Embed Size (px)
Citation preview
STUDY OF SLUG CONTROL TECHNIQUES IN PIPELINE SYSTEMS
JOSÉ L. A,VIDAL
Petrobrás Research Center - CENPES/PDEP/TOOL
Av.Horácio de Macedo 950- Cidade Universitária – 21941-915 -Rio de Janeiro-RJ
E-mail:[email protected]
PAULO C.C, MONTEIRO
Ocean Engineering Department, Federal University of Rio de Janeiro, COPPE/UFRJ
Cidade Universitária – Bloco C - Ilha do Fundão— 21945-970- Rio de Janeiro-RJ
E-mail:[email protected]
Abstract - Severe slugging may occur at low flow rate conditions when a downward inclined pipeline is followed by a vertical
riser. This phenomenon is undesirable for offshore oil and gas production due to large pressure and flow rate fluctuations. It is of
great technological relevance to develop reliable and economical means of severe slugging mitigation. This study aims to develop
an automated control system to detect and mitigate the formation of severe slugging through a choke valve and a series of sensors.
As a first step, an overall flow map is generated to indicate the region within which severe slugging may occur based on Boe’s
criterion (Boe, A, 1981 ) and Taitel’s model (Taitel, Y.,1986; Taitel, Y., 1990). It was possible to obtain different (Taitel, Y., 1986) flow patterns by controlling the rate of water and gas injection. The aim of this paper is, however, the formation of severe slugs
and study of mitigation techniques. In the control part, we used a choke valve controlled by software which is in feedback with
data from a system with pressure, temperature, flow, which are able to measure even small changes in the relevant parameters to
the model. A two-phase flow loop was built for the study of severe slugging in pipeline-riser system with air and water as work
fluids. The inner diameter of riser and flowline is 76.2 mm. The riser is 20 meters high and the flowline is 15 meters long and
could be inclined upward or downward up to 8-degree. It has been shown by experiments how riser slugging can be controlled by
automated control system.
Keywords-severe slugging, slug control, flow, flowline, hydrodynamic, slug periods, slug length, riser, control algorithm,
automatic control.
1 Introduction
The multiphase flow occurs in many processes in the
oil industry, in the production and transportation, wells
and the links between these and platforms. Several
studies have been undertaken in order to predict their
behavior as this has a large effect on the productivity and
safety of equipment (Schmidt, 1985; Pots, B. F. M.
,1987).
The flow of gas and liquid simultaneously have
various types of configurations or patterns, depending on
the operating variables, speed, fluid pressure, pipe
diameter and inclination angle. It is known that the
equations that predict the behavior of fluids vary
considerably according to the type of flow. Among the
flow patterns, the intermittent was the object of study of
this project. According to the literature, the intermittent
flow can be subdivided into plug flow and slug flow. The
plug flow occurs in general for low flow rates and bubble
are free flowing inside the liquid. At high flow rates of
gas bubbles have small size as the fluid was aerated. The
slug flow may also be divided into two types, the
hydrodynamic and severe. The hydrodynamic slug can
form in horizontal sections or wells and risers. The
severe intermittence or slug occurs from the
accumulation of fluid in the sections downhill by
gravitational effect. This phenomenon can occur in
pipeline-riser systems, which the pipeline is downhill
and soon rises to the riser. At low flow rates of liquid
and gas, due to the accumulation of fluid in the riser
blocking the passage of gas, resulting in its compression.
When the gas pressure exceeds the hydrodynamic
pressure of the fluid in the riser, the gas arises expanding
and pushing the fluid in the column to the separator. This
phenomenon results in periods without production
followed by a large amount of liquid and gas as well as
sudden changes in pressure.
Slug, therefore is the formation of large gas bubbles
in a flow regime that occurs within the multiphase
pipeline transportation and production of hydrocarbons.
Characterized by large variations in flow and pressure
occurring in the whole process of production and
Anais do XX Congresso Brasileiro de Automática Belo Horizonte, MG, 20 a 24 de Setembro de 2014
3382
transportation. The slugs generate undesirable
consequences in the process of oil production, such as no
oil periods, followed by high oil production within the
separator. The reduction in production capacity,
emergency stop on the platform due to the high liquid
level in the separator, corrosion and wear to the process
equipment and high maintenance costs are some
consequences of this phenomenon. The development of
slug flow regime starts from stratified flow due to two
factors: the natural growth of small disturbances present
in the flow or due to accumulation of fluid caused by
slope changes in the profile of the duct. A scenario in
which this can occur comes from wells whose production
line slope is downward, still associated to the presence of
stratified flow regime.
A few systematic studies have been conducted to
account for the changes in the operational conditions
when applying methods to eliminate severe slugging
(Jansen, F. E. & Shoham , 1994), the most used technics
for eliminations are gas injection (gas lift) (Hill, T. J.
