GAS INJECTION AT RISER BASE SOLVES SLUGGING, FLOW PROBLEMS

Feb. 26, 1990
T.J. Hill BP Research Sunbury Research Centre Sunbury-on-Thames, U.K. Test-rig studies and operating experience by BP International have shown that riser-base gas injection can help prevent severe slugging and smoothen start-up transients in subsea pipelines. The studies were based on predictions that multiphase lines proposed as part of the Forties field extension would be subject to severe slugging.
T.J. Hill
BP Research
Sunbury Research Centre
Sunbury-on-Thames, U.K.

Test-rig studies and operating experience by BP International have shown that riser-base gas injection can help prevent severe slugging and smoothen start-up transients in subsea pipelines.

The studies were based on predictions that multiphase lines proposed as part of the Forties field extension would be subject to severe slugging.

Test-rig studies were undertaken to analyze severe slugging and to demonstrate gas injection as a means of prevention. Following scale-up studies, riser-base gas injection systems were installed on both 12-in. production and 6-in. test lines. Comprehensive instrumentation was also put onto the lines.

PROPOSED DEVELOPMENT

During 1983-84, a study was undertaken by BP's engineering and technical center on the operability of the multiphase flow lines in the proposed development scheme for the southeast extension to the Forties reservoir (SE Forties).

The scheme then comprised two subsea production templates tied back to the Forties Alpha platform by two flow line loops, one of 250 mm (10 in.) diameter, the other of 200 mm (8 in.).

The study concluded that severe slugging would be a potential problem with the pipeline-riser geometry of the two flow line loops at low production rates because the flow lines would slope down to the risers.

This phenomenon had been previously identified by Schmidt' as a cyclic flow-rate variation, resulting in periods both of no flow and of very high rates substantially greater than the time average.

Fig. 1 illustrates the stages of a severe-slugging cycle.

For reasons unrelated to the multiphase flow lines, the final development scheme was altered to a fixed minimum-facilities platform with one 300 mm (12 in.) and one 150-mm (6-in.) line, each carrying unseparated reservoir fluids, tied back to the Forties Alpha platform.

The operability studies on these lines also concluded that severe slugging could occur at low production rates.

At this stage it was decided that a detailed investigation of the severe-slugging phenomenon was required, including consideration of its occurrence, effects, and elimination.

During this study a request was made back to the SE Forties field development project for the inclusion of an instrumentation and data-acquisition system in the final offshore scheme to allow comparison of the "real life" performance with the design.

OPERABILITY, RIG STUDIES

The work by Schmidt at the University of Tulsa's fluid flow projects (Tuffp) had led to a criterion for predicting the onset of severe slugging, and a computer code was available to Tuffp members for predicting cycle times at different gas and liquid flow rates.

For the 10 and 8-in. lines, the prediction showed the onset of severe slugging at production rates of 2,020 cu m/day (1 2,700 b/d) and 1, 1 60 cu m/day (7,300 b/d), respectively.

For the 12 and 6-in. combination, the severe-slugging boundary was at 3,980 and 590 cu m/day (25,000 and 3,700 b/d). In the latter case these rates were originally predicted for 1993-94 onwards.

The resultant cycle times for the 12-in. line were predicted to be as high as 83 min. For a substantial portion of this time there would be no fluids leaving the riser, followed by a short period of very high production rate.

Test-rig work was commissioned at the BP Research Centre to investigate and confirm the previous descriptions of severe slugging, determine the resultant flowrate variations, and investigate methods of eliminating the phenomenon.

A 50-mm (2-in.) diameter rig, with a 50-m (150-ft) pipeline and a 15-m (50-ft) riser, using air and water as the fluids, and venting to atmosphere, was constructed at the Research Centre.

The pipeline was inclined 2 down towards the base of the riser (Fig. 2). The test rig was inclined more steeply than the intended Forties lines to compensate for its lower L/D ratio.

With the complete range of flow rates available on the rig, a flow-regime map was plotted for the line geometry (described presently). This is shown as Fig. 3, with the severe-slugging regime to the bottom left, at the lower gas and liquid flow rates. The TTss boundary due to Pots2 is marked for comparison at one value of pipeline void fraction.

The severity of the effect of the cycle is quantified in two ways.

First, there is the cycle time, which determined the length of time that the downstream processing facilities would not be receiving fluids.

Second, there is the maximum liquid-production rate during the cycle, coupled with the volume of liquid produced at this rate. This affects the performance of the separation system and the pipework restraints required.

The rig was then used to quantify these effects at different gas and liquid flow-rate combinations.

The variations of cycle times and maximum length of the severe slug are plotted in Fig. 4. This indicates that as production rates drop, for a given in situ GOR, the severe-slugging cycle time increases.

