TECHNOLOGY Simulator predicts transient flow for Malaysian subsea pipeline

April 15, 1996
Anis A. Inayat-Hussain, Mohd Sapihie Ayob, Ahmad Bazlee Mat Zain Petronas Research & Scientific Services Sdn. Bhd. Kuala Lumpur In a step towards acquiring in-house capability in multiphase flow technology, Petronas Research & Scientific Services Sdn. Bhd., Kuala Lumpur, has developed two-phase flow simulation software for analyzing slow gas-condensate transient flow. Unlike its general-purpose contemporaries-TACITE, 1 OLGA, 2
Anis A. Inayat-Hussain,
Mohd Sapihie Ayob,
Ahmad Bazlee Mat Zain

Petronas Research & Scientific Services Sdn. Bhd.
Kuala Lumpur

In a step towards acquiring in-house capability in multiphase flow technology, Petronas Research & Scientific Services Sdn. Bhd., Kuala Lumpur, has developed two-phase flow simulation software for analyzing slow gas-condensate transient flow.

Unlike its general-purpose contemporaries-TACITE,1 OLGA,2 Traflow (OGJ, Jan. 3, 1994, p. 42; OGJ, Jan. 10, 1994, p. 52), and PLAC (AEA Technology, U.K.)-ABASs is a dedicated software for slow transient flows generated during pigging operations in the Duyong network, offshore Malaysia. This network links the Duyong and Bekok fields to the onshore gas terminal (OGT) on the east coast of peninsular Malaysia (Fig. 1 [71592 bytes]).

It predicts the steady-state pressure drop vs. flow rates, condensate volume in the network, pigging dynamics including volume of produced slug, and the condensate build-up following pigging.

The predictions of ABASs have been verified against field data obtained from the Duyong network.

Presented here is an overview of the development, verification, and application of the ABASs software. Field data are presented for verification of the software, and several operational scenarios are simulated using the software.

The field data and simulation study documented here will provide software users and developers with a further set of results on which to benchmark their own software and two-phase pipeline operating guidelines.

Multiphase challenges

The complex behavior of multiphase fluid flow continues to pose a major challenge in the optimal design of pipeline networks for gathering and transporting oil and gas.

Traditionally, the hydraulics design of such networks has relied largely on empirical correlations, but the last decade has seen an increase in the application of advanced simulation techniques based on mechanistic models.

The mechanistic approach exploits the fundamental physics of multiphase flow and provides improved predictions and deeper insight into the performance of pipeline networks under diverse flow scenarios. Of particular interest are transient-flow scenarios arising from interruptions to flow, pipeline ruptures, liquid surges, and pigging operations.

Inadequate design for these transient situations can lead to frequent unscheduled shutdowns of the pipeline network and associated processing.

Development of ABASs was commissioned by Petronas Carigali Sdn. Bhd. for use as an operations analysis tool for the Duyong pipeline network.

Gas and condensate produced from the Duyong field are separated and processed at the Duyong central processing platform. The separate streams are then re-combined to flow in a single pipeline to the Sotong collection platform (SCP) which has no processing.

At SCP, the Duyong gas-condensate merges with the gas-condensate stream from the Bekok platform; the resulting mixture flows to the OGT.

The liquid-control facilities in the network include two pig receivers and one pig launcher at SCP as well as liquid receiving facilities at OGT. Pigging is undertaken regularly to remove accumulated condensate from the pipeline network and maintain gas production from the network.

Model's assumptions

The ABASs software is based on a mechanistic model which has been implemented in a Fortran 77 computer program operating in the Unix environment. The input parameters and output variables of ABASs are shown in Table 1 [25016 bytes].

Key assumptions of the model are the following:

  • The two-phase flow in each branch of the pipeline network is one-dimensional.

  • There is no flow reversal.

  • The two-phase flow model is based on the two-fluid formulation within Taitel's simplified transient approach.3 Separate mass and momentum conservation equations are applied to the gas and condensate phases.