1989; Hill, T. J. 1990) and choke valve systems
(Schmidt, Z.1979) . This work deals this problem using
choke valve procedure through automatization control
and algorithm and cascade PID.
2 Experiments
As proposed in this article, mitigation slug flow is the
main objective of this project. To achieve it was
necessary to conduct a study of the hydrodynamics of the
flow. This phase, coordinated by Prof. Su Jian, was
aimed to determine the conditions necessary for
producing a slug flow regime in the duct system. The
system characteristics such as pressure drop, inclination,
height of vertical section, length inclined section and
flow of gas and liquid. Those parameters were used to
perform the simulations. The simulations aimed at
guiding the experimental tests, providing the flow
parameters of water and air which were used in the tests.
Initially we attempted to reproduce the types of flows in
the system until the slug flow regime, and then used the
automatic control valve to mitigate them. The loop
system can be divided into three parts: pipes, pump and
injection system and acquisition/control system.
This setup consists of a PVC and acrylic pipes with 4
inch of diameter and total length of 120 m. It has a
vertical column with 20 m high and a water reservoir
representing the riser-separator system, an inclinable
section with 16 m long that can be inclined by ± 8ºC
representing the region between the flowline and the
touch-down point as shown in Figure 1 and Figure 2.
Figure 1: Vertical column with 20 meters high.
Figure 2: Inclinable section with 16 meters long.
The injection system is divided into two: the
circulation of water and injection of compressed air. A
40 hp computer-controlled pump makes the circulation
of water. A compressor and a computer-controlled flow
valve perform the compressed air injection system.
The acquisition system is based on a set of two
multivariable sensors (pressure, flow and temperature)
and two controllers for the valves, one for compressed
air and the other for the water choke valve that used for
slug mitigate and control. A PLC (Programmable Logic
Controller) using PID and Cascade PID link all the
sensors and controllers for parameters control.
Anais do XX Congresso Brasileiro de Automática Belo Horizonte, MG, 20 a 24 de Setembro de 2014
3383
3 Control Algorithm
There are two possible types of control for this setup:
Bottom Control and Platform control. In both cases, the
control is performed by cascade PID with the pressure in
master loop and flow in slave loop.This configurations
aims to prevent that the control valve produces pressure
peaks that may result in damage to the pipe.
3.1 Bottom Control
In bottom control configuration the control valve, the
pressure and flow sensors are on the seabed by means of
sensors located at the wellhead. Presents the advantage
of the ability to anticipate the slug before reaching the
platform. Figure 3 shows a schematic drawing of the
control in the bottom.
Figure 3: Schematic of bottom control.
3.2 Platform Control
In this control way, all the sensing are accomplished by
measuring pressure and flow at the top, in platform. Has
the advantage of easy access to the sensor network,
however, requires the control to be fast in mitigating
slugs. The schematic of the platform control is shown in
Figure 4.
Figure 4: Schematic of platform control.
4 Results
To identify the flow conditions that produce slugs,
tests were performed with different water flows keeping
the gas flow constant. The results for flow are shown in
Figure 5. It is possible to see the flow behavior of two-
phase fluid flow in different water flow rates for the
same gas flow. We observe that occur at low flows rates
large fluctuations in flow, these represent the slug, and
tend to diminish with increasing pump flow rate.
0 5 10 15 20 25 30 35 40 45 50
5
10
15
20
25
Flo
w (
m3/h
)
Time (min)
7.42 m3/h
7.63 m3/h
7.85 m3/h
8.28 m3/h
8.96 m3/h
9.35 m3/h
9.85 m3/h
Air Flow = 10 l/min
Figure 5: Flow conditions with different pump flows.
Anais do XX Congresso Brasileiro de Automática Belo Horizonte, MG, 20 a 24 de Setembro de 2014
3384
0 10 20 30 40 50
20
40
60
80
100
120
140
7.42 m3/h
7.63 m3/h
7.85 m3/h
8.28 m3/h
8.96 m3/h
9.35 m3/h
9.85 m3/h
Air Flow = 10 l/min
Time (min)
Pre
ssu
re (
psi)
Figure 6: Pressure conditions with different pump flows.
We also examine the behavior of the pressure in an
equivalent position at the bottom, or near the riser while
maintaining the gas flow rate constant. The result is
shown in Fig. 6. It is also observed that the presence of
fluctuations decrease with increasing pump flow.
Other results were obtained for different water flows.