The liquid volume produced during one cycle is plotted against time on Fig. 5. This clearly indicates the no-flow period, the period of steady liquid production, and the subsequent high rate liquid production as the gas breaks through into the riser.

During these investigations a dynamic model of the response of the separation facilities was set up and run to determine their behavior when forced by a severe-slugging cycle.

This cycle itself was reproduced by a modified version of the original Tuffp software. Further account of this work is given in Mackay.3 The Pots criterion has since also been shown to predict that severe slugging could occur in these lines.

The topsides simulations concluded that there was a remote chance of high liquid level alarms, and of liquid carryover, from the separation system during the production stage of the severe-slugging cycle. The test-rig investigation was therefore continued into possible methods for preventing the occurrence of severe slugging.

RISER-BASE GAS INJECTION

The preferred method for the first investigation was that of riser-base gas injection as suggested from discussions between BP and Tuffp. Later work, not covered here, was also conducted into riser-top choking as an alternative method suitable in some circumstances (similar to studies undertaken later by Shell 2).

Increasing amounts of gas being injected through a simple tee into the base of the riser were shown to reduce the extent of the severe-slugging regime and its severity in the area where it was still occurring.

Fig. 6 shows the reduction in the severe-slugging regime for the addition of injection gas at a superficial velocity of 1. 1 5 m/sec. With the addition of sufficient gas at the riser base, severe slugging could be eliminated at all flow line gas and liquid-flow rates.

The condition for prevention of severe slugging was that enough gas be injected to bring the total gas and liquid flow rates in the riser into the annular flow regime, i.e., continuous liquid transport up the riser, thus preventing the riser-base liquid accumulation necessary for the onset of the severe-slugging cycle.

The test-rig studies showed that the plot of liquid volume produced vs. time could be made into a straight line (Fig. 7) at the pipeline flow rates which, without gas injection, led to the severe-slugging cycle shown in Fig. 5.

With the principle of riser-base gas injection now demonstrated, further modifications to the severe-slugging software were made to allow scale-up of the gas-injection rates required for the conditions in the SE Forties lines.

GAS-INJECTION INSTALLATION

The scale-up calculations indicated that gas-injection rates of 210,000 and 56,000 std. cu m/day (7.5 and 2.0 MMscfd) would be required to prevent the occurrence of severe slugging in the 12 and 6-in. lines, respectively.

The two lines from the new minimum-facilities platform, now known as Forties Echo, were to be connected to the existing production and test headers on the Forties Alpha platform.

This platform already had gas-compression facilities for the removal of NGL from the produced gas. After the NGL plant, any remaining gas is flared.

The available compression capacity was shown to be sufficient to be able to supply the injection requirements of the two multiphase pipeline risers in terms of both flow rate and pressure.

During platform start-up, the first gas would have to be drawn from Forties Alpha production and recycled until a sufficient flow rate could be passed into the injection system.

Flow-rate control and metering of the gas injection are both carried out topsides before gas passes into 75 and 50-mm (3 and 2-in.) lines down to the injection tees at the base of the 12 and 6-in. risers, respectively.

The final design of injection tee was optimized with a full-scale model of the 6-in. tee on another test rig at the Research Centre. The 2-in. rig had used a simple dead tee which may have had undesirable possibilities of erosion in actual field use.

Therefore, a design was adopted utilizing a series of holes in the main riser pipe encased in a sleeve to provide an even distribution of the injection gas into the riser. As a guideline, the API 14E erosion equation was used to determine the total hole area into the riser.

GAS-INJECTION INTENTION

A series of courses was held both on and offshore to describe the reasons behind the installation of the gas-injection system. The primary purpose of the design is to eliminate severe slugging which could occur when production rates drop and as water cut increases.

The use of gas injection as an aid during start-up was also described in the courses. The extra gas would be available to lift the first liquid arriving at the riser base, thus preventing the buildup of liquid which could otherwise result in a severe slug as the production rate increases.

The combination of the many different demands on the offshore operations staff during the start-up of a new production facility and the absence of any of those involved in the design and experimentation on the gas-injection system during its start-up has, however, led to a number of situations that may otherwise have been avoided.

Discussion of these follows comments on the instrumentation now in place to monitor the multiphase flow lines and their associated gas-injection systems.

DATA ACQUISITION

A request was made to have instrumentation and an associated data-acquisition system installed to monitor the behavior of the multiphase flow in the two flow lines.

During the installation of the Forties Echo platform and of the tie-ins on Forties Alpha, subsea pressure transducers were located at the bases of the four vertical flow line sections.

Temperature and pressure measurements at the top of the vertical sections were also provided. Signals from these instruments were routed not only to the central control room but also to a datalogging unit dedicated to monitoring the multiphase lines.

