Flow modeling to reduce CAPEX and OPEX on wet gas pipeline system

Fabien Carimalo, Raphael Hauguel*, Igor Fouché, Thomas Chrétien, Carine Ryszfeld GDF SUEZ, Research and Innovation Division, 361 av du Président Wilson, B.P. 33, 93211 Saint Denis

la plaine cedex, France

*Contact author : raphael.hauguel@gazdefrance.com

© Copyright 2008 IGRC2008

© Copyright 2008 IGRC2008

ABSTRACT

On underground aquifer gas storages, wet gas and liquid water are both carried away by gathering lines from wellheads to gas treatment stations. Internal corrosion of these lines, due to liquid water, is the main time related damaging process. In order to develop corrosion integrity management plans, it is essential to identify the most likely locations of internal corrosion. Water accumulation is one of the most influential factors in assessing the corrosion threat along a pipeline.

In this context, flow models implemented in OLGATM1 software were tested with two field

measurements performed on Gaz de France facilities and gave relevant answers on the understanding of the hydrodynamic in a wet gas gathering line. Moreover, a correspondence between the water accumulation zones predicted by the software and the corrosion zones measured by a detection pig was also clearly highlighted for one pipeline.

The goal of this paper is to illustrate how the multiphase flow software can be

advantageously used for optimizing wet gas pipeline design and increase water management on these gathering lines. Different works were made as the studies of pipelines diameter to minimize the presence of water in respect of pressure losses and gas velocities constraints. Another studies were consisted in compare results from simulations and experimental data concerning corrosion to improve inspections management plans. In parallel the round of manual purges, in respect of normal operation conditions, can be improved by analysing multiphase flow simulations prediction.

Then, by strengthening data field comparisons, OLGA software results will help to reduce

CAPEX and OPEX for the wet gas system sizing and operation strategies.

1 Software trademark

INTRODUCTION

In France, natural gas supply comes from various production companies through specific contracts called “take or pay” which impose customers to take a fixed and constant gas volume. That forces GDF SUEZ Group to strictly manage the amount of gas all over the year even during low periods (i.e. consumption is five times greater in January than in August). Thus, storages constitute a solution for natural gas surplus received during the summer which is stored in order to anticipate the gas supply during high consumption in winter. In order to manage this natural gas surplus GDF SUEZ operates several underground aquifer storages (Figure 1)

Figure 1: Schema of an aquifer storage

Under normal operating conditions, during summertime, natural dry gas is injected by

compressor stations through wells drilled into the ground. During winter, the flow direction is reversed through gathering lines inserted between the storage (wells) and the gas de-hydration treatment1 plants and/or gas compressor stations. In this case, gas phase is saturated with water caused by contact with the lower aquifer. In addition, liquid water is carried out by the flow from the liner through collecting network, to liquid/gas separator2 (also called drips) (Figure 2). Moreover, depending on the temperature and pressure gradient in the pipe, some water may condense. This condensation increases the amount of liquid water arriving at the separators.

These wet gas gathering lines are characterized by large seasonal variations of the gas-to-liquid volume ratio with possible concomitant variations of the water composition (through variations of the condensing water/formation water ratio). These variations also come with seasonal reversal of the flow direction (injection versus retrieval regime), as well as some shutdowns periods between the injection and the retrieval periods. Contrary to gas transmission pipelines where liquid water contamination is episodic (dry gas), water gas ratio called WGR (Water Gas Ratio) is a key parameter which could not be neglected by operators on storage plants.