It should be noted that two different flow rates were
measured, the water flow at the pump outlet (pump flow)
and other measures in the study region (two phase flow),
this is justified by the fact that during the occurrence of
slug biphasic flow variation is considerably large to be
used as a defined parameter. Interestingly, however,
analyze the variation of these parameters (two phase
flow and pressure) rather than the average over the
occurrence of slug flow. Based on the results shown in
Fig. 5 and 6 was possible to determine correlation with
the flow of water and air of great interest parameters:
duration and period of the slug, through this analysis was
possible to observe the effect of control used in the
mitigation.
Figure 7 shows the variation of the period of the
slugs due to the increase of pump flow and gas flow. It is
observed that there is a decrease in function of increased
pump flow as the flow of gas, this behavior occurs until
there is no more slugs and the flow becomes a sparse
bubble.
8 10 12 14 16 18
60
80
100
120
140
160
180
200
7.42 m3/h
7.63 m3/h
7.85 m3/h
8.28 m3/h
Air Flow (l/min)
Slu
g P
eri
od
(s)
Figure 7: Slug periods for different flows of water and gas.
8 10 12 14 16 18
50
55
60
65
70
75
80
7.42 m3/h
7.63 m3/h
7.85 m3/h
8.28 m3/h
Air Flow (l/min)
Slu
g L
en
gh
t (s
)
Figure 8: Slug lengths for different flows of water and gas.
The same applies to the length of the slug, this
becomes increasingly faster until they can no longer be
observed. The results are shown in Fig. 8.
Another form of analysis is the use of variation of
flow and pressure instead of absolute measurements, this
method avoids measurement errors due to lack of
calibration of the sensors used on platforms or oil and
gas production plants. The following results, shown in
Fig. 9 shows the evolution of the two-phase flow
variation measurements at the base of the riser as
function of air flow. The results show the decrease of
flow variations with increasing pump flow, meaning that
increasing the velocity of the air-water mixture obtains a
reduction of the effects of slugging. On the other hand,
the increase of the gas injection also causes an increase
in the effect of variation in slug flow of the mixture.
Anais do XX Congresso Brasileiro de Automática Belo Horizonte, MG, 20 a 24 de Setembro de 2014
3385
8 10 12 14 16 18
0.5
1.0
1.5
2.0
2.5
3.0 7.42 m
3/h
7.63 m3/h
7.85 m3/h
8.28 m3/h
Air flow (l/min)
Flo
w v
ari
atio
n (
m3/h
)
Figure 9: Flow variation for different flows of water and gas.
8 10 12 14 16 18 20
6
8
10
12
14
16
18 7.42 m
3/h
7.63 m3/h
7.85 m3/h
8.28 m3/h
Pre
ssu
re v
ari
atio
n (
psi)
Air flow (l/min)
Figure 10: Pressure variation for different flows of water and
gas.
In Figure 10, there is a similar behavior to the
pressure variation due to increases in pump flow and the
injection of gas, the first one causes a decrease of the
pressure variation at the base of the riser and the second
causes an increase in pressure variation.
Through testing, it was possible to identify necessary
conditions for reproducibility of the slug. During control
tests, it was decided to use the combination of water flow
and gas that provides of longer length slugs, this is due to
the fact that, in general, the slugs that occur in
production plants are long lasting.
5 Control Tests
For reasons of comparison, tests were performed
initially with manual actuation of the choke valve. This
method is commonly used in production plants. The flow
parameters of the pump and gas injection were used in
all the control tests were respectively: 7.42 m3/h and 10
liters/min. The valve actuation pattern used for manual
control was square wave type with the same length of
slug.
5.1 Manual Control
The results presented below show comparisons
between variations in flow and pressure with and without
the use of manual control. Figure 11 shows flow during
slug cycles for cases with and without manual control
and Fig. 12 shows the results for pressure. The result
shows that there was no great differences between this
with or without control.
0 5 10 15 20 25 30 35 40 45 50
5
6
7
8
9
10
No Control
Manual Control
Flo
w (
m3/h
)
Time (min)
Figure 11: Flow comparison with and without manual control.
0 5 10 15 20 25 30 35 40 45 50
20
25
30
35
40
No Control
Manual Control
Pre
ssu
re (
psi)
Time(min)
Figure 12: Pressure comparison with and without manual
control.
Figure 13 and 14 shows the average flow rate over a
period of 10 min for the cases with control and without
control respectively. These results aims to observe
whether there was a decrease of mean flow caused by the
pressure drop resulting from choke valve actuation.
Anais do XX Congresso Brasileiro de Automática Belo Horizonte, MG, 20 a 24 de Setembro de 2014
3386
10 12 14 16 18 20
0
2
4
6
8
10
Flo
w (
m3/h
)
Time (min)
Average Flow = 7.0 m3/h
No Control
Figure 13: Average flow without control.
10 12 14 16 18 20
0
2
4
6
8
10
Flo
w (
m3/h
)
Manual control
Average Flow = 6.9 m3/h
Time (min)
Area=4181
dx=599.6
Figure 14: Average flow with manual control.
The comparative results show that the activation of the
valve causes a decrease of peak flow as shown in Fig. 11
and very few changes in pressure show in Fig. 12 and
results in a slight reduction in the mean flow.
5.2 Automatic Control
The automatic control uses the same conditions for
all tests, as reported previously. The configuration
chosen for the tests was the use of the bottom control.
The process of automatic control is based on the use of
two meshes, master (pressure) and slave (flow). This is
possible by reading these previous parameters and the
use of it in a feedback process. The following results in
Fig. 15 show the comparison between flow with and
without automatic control process. Figure 16 shows the
same result for pressure. One can observe the effect of
automatic control of flow peaks, it is clear that they are
considerably decreased. The same occurs with the
pressure. One can also observe that the control requires a
few minutes to get in tune with the slug cycles from
which it becomes effective.
0 5 10 15 20 25 30 35 40 45 50
5
6
7
8
9
10
11
12
No control
Automatic control
Flo
w (
m3/h
)
Time (min)
Figure 15: Flow with and without automatic control.
0 5 10 15 20 25 30 35 40 45 50
20
25
30
35
40
No Control
Automatic control
Pre
ssu
re (
psi)
Time (min)
Figure 16: Pressure with and without automatic control.
From the point of view of the average flow also in an
interval of 10 min, Fig. 17 shows the average flow
without control process and Fig. 18 with automatic
control process. It can be observed that there was an
increase in average flow in comparison with no control
process in contrary to what was observed in the manual
control. Because of the process control, there was an
increase in average flow of about 8%.
Anais do XX Congresso Brasileiro de Automática Belo Horizonte, MG, 20 a 24 de Setembro de 2014
3387
10 12 14 16 18 20
0
2
4
6
8
10
Flo
w (
m3/h
)
Time (min)
Area=4211
dx=598.1
Average Flow = 7.0 m3/h
No control
Figure 17: Average flow without control.
10 12 14 16 18 20
0
2
4
6
8
10
Automatic control
Average Flow = 8.10 m3/h
Flo
w (
m3/h
)
Time (min)
Area=4863
dx=599.6
Figure 18: Average flow with control.
6 Conclusions
Tests performed in the experimental setup allowed to
produces slugs with different flow characteristics,
pressure, period and length. It was possible to perform
three different comparisons: manual control vs. no
control, automatic control vs. no control and automatic
control vs. manual control. The tests showed that a
manual control results in negligible decrease of pressure
peaks, however slightly decreasing the average flow rate,
or decreasing the production. The automated control was
allowed to mitigate slug in addition to decreasing the
pressure peaks up to 57%, decrease flow peaks up to
50% and an increase the average flow rate up to 8% in
comparison with the results without any slug control.
These results show that the control system developed in
this project is able to tune into the slugs cycles and
mitigate them. As future tasks will be necessary to
perform further tests with different features of slug flow
conditions and to test with the platform configuration in
order to determine the benefits and limitations of the
technique.
Acknowledgments
The authors would like to thank the financial support
from PETROBRAS S.A.
References
Boe, A. 1981 Severe slugging characteristics, selected
topics in two-phase flow, NTH, Trondheim,
Norway.
Taitel, Y. 1986 Stability of severe slugging. Int. J.
Multiphase Flow 2, 203-217.
Taitel, Y., Vierkandt, S., Shoham, O. & Brill, J. P. 1990
Severe slugging in a pipeline-riser system,
experiments and modeling. Int. J. Multiphase Flow
16, 57-68.
Schmidt, Z., Doty, D. R. & Dutta-Roy, K. 1985 Severe
slugging in offshore pipeline riser-pipe systems.
Soc. Petrol. Engs. J. 25, 27-38.
Pots, B. F. M., Bromilow, I. G. & Konijn, M. J. W. F.
1987 Severe slug flow in offshore flow-line/riser
systems, SPE 13723. SPE Prod. Eng. 2, 319-324.
Jansen, F. E. & Shoham, O. 1994. Methods for
eliminating pipeline-riser flow instabilities, SPE
27867, presented at SPE western regional meeting,
Long Beach (March 23-25), 193-204.
Hill, T. J. 1989 Riser-base gas injection into the S.E.
Forties line. Proc. 4th Int. Conf. BHRA, pp. 133-
148.
Hill, T. J. 1990 Gas injection at riser base solves
slugging, flow problems. Oil & Gas J. 26, 88-92.
Schmidt, Z., Brill, J. P. & Beggs, H. D. 1979 Choking
can eliminate severe pipeline slugging. Oil & Gas J.
12, 230-238.
Anais do XX Congresso Brasileiro de Automática Belo Horizonte, MG, 20 a 24 de Setembro de 2014
3388