As well as these pressure and temperature measurements, the logging system was designed to accept two density measurements (Fig. 8).

During Apr. 14-22, 1988, two topsides gamma-ray density gauges were temporarily mounted onto the 12 and 6-in. lines in turn.

The logging system was used to record pressure, temperature, and density over a range of flow rates in each of the lines. This enabled identification of the flow regime and, where appropriate, characterization of the slugging floW.4

In early May a subsea density gauge was permanently installed on the 12-in. line at the base of Forties Alpha and wired onto the logging system.

Later in the year a topsides density gauge was permanently installed, also on the 12-in. line, to allow monitoring of slugs passing up the riser.

Information from the April trials has been used to assess the performance of the flow lines and of the gas-injection technique for assisting in start-up.

OPERATIONS

There are two aspects to the operational experience gained during the multiphase flow trials of April 1988.

Flow regimes at a number of flow conditions were identified, with data collected for future detailed characterization, and three line start-ups were monitored, with different gas-injection procedures each time.

The variation of flow regime with flow rate is illustrated in Fig. 9. This shows the density traces from the 6-in. line at 1,700, 1,100, 700, and 525 cu m/day (10,700, 6,900, 4,400, and 3,300 b/d), showing the change from regular frothy slugs, to regular more dense slugs, and then to irregular surges.

The last condition is on the border of the severe-slugging regime.

During commissioning of the multiphase lines in late 1986 to early 1987, several very large severe slugs were reported.

These are likely to have occurred as a result of inadequate rates of injection gas being used.

During the April trials, three line start-up conditions were monitored. These are illustrated by Fig. 1 0.

The first of these used medium flow rates of gas injection, with the test line crossed over into the production line. The second start-up used only very low settings of the gas injection, again with the test line discharging into the production header.

The third start-up used high gas-injection rates and flowed the test line into the test separator rather than into the production header.

By far the least smooth start-up was the second one with the low gas-injection rates.

There is less to choose between the first and third start-UPS.

Some surging is inevitable during start-up and further tests are necessary to determine the optimum gas-injection rates and pipework configuration.

Following the initial surge, however, the transition into steady slug flow is smoother with the higher gas-injection rate, which prevents long slugs formed in the pipeline from stalling in the base of the riser.

In addition to the start-up of the 12-in. line, the gas injection was used to assist in kicking off a well that had "died" during the first shutdown. The well was not strong enough to overcome the back pressure from the separators.

The procedure was to open the well test line to flare-base pressure on Forties Alpha and then open the wellhead choke on Forties Echo. At this point the well began to flow into the test line.

When liquid reached the base of the riser at Forties Alpha, however, the flow rate was insufficient to drive it up the riser. Therefore, liquid accumulated at the riser base, and the resultant head increase was sufficient to cause the well to stop flowing.

It was then suggested that gas injection should be used to lift out liquid from the riser base as it arrived, thus preventing the head increase. The well start-up procedure was then repeated with the maximum gas-injection rate into the base of the test line.

The well began to flow as before. But when the liquid reached the riser base, it was lifted out, allowing the well to keep on flowing and recover to the point at which it could be rerouted at Forties Echo back into the 12-in. production line.

It was not possible to cut back the rates in either the 6 or 12-in. lines down to below the predicted severe-slugging boundary for fear of killing one or more wells.

The benefits of the gas-injection system for start-up are evident, however, and indicate the necessity of the system for future years as production rates drop and water cuts increase.

Further analysis of the April tests has continued, with additional data collection being undertaken at regular intervals over the coming years.

ACKNOWLEDGMENT

This article has been produced with the permission of the British Petroleum Co. plc and with the assistance of the Forties field group of BP Petroleum Development Ltd., Dyce.

REFERENCES

  1. Schmidt, Z., Brill, J.P., Beggs, H.D., "Experimental Study of Severe Slugging in a Two Phase Pipeline-Riser Pipe System," SPE Paper 8306. 1979 Annual SPE Technical Conference and Exhibition.

  2. Pots, B.F.M., Bromilow, I.G., and Konijn, M.J.W.F., "Severe Slug Flow in Offshore Flowline Riser Systems," SPE Paper 13723, 1985 Middle East Oil Show, Mar. 11-14, 1985, Bahrain.

  3. Mackay, D.C., "Dynamic simulation of the effects of slugging flow on process plant-a design study," OGJ, Sept. 14, 1987, pp. 67-72.

  4. Hill, T.J., "Multiphase Flow Field Trials on BP's Magnus Platform," JERT 109 (September 1987), pp. 142-147.

Copyright 1990 Oil & Gas Journal. All Rights Reserved.