By-pass pig passes test for two-phase pipelines

H.L. Wu
Wu Flow Consultancy B.V.
The Netherlands

G. van Spronsen
Shell International Exploration & Production B.V.
The Netherlands

E.H. Klaus
BEB Erdgas & Erdöl Gmbh
Hannover, Germany

D.M. Stewart
Shell Todd Oil Services Ltd.
New Zealand

The method can widen the possibility of applying two-phase flow pipeline transportation to cases in which separator or slug-catcher capacity is limited by practicality or cost.

Pigging two-phase pipelines normally generates large liquid slug volumes in front of the pig. These require large separators or slug catchers. Using a high by-pass pig to disperse the liquid and reduce the maximum liquid production rate before pig arrival has been investigated by Shell Exploration & Production companies.

A simulation model of the dynamics of the pig and related two-phase flow behavior in the pipeline was used to predict the performance of by-pass pigs. Field trials in a dry-gas pipeline were carried out to provide friction data and to validate the model.

It was then used to explore operating possibilities in a two-phase line which led to the follow-up trial in a 15.6 km 20-in. OD two-phase offshore interfield pipeline with risers. Whereas the volume of liquid swept in front of the pig would be 179 cu m if the by-pass fraction were 0, a reduction of 70% to 53 cu m was achieved in the field with a by-pass fraction of 10%.

("By-pass fraction" is the effective by-pass passage cross-section area divided by cross-section of the pipe.)

The predicted mobility of the high by-pass pig in the pipeline and risers was verified and the beneficial effects due to the by-pass concept exceeded the prediction of the simplified model.

Liquid removal

In two-phase flow pipeline systems, liquid removal by pipeline pig is often required. To avoid the need for large slug-catching equipment for temporary storage at the terminal, it is desirable to be able to even out the swept volume of liquid arriving in front of the pig so that the peak liquid-arrival rate can be lessened.

One means to achieve this is to reduce the pig velocity which will reduce the volume of liquid pile-up ahead.

For example, if the pig's speed were reduced to equal the actual liquid velocity in the line, no pile up of liquid in front of the pig will occur. It is possible to reduce pig velocity with by-passing, by which a passage through the pig is made to allow gas to pass.

This approach is particularly attractive because the by-pass gas itself will also contribute in minimizing the liquid peak by its penetration through the liquid or slug in front of the pig.

Depending on the application, relatively high by-pass fractions will be necessary. The risk of standstill is a concern because of the resulting reduction in the force available for propagation.

Development

The disc-type by-pass pig shown in Fig. 1 [25841 bytes] is designed and manufactured by Kopp for pipelines 20-in. OD or larger. It has been modified to include a by-pass passage connecting the rear end with the front through the center cylinder.

The gas leaving the front is deflected towards the wall by the front plate. A special feature of this pig is the possibility of changing the by-pass passage size by altering the spacers supporting the front plate. The weight of the pig is 120 kg.

Development of the by-pass pig technique for a two-phase flow pipeline involved different stages. The preliminary part of the work, predicting the dynamics of the pig followed by a field test in a gas line, has been described elsewhere.1

A model was at first created to simulate the balance of forces on a by-pass pig and the dynamics of its motion. The hydraulics of two-phase flow in the pipeline were also included in the program.

This simulation and an assumed friction behavior made it possible to estimate the range of by-pass fraction of a pig for operation without the risk of standstill.

Subsequently, two field trials in a dry-gas land line of 0.49 m OD and 35 km length were carried out with by-pass fractions of 10 and 15%.

The purpose was to verify the model and to collect data on pig friction behavior. The land line was convenient for this reason because the speed of the pig could be measured at different locations along the line.

The main results from the dry-gas pipeline field trial were:1

• Pigs with by-pass openings of up to 15% can operate without problems of stoppage in the gas pipeline of 20 in. diameter.

• Considerable wear occurred on the seal material, particularly during the early part of the journey as was also revealed by the correspondingly lower speed of the pig.

• For new seals on the pig, friction appeared to be greater in the beginning and decayed logarithmically with distance traveled.

• Validity of the model as a means to support the development effort was proved.

Field trial

The pipeline chosen for this field test was the interfield gas and condensate pipeline in the Maui fields offshore the northern island of New Zealand.

Fig. 2 [39227 bytes] presents a scheme of the system. The 0.464 m OD pipeline between the platforms MPB and MPA crosses a distance of 15.6 km horizontally at a depth of 100 m with few elevation changes. The flow rate for the test was 87 std. cu m/sec.

The condensate-to-gas ratio of the stream was 280 cu m/mm/std. cu m. With the pressure and temperature at MPB being 17 MPa (170 bara) and 70° C., the averaged superficial gas and liquid velocities in the line were 3.10 and 0.12 m/sec, respectively.

Under equilibrium flow condition, the corresponding liquid hold-up volume in the pipeline was 264 cu m and the actual liquid velocity in the pipeline was 1.09 m/sec. The capacity of the slugcatcher at MPA was only 25 cu m and the maximum continuous liquid processing rate without flooding the slug catcher was 0.08 cu m/sec.

In anticipation that friction force on the pig in the two-phase line would be less than observed in the gas-pipeline trial because of the lubricating effect by the liquid, a smaller by-pass fraction of 10% was chosen for this test.

The flow condition of the line was allowed to stabilize before the pig was launched. To avoid flooding of the slugcatcher at MPA, the test procedure was planned so that gas flow at MPA was to be reduced by 75% as soon as the front of the slug arrived. This could be detected from a pressure dip at MPA as a result of the change in static head when the slug entered the riser.

The gas flow rate was regulated by controlling the throughput of the low-temperature separation units which were operating in parallel. After arrival of the slug front in the separator, the gas flow was to be ramped up gradually to the original flow rate. This would be accomplished by ensuring that the rate of liquid entry into the separator did not exceed the maximum.

Fig. 3 [51255 bytes] shows the gas and liquid arrival rates measured at MPA during the field trial. The pig was launched at time 0 sec, and the slug front arrived at MPA after 6,900 sec. Subsequently, the gas flow was reduced and ramped up later as planned.

The slug volume arriving in front of the pig was 53 cu m. The residence time of the pig was 8,900 sec with an average pig speed of 1.76 m/sec.

As predicted, the high by-pass fraction of the pig did not affect the upward motion of the pig in the vertical riser.¹ After its arrival at MPA, a typical dry period of about 1,250 sec followed without liquid flow.

The liquid flow rate was measured by indirect means at the outlet of the separator with an expected accuracy of ±20%.

Comparison with prediction

A surprising difference between the performance measured (Fig. 3 [51255 bytes]) and predicted (Fig. 4 [49018 bytes]) was the smaller swept volume of liquid arriving ahead of the pig, that is, 53 compared to the predicted 140 cu m.

Table 1 [8413 bytes] presents these data together with the values of gas starvation during pig arrival.

Corresponding values for a pig without by-pass under the same flow condition are also included for comparison. They reveal that the improvements in the reduction in swept slug volume as well as gas starvation are better than 70% (Table 1).

To investigate the reason for the low swept slug volume, it would have been useful if the pig velocity variation had been monitored but this was not possible.

Instead of pig-velocity data, the pressure drops over the pig, derived from recorded MPB and MPA pressures, are compared with the predicted values as shown in Fig. 5 [37209 bytes].

The corresponding predicted pig-velocity variation is plotted in Fig. 4. The measured pressure drop makes evident that the typical "stick slip" phenomenon of the pig occurred at least during the first 1,500 sec.

Although the predicted and measured residence times of the pig were roughly equal, the actual pig velocity must have differed from the predicted not only because of the difference in pressure drop over the pig in the beginning of the journey but more importantly because of "stick-slip" behavior.

Inspection of the pig after the trial revealed evidence that temperature of some parts of the seal had exceeded 100° C. It is postulated that the over heating to greater than the inlet stream temperature of 70° C. must have been caused by friction generated at high velocities.

It is not uncommon that "stick-slip" can be accompanied by large velocity excursions,2 and the overheating of material may also have exacerbated this effect.

The observed evidence makes it reasonable to assume that the "stick-slip" action involving several force reversals must have drastically reduced the average pig velocity and most probably to below the equilibrium liquid velocity of 1.09 m/sec. This will have had a strong tendency to even out the liquid swept up in front of the pig.

In the latter part of the journey, the velocity must have increased to a level greater than the overall average speed of 1.76 m/sec but apparently this did not result in significant slug build-up.

The occurrence of "stick-slip" may be minimized in future after modifications to reduce the effects of seal wear and overheating. If speed reduction will be lessened by the elimination of "stick-slip," the possibility still remains to reduce pig speed again by increasing the by-pass fraction to 15% without the risk of standstill as was indicated by the results of the trial in the dry gas pipeline.

Therefore expectations for the next trial in the Maui pipeline are good.

Although further tests and refinement of the technique will be required, it is anticipated that the by-pass pig will eventually enable a relaxation on the requirement for excessive slug catcher volumes in future pipeline designs.

Acknowledgment

The authors wish to thank the companies supporting this investigation, Shell International Exploration & Production b.v., B.E.B. Erdgas & Erd"l Gmbh, and Shell Todd Oil Services Ltd., for their permission to present this article; Kopp Gmbh for technical support; and N. Trompe for contributing to the computer simulation effort.

References

1 Wu, H.L., van Spronsen, G., Klaus, E., and Stewart, D.M., "By-pass Pigs for Two-phase Flow Pipelines," 7th International Conference on Multiphase Production, Cannes, June 7-9, 1995.

2. Out, J.M.M., "On the Dynamics of Pig Slug Trains in Gas Pipelines," 12th International Conference on Offshore Mechanics & Arctic Engineering, 1993.

Based on a presentation to the 75th Annual Convention of the Gas Processors Association, Mar. 11-13, Denver.

The Authors