rapid rise in the oil phase velocity and a slight fall in total pressure loss due
to a reduction in the frictional component. More significant there was an increase in
the fluctuations in the measured total pressure drop caused by the rapid movement of
the oil annular rings. At f, = 0.3 the oil layer eventually completely covered the water
annulus as the oil rate was such that individual rings of oil with spaces in between
could no longer be accommodated. The oil moved as a relatively thin annular film
with waves moving over the surface continuing to give a fluctuating pressure drop
pattern because of the movement of the waves over the oil surface. Thus at f,
= 0.3
G.S. Woods et al.
Figure 3. Actual superficial velocities versus
- oil/liquid ratiofor-vertical air-wateroilflow. Diameter = 0.026 m, VSL= 0.0628 m s-1,
VSG= 20 m s-l.
Hordup in vertical three phase flow
Figure 4.
Pressure drops versus oil/liquid ratio for vertical air-water-oil flow.
Diameter = 0.026 m, VSL= 0.0628 m s-], VSG = 20 m s-].
G.S. Woods et al.
the shielding of the underlying water annulus from any direct contact of the gas was
complete. The result was that the water film velocity started to fall rapidly and its
thickness increased, while the pressure loss fell in all its component parts. However
as waves were now formed on the surface of the now continuous thin oil annular film
there was a dramatic increase in oil velocity and a fall in oil thickness due to the
scouring action of the wave troughs on the underlying oil film. The overall result was
that the holdup
RLTcame to a maximum at f ,= 0.3.
Its subsequent fall was caused
by a dramatic fall in R m as the waves carried the oil phase more readily over the oily
water annulus. In this region the water holdup
remained virtually unchanged
due to a balance between a rise in the water thickness and a fall in its velocity as f,
went up from 0.3 to 0.375. Over this f, range the pressure drop commenced to fall
rapidly because the frictional component was reduced by the oil completely covering
the water annulus. Also the oil phase velocity tended towards a maximum value and
the holdup
RLTwent through a minimum at f, = 0.375.
Between f, -0.375 and 0.5 the velocity of the oil film maximised while the
water film velocity went through a minimum. In this region all the film thicknesses
rose steadily with the total and water films coming to maximum values. The steady
buildup of the oil film in this region led to an increasing shielding of the water
At f,= 0.5 a dramatic change took place as the oil film passed from laminar to
turbulent flow that resulted in increased momentum transfer to the water annulus.
The result was maxima in water and total holdup, head pressure loss, water and total
film thicknesses and oil velocity, while the frictional pressure loss and water and total
velocities went through minima.
Further the excess pressure loss fluctuations
observed at the initial addition of oil to the two phase system ceased dramatically at
f ,= 0.5.
From f,
0.5 to 0.625 the holdup RLT fell dramatically as the water phase
(which up to this stage was itself in turbulent flow) was increasingly thinned and
Holdup in vertical three phaseflow
speeded up by increasing momentum transfer from the gas phase through the oil
annular film.
At f,
0.625 the flow regime began to change as the oil film commenced to
break through the water annulus and join onto the wall. The flow regime began to
change from water dominated (WD) water annulus plus oil annular ripple flow to oil
dominated (OD) broken annulus flow. Short strips of the water annulus de-laminated
from the pipe wall and were replaced by oil, later to be pasted over by water. This
stripping and pasting of the phases consumed energy so the pressure loss rose in all
its components. Further, the oil and total holdup, velocities and thickness rose
dramatically while the water phase holdup and thickness fell away as its velocity rose
sharply. Eventually at f
0.825 the thicknesses and holdups of total liquid and oil,
and pressure drops peaked and their velocities minimised as the OD dispersed annular
regime was formed. Thereafter the total holdup and pressure drop reduced to the two
phase aidoil condition. In this f, regime the water tended to move as a dispersion in
the predominantly oil phase but with more concentration of water droplets towards
the annular film surface of the dispersion.
Thus the complex form of the total holdup relation shown in Figure 1 can be
explained in terms of the annular/annulus nature of the two phases present. It is
significant that the f ,= 0.5 point could be viewed not only as the potential inversion
point (if surface tension and viscosity effect, etc., of the two phases were
insignificant), but also in reality as the point in which the annulus phase moves from
being protected by the overlying annular layer to being increasingly erroded by
momentum transfer from the gas through the annular layer. Moreover the WD
annulus water/annular oil plus ripple regime was more complex than first supposed
forming virtually two regimes; one in which the annulus was protected and the other
beyond f,
= 0.5 where it was
G.S.Woods et al.
Data on vertical three phase holdup in annular flow shows that a complex series of
patterns are formed depending on oil/liquid ratio. At low fo values up until 0.3 the oil
proceeded to cover over the water annulus layer protecting it from the influence of
the gas scouring action so liquid holdup rose steadily to a maximum. At f, = 0.3 the
oil completely covered the water annulus but while the water layer continued to be
protected the oil annular layer velocity increased dramatically leading to an overall
fall in liquid holdup to a maximum at fo = 0.375. Beyond this point the oil annular
layer began to build steadily and increasing protected the annulus water layer so
rose to a maximum at f, = 0.5. At this point the annular oil layer became turbulent
and the annulus water layer thinned rapidly due to increased momentum transfer.
RLTfell at a minimum at f ,= 0.625 where the OD broken annular regime
commenced to be formed. The holdup rose dramatically to a maximum at f ,= 0.823
where the OD dispersed annular regime of water droplets in oil medium was formed.
The holdup then fell away to the air/oil two phase value at f, = 1.O. The liquid holdup
can therefore be explained on the basis of the patterns formed by the two phases.
= annular
AS = annulus
fo = oivliquid volumetric ratio
L = length m
OD = oil dominated
P = pressure kg m-' s-*
R = ripple
R =
v =
thickness m
velocity m s- 1
water dominated
Holdup in vertical three phase flow
F.H.and Carpenter, P.G. The multiphase flow of gas, oil and water
through vertical flow strings with application to the design of gas-lift installations.
Drill Prod. hact. 257-3 17 (1952).
Tek, M.R. Multiphase flow of water, oil and natural gas through vertical flow
strings. J. Pet. Tech. 1029-1036 (1961).
Francher, G.H.and Brown, K.E. Prediction of pressure gradients for multiphase
flow in tubing. SOC.Pet. Eng. J. 59-69 (1963).
Shean, A.R. Pressure drop and phase fraction in oil-water-air vertical pipe flow.
PhD Thesis MIT ( 1976).
Govier, G,W., Sullivan, G.A. and Wood R.K. The upward vertical flow of oilwater mixtures. Can. J. Chem. Eng. 39 67-75 (1961).
Pleshko, A. and Shama, M. An experimental study of vertical three-phase (oilair-water) upward flows. Pet. Eng. Dep. Univ. Wyoming (1990).
Taitel, Y, Barnea, D. and Dukler, A.E. Modelling of flow pattern transitions for
steady upwards gas-liquid flow in vertical tubes. A.1.Ch.E. J. 26 345-354 (1980).
Woods, G.S. Studies in vertical multiphase flow. Ph.D. Thesis, Queen’s
University of Belfast (1 996).
1. Poettman,
Без категории
Размер файла
341 Кб
flow, holdup, three, phase, vertical
Пожаловаться на содержимое документа