In the Taitel approach, the dynamics of the flow are simulated by retaining the time derivative in only the condensate mass-conservation equation. This approximation is strictly valid for high gas-to-condensate ratios and slow transients as, for example, those observed during pigging.

Petronas' simulation studies indicate a range of validity which extends up to a condensate-to-gas ratio of 150 st-tk bbl/MMscf.

  • The correlations used are the usual Fanning friction factor for the friction between the fluids and the pipe wall, the Gregory correlation for the condensate hold-up in the slug ahead of the pig,4 and the Andritsos-Hanratty correlation for the interfacial friction between the gas and condensate.5

These correlations provide the necessary closure relations for the mass and momentum conservation equations.

  • The flow-pattern transition model is as reported in Minami.6

  • The seabed topography factor is taken into account by including the elevation of the outlets of each segment for a given pipe branch in the network.

Table 2 [13608 bytes] gives the pipeline geometry including the maximum elevation difference along each branch.

  • The temperature of the two-phase mixture varies along a pipeline branch.

This temperature profile, assumed steady, is calculated from a heat balance between the rate of change of internal energy of the mixture and the rate of heat transfer by fluid flow in the pipeline, heat conduction through pipe wall, and natural convective heat transfer from outer pipe wall by sea water.

The temperature calculation is de-coupled from the two-phase flow mass and momentum calculations; that is, the energy conservation equation is not solved.

Because of this simplification, the model is not applicable to fast transient processes, such as pipeline blowdown or rupture, which involve the conversion of gas internal energy to mechanical flow energy (Joule-Thomson process).

  • The thermodynamic properties of fluids entering the Duyong-to-SCP and Bekok-to-SCP pipelines are flashed at a range of temperatures and pressures using the PVT package CMGPROP (Computer Modelling Group, Calgary), based on the Peng-Robinson equation of state.

This calculation takes as input the wellstream compositions of the Duyong and Bekok gas-condensates (Table 3 [20079 bytes]) and yields the gas and condensate densities, dynamic viscosities, specific heats, volume percent, and the gas compressibility factor (Z factor). The Bekok composition already includes dry gas from the nearby Seligi gas field.

The properties of the combined Duyong and Bekok fluid mixture in the SCP-to-OGT pipeline are taken as a weighted arithmetic average of the properties of the Duyong and Bekok fluids weighted by the inlet-gas mass flow rates (for the case of gas properties) and by the inlet-condensate mass flow rates for the case of the condensate properties.

  • There is mass transfer between the gas and condensate phases.

  • The pressure and mass flow rate are continuous at SCP.

  • The pigging model is based on Minami's formulation.7

Pigs can be launched independently in each branch of the pipeline network. At any given time, only one pig can be in each branch. The pigs are assumed to be 100% efficient with negligible pressure drop across them.

Verification

A series of simulation studies for verification has been undertaken to check the steady-state flow and pigging dynamics capabilities of the ABASs program.

Table 4 [24001 bytes] is a comparison of ABASs predictions vs. field measurements for the steady-state absolute pressure at the Duyong, Bekok, and Sotong platforms. The errors shown are absolute percent errors.

For the case of the Duyong platform, the predictions agree with the observed pressures to within an average error of about 2%. The observed pressures were recorded for the 6 days of 1994 (Table 4 [24001 bytes]) along with the corresponding operating conditions (Table 5 [21288 bytes]).

It is also interesting to compare the observed pressures against simulations based on a commercial software. Table 6 [20470 bytes] gives the simulation results obtained from the software package PIPE SIM (Windows Release 2.73, 1994, Baker Jardine & Associates) using the Beggs-Brill (B-B) and Baker-Jardine (B-J) correlations; the operating conditions of the pipeline network are the same as previously indicated.

The average prediction errors in this case are clearly greater than for the case of ABASs.