Gas Transport

Gas Storage in aquifer

Surge points (manual or automatic)

Well part

Surface utilities

Separator

Accumulated water in low points

Collecting network

Central Processing Facilities

Gas Transport

Gas Storage in aquifer

Surge points (manual or automatic)

Well part

Surface utilities

Separator

Accumulated water in low points

Collecting network

Central Processing Facilities

Gas Storage in aquifer

Surge points (manual or automatic)

Well part

Surface utilities

Separator

Accumulated water in low points

Collecting network

Central Processing Facilities

Figure 2: Simplified description of the surface uti lities, from the reservoir to the central processin g facilities

Internal corrosion of these lines walls, enhanced by liquid water and wet gas, is the primary

time related damaging process which may endanger pipe integrity. Besides, locating internally corroded pipe is all the more difficult as it requires internal detection methods. Thus, to optimize wet gas pipeline water management (and subsequent corrosion issues) and decrease water accumulation at low spots, multi-phase flow calculations (water and wet gas) are performed using OLGA software3. This tool was chosen because the implemented closure laws were validated through large-scale laboratory results4,5. Moreover, the models were continually improved through the OVIP(2) project6,7,8,9 sponsored by eleven oil and gas companies.

FLOW MODELING AVAILABLE FOR WET GAS GATHERING LINES : COMPARISON BETWEEN SIMULATIONS AND MEASUREMENT DATA

Two tests were made to evaluate the software efficiency by simulating correctly the multi-phase flows in gathering pipelines. Simulations results were compared with field measurements performed on GDFSUEZ storage (named storage A).

The first test was a progressive increase of the gas flow rate and the second one was an

abrupt increase of the gas flow rate. The results are summarized on the following figures. Because the water volume on drips cannot be easily known (between 40 l and 70l during the draining), this water volume is counted on sites by considering these two limits (Fig. 3 and 4).

2 OLGA Verification and Improvement Program

Test 1 - 15/02/2006 - Storage A

0

2000

4000

6000

8000

10000

12000

14000

0 2 4 6 8 10 12

Time (h)

Gas

Flo

w r

ate

in (

n)m

3/h

0

50

100

150

200

250

300

350

400

450

Wat

er v

olum

e (l)

Gas Flow rate

Water arrival at the separator (hyp=surge volume = 70l)

Water arrival at the separator (hyp : surge volume = 40l)

Water arrival at the separator (OLGA results)

Accumulated water in the gathering line (OLGA results)

Figure 3: TEST 1 - Simulations results compared wit h experimental data. This graphic represents the ga s flow rate and the water volume at the drips versus time

Test 2 - 14/03/2006 - Storage A

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

0 5 10 15 20 25 30 35

Time (h)

Gas

Flo

w r

ate

(n)m

3/h

0

100

200

300

400

500

600

Wat

er v

olum

e (l)

Gas F lo w rate

Water arrival at the separato r (hyp : surge vo lume = 70 l)

Water arrival at the separato r (hyp : surge vo lume = 40 l)

A ccumulated water in the gathering line WGR 2.2 (OLG A results)

Water arrival at the separato r WGR 2.2 (OLGA R esult s)

A ccumulated water in the gathering line WGR 3 (OLGA R esults)

Water arriva l at the separato r WGR 3 (OLGA R esults)

Figure 4: TEST 2 - Simulations results compared wit h experimental data. This graphic represents the ga s flow rate and the water volume at the drips versus time

The test 1 (Fig. 3) confirms that simulations results are in a fairly good accordance with the water arrivals measured in these conditions:

• No water was collected at the separator during the 9 first hours of the scenario

• Starting at 9 hours, the amount of water lies between the two limits given by the hypothesis on the purge volume (40l and 70l)

The test 2 (Fig. 4) shows that:

• For a chosen WGR (Water Gas ratio) of 2.2 g/(n)m3 (n means standard conditions at 0°c and 1 bar) the timing of the water arrivals in the separator corresponds to these predicted on the gathering line as a consequence of the abrupt increase in the gas flow rate. However, the amount of total water arrivals predicted by simulations at the separator for this case was 200 L, which is less than the water arrivals measurements based on a minimum purge volume of 40 L (280 L).

• For a WGR equal to 3 g/(n)m3, the arrivals of water predicted with the separator are closer to reality. In fact, simulation results have predicted a total of 300 L of water upset. This is in between the two estimated water volumes: 280 L and 480 L.

Therefore, the simulation results and the measurements on site are coherent. The dynamics of physics are well retranscribed. The actual ratio ranges between 2,2 and 3 g/(n)m3, which is consistent with the geologist’s estimations (Exploration & Production department of GDF SUEZ).

Conclusions: The two scenarios tested in the field confirmed the accuracy of OLGA

software concerning the hydrodynamic in a wet gas gathering line. These two experimental studies made GDF SUEZ confident in the use of OLGA software to optimize wet gas pipeline water

management10,11.

FLOW MODELING TO IMPROVE PIPELINES INSPECTION PLANS : COMPARISON BETWEEN SIMULATIONS RESULTS AND CORROSIO N MEASUREMENTS

In order to develop corrosion integrity management plans, it is essential to identify the most likely locations of internal corrosion. Water accumulation is one of the most influential factors in assessing the corrosion threat along a pipeline. Thus, if comparisons between multiphase flow simulations (to predict water accumulation) and corrosion default measurements give accurate results, simulations will allow to decrease inspection costs focusing only on critical points (or corroded parts of pipes).

Corrosion measurements along a whole pipe were been made with detection pig on

another aquifer storage named B. The figure 5 represents the corrosion defaults position on the pipeline circumference. These data were compared to simulations results (i.e. water accumulation along the pipeline) for different scenarios of gas production (Fig. 6). Several comparisons were made for different pipelines profiles.

Position of the corrosion defaults on the pipeline circumference

0:00

1:12

2:24

3:36

4:48

6:00

7:12

8:24

9:36

10:48

12:00

0 500 1000 1500 2000 2500 3000 3500

Distance (m)

Pos

ition

of t

he c

orro

sion

def

aults

(h)

Superior generator

Lower generator

Superior generator

Depth of corrosion default

entre 10 et 20 %entre 20 et 30 %> 30 %

Figure 5: Pipe 1 - Position of corrosion defaults on the circum ference. The colour represents the corrosion default depth

Corrosion measurement along pipe 1 versus OLGA resu lts Psep = 80 bar, Q = 7000 (n)m3/h, WGR = 50 g/(n)m3

0

20

40

60

80

100

120

0 500 1000 1500 2000 2500 3000

Distance (m)

Wat

er H

old-

up, C

orro

sion

Dep

th (

%)

0

10

20

30

40

50

60

70

80

90

Profile elevation (m

)

Water Hold-up in shut in conditions (%) Maximum depth of corrosion (%)

Pipe1 profile Manual Purges

Water hold-up in production conditions (%)

Zone 2 Zone 3 Zone 4 Zone 7

Peak of corrosion 1

Peak of corrosion 2

Peak of corrosion 3

Permanent Water Accumulation Zone

Figure 6: Pipe 1 - Corrosion measurements along pipe versus water hold -up predictions (which indicates water and gas distribution across pipe section)

The Figure 6 shows that accumulated water zones predicted by the software correspond to zones where the corrosion defaults are important:

• 3 main peaks of corrosion can be identified. These zones correspond of permanent accumulation of water predicted by simulations

• 3 peaks on 6 (50 % of peaks) are predicted

• the area impacted by high level of corrosion corresponds perfectly where continuous water accumulation is foreseen by software

Conclusions: In case of technical limitations, it will be possible to decrease the costs of inspection by building arguments strengthened by an analysis of various feedbacks to avoid the administrative constraints.

GDF SUEZ METHOD TO IDENTIFY CORROSION RISK ON WET GAS PIPELINES

A study was carried out using OLGA software to identify the risk of corrosion along several pipelines on storage B. This method allows operators to improve inspection management and reduce the costs due to corrosion detection methods.

Three levels of risk have been defined in agreement with the storage operator:

• risk 0 : where there is no accumulation of water,

• risk 1 : where there is water transient accumulation due to gas flow rates, which sweeps the water along the pipeline (maximum gas flow)

• risk 2 : there is constant water accumulation for any operating gas flow rate

Profile and purges location for the pipeline

-10

0

10

20

30

40

50

60

70

80

90

0 500 1000 1500 2000

Pipeline length (m)

Rel

ativ

e al

titud

e (m

)

Risk 0

Manual purges

Automatic purge

Hypothesis

- Maximal gas flow rate = 25024 (n)m3/h- Mean gas flow rate = 18204 (n)m3/h- WGR = 10 g/(n)m3- Pressure at separator level = 55 bar- Temperature at separator level = 25°C

R2

R1 = Risk 1

R2 = Risk 2

R1 R1

Figure 7: Simulation results for one profile. The level of risk can be identified on the pipe. According to the time of wetness in pipes, the type of purges along the profile can be advised

An example of results is shown on the figure 7. This graphic represents a particular topographic profile due to the presence of a hill. Indeed, on these pipelines, water carried out by the wet gas tends to accumulate along the hill. Thus, the risk of corrosion is very high in this area (red part of the profile between the two vertical dotted line on the graph).

The studies show that, at the bottom of the hill, an automatic purge (natural separator) must be installed whereas in top of the hill only manual surges equip the low spots are needed. The installation of an automatic purge at the bottom of the hill allows to reduce seriously water volume in the pipe on the higher part of the hill. Then, the risk of corrosion will decrease seriously.

By considering an automatic purge located at the bottom of the hill with an efficiency of 95%

(reduction of the WGR value in simulations), new simulations were realized to evaluate the number of days where water can be accumulated at low spots downstream the hill. An example of results is shown on figure 8.

Operating conditions (P,Q) - 2004/2005 - Storage B

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

22000

24000

45 50 55 60 65 70 75 80 85 90 95

Well-head pressure (bar)

Gas

flow

rat

e ((

n)m

3/h)

Operating point

Critical flow rate (manual purgesdownstream from the hill)

Water sweeping

Water accumulation

Figure 8: – Operating conditions (P,Q) for season 2004-2005, storage B an d accumulation conditions

In parallel, the time of wetness in the gathering lines can be identified. If this time is short according normal operating conditions, new recommendations can be made to remove manual purges at these low points. Moreover, only one separator can be installed for two or three pipes before gas treatment station.

Conclusions: the inspection strategy of gathering lines was optimized by evaluating a

relative time of wetness in the gathering lines. This work allows to reduce operating costs by decreasing the number of inspections and the number of manual purges along profiles where no water accumulation is predicted. At least, concerning the new pipelines, the number of manual purges can be reduced (lower CAPEX & OPEX).

OPTIMIZING WET GAS PIPELINE DESIGN TO IMPROVE WATER MANAGEMENT

Flow modeling was used to determine the optimum diameter of new pipes on storage B. The goal of this work is to determine the diameter for decreasing water accumulation in pipes according to the operating condition (mass flow rate, water gas ratio, …). Two internal diameters (Fig. 9) were considered in our simulations (DI = 150 mm and DI = 200 mm).

DI = 200 mm

DI = 150 mm

Figure 9: Influence of the internal diameter value on water accumulatio n for a chosen scenario. The blue colour represents the water holdup along the profile

The figure 9 shows that the water volume is low when the diameter of the pipe is equal to

150mm. Parametric studies for three representative profiles have allowed the comparison of the various pressure losses obtained for either DI150 or DI200. The table 1 summarizes the results of multiphase flow simulations for one well and for the worse gas extraction conditions measured on sites

Synthesis DN150 DN200

Max pressure loss (bar) 5.4 5.6 Max gas velocity (m/s) 8.1 5.3 Max water volume (L) 9670 22600

Table 1: Synthesis of the analysis for well 1, Storage B

A DI150 pipeline creates an acceptable pressure loss for all considered scenarios and an acceptable fluid velocity due to acoustic constraints. Moreover, for all the considered cases, in steady state mode, the quantity of water is considerably reduced in the pipeline. This study validates the sizing of the new pipelines in DI150, minimizing the presence of water in the gathering lines.

Conclusions: The reduction of the pipe diameter allows to reduce:

• The risk of corrosion related to the presence of water

• The risk of hydrates formation, because hydrates formation is partly due to the presence of water in the network

• Slug low risk which can damage surface utilities

• The installation of surge points because water is continuously carried away through the pipe

CONCLUSIONS

Two experimental studies on storage A made GDF SUEZ confident in the use of OLGA software to model water gas flow and validated its use to optimize wet gas pipeline water management.

Comparison of corrosion measurement data along profile versus OLGA water accumulation prediction results for one pipe of storage B gave good results. Then, studies were made to evaluate the risk of corrosion on several pipelines. Moreover, the evaluation of the time of wetness allow to determine the type of purges which must be installed on new gathering lines.

At least, the optimum diameter on new pipes can be determined to reduce drastically the amount of water in respect of pressure losses and gas velocities constraints.

Thus, having a good understanding of hydrodynamic flows on collecting network based on OLGA results will help to reduce CAPEX and OPEX for the wet gas system sizing and operation strategies, which is completely new for the operator.

ACKNOWLEDGEMENTS

The work has been conducted under the Research and Innovation Division of GDF SUEZ, which client of the project was the Major Infrastructures Division of GDF SUEZ. The authors express their special thanks to the co-workers from the underground storages facilities.

REFERENCES

1. A. Kohl and F. Riesenfeld, Gas Purification (Houston, TX: Gulf Publishing Co., 1985)

2. T.G. Braga and R.G. Asperger, ‘Engineering Consideration for Corrosion Monitoring of Gas Gathering Pipeline Systems,’ CORROSION/87 Proceedings, NACE International, Houston, TX (1987)

3. Bendiksen, K., Malnes, D., Moe, R., Nuland, S., “The dynamic two Phase Flow Model OLGA : Theory and Application“, SPE Production Engineering, May 1991, 171

4. Shea R., Rasmussen J., Hedne P., Malnes D., “Holdup predictions for wet-gas pipelines compared“, Oil and Gas Journal, May 19 (1997)

5. Alvarez C., Al-Awwami M.H., “Wet crude transport through a complex hilly terrain pipeline network”, 74th Annual technical conference and exhibition, SPE 56463 (Houston, TX, Society of Petroleum Engineers, October 3-6, 1999).

6. Richard H. Shea, OVIP – Comparison of BHRG Pipeline/Riser Data, Scandpower technical note no. 34.010.011/R3

7. Richard H. Shea, Jan Nossen, OVIP 2004-2006 Project,Tulsa Univ.Well data bank – effort to improve predictions , Scandpower technical note no. 34.010.006

8. Nossen, J., Shea, R.H., Rasmussen, J., “New Developments in Flow modeling and field Data Verification“, BHR Group 2000 Multiphase Technology, 209-222

9. Langsoft M., Holm H., “Liquid accumulation in gas-condensate pipelines – an experimental study,”, BHR Group 2007 Multiphase production technology,p33-46

10. Carimalo F, Fouche I, Hauguel R, Campaignolle X, Chrétien T, Meyer M, “Flow modeling to optimize wet gas pipeline water management”, Paper No. 08137, CORROSION 2008 Proceedings, NACE International

11. Hauguel R, Lajoie A, Carimalo F, Campaignolle X, Chrétien T, Meyer M, “Water Accumulation assessment in wet gas pipelines”, Paper No. 08138, CORROSION 2008 Proceedings, NACE International

12. Magnus Nordsveen, Comparison of OLGA 2000 v4.06, v4.12, and 4.16.8 with Field Data , Scandpower technical note no. 34.011.011/R2

ANNEX 1: MODELING TOOL DESCRIPTION

OLGA software is a simulator for the engineering of water, oil and gas flow of oil, in wells, pipelines, and receiving facilities. The OLGA model was first described by Bendiksen et al and is a two-fluid model which solves separate continuity equations for gas, liquid bulk and liquid droplets, two momentum equations for gas, together with possible liquid droplets, and a separate one for liquid film at the wall, and at last one energy-conservation equation for the mixture of gas and liquid. This tool has been chosen because the implemented closure laws12 were validated through large-scale laboratory results and field data. Moreover, the models have been continually improved through the OVIP(3) project sponsored by eleven oil and gas companies.

A fluid characterisation is necessary to fluid flow simulations. The PVTsim TM4 package from Calsep is used within the OLGA Software Suite and provides capabilities for fluid flow with the Peng-Robinson (PR) equation.

3 OLGA Verification and Improvement Program 4 Software trademark