The pigging capabilities of ABASs have also been verified. Table 7 [22775 bytes] gives a comparison between the predictions of ABASs and field data recorded for two pigging operations in December 1994. (Table 8 [15715 bytes] presents the operating conditions of the pipeline network.)

The results in Table 7 [15715 bytes] show that ABASs is able to predict the pigging dynamics in the Duyong network to within an average of 6% of the observed behavior.

Simulation studies

ABASs has been used to analyze various operating scenarios and determine the optimal operating conditions corresponding to maximum gas throughput, minimum condensate hold-up in the SCP-OGT pipeline, and optimal pigging frequencies for the Duyong network.

The results of these simulation studies include the following:

  • The maximum volume of the slug generated during the pigging process is approximately 25% of the total volume of condensate in the pipeline network prior to pigging.

  • Steady-state condensate hold-up volume in the network is recovered about 1.5 days after the pigs are launched.

  • Simulations of the condensate hold-up volume as a function of the combined Duyong and Bekok gas flow rates for a range of OGT pressures demonstrate the existence of an optimal flow rate of gas corresponding to a minimum hold-up volume (Fig. 2 [36830 bytes]).

In these simulations, it is assumed that the total gas flow is split equally across the Duyong-to-SCP and Bekok-to-SCP pipelines, and that the condensate flow rates are proportional to the gas flow rates.

At a total gas flow rate of 440 MMscfd, the condensate flow rates are 26,000 st-tk bbl/day and 2,800 st-tk bbl/day, respectively, for the Duyong and Bekok platforms.

  • For given gas and condensate flow rates into the network, the gas throughput of the network can be increased by about 2% by a suitable choice of the OGT landing pressure. This choice corresponds to a minimum retrograde condensation in the pipeline network.

For future field developments offshore Malaysia, the ABASs program will be significantly enhanced to provide solutions to a broader range of pipeline transport problems.

These enhancements will involve development of robust models and numerical schemes for handling liquid transients in gas trunklines, preferential segregation of phases during the merging of pipeline flows at manifolds, two-phase waxy crude flows, as well as flows along steep terrain.

Acknowledgments

The authors wish to thank Ikhlas Hj. Abdul Rahman, Petronas Carigali Sdn. Bhd., for his support of this work and Nasir Hj. Darman for assistance with the CMGPROP simulations. Permission to publish this article from Petronas Research & Scientific Services is acknowledged.

References

1. Pauchon, C.L., Dhulesia, H., Cirlot, G.B., and Fabre, J., "TACITE: A Transient Tool for Multiphase Pipeline and Well Simulation," SPE Annual Technical Conference and Exhibition, New Orleans, Sept. 25-28, 1994.

2. Bendiksen, K.H., Malnes, D., Moe, R., and Nuland, S., "The Dynamic Two-Fluid Model OLGA: Theory and Application," SPE Production Engineering, May 1991.

3. Taitel, Y., Shoham, O., and Brill, J.P., "Simplified Transient Solution and Simulation of Two-Phase Flow in Pipelines," Chemical Engineering Science, Vol. 44 (1989), pp. 1353-59.

4. Gregory, G.A., Nicholson, M.K., and Aziz, K., "Correlation of the Liquid Volume Fraction in the Slug for Horizontal Gas-Liquid Slug Flow," Int. J. Multiphase Flow, Vol. 4 (1978), pp. 33-39.

5. Andritsos, N., and Hanratty, T.J., "Influence of Interfacial Waves in Stratified Gas-Liquid Flows," AIChE J., Vol. 33 (1987), pp. 444-54.

6. Minami, K., and Shoham, O. "Transient Two-Phase Flow Behaviour in Pipelines - Experiment and Modelling," Int. J. Multiphase Flow, Vol. 20 (1994), pp. 739-52.

7. Minami, K., and Shoham, O., "Pigging Dynamics in Two-Phase Flow Pipelines: Experiment and Modelling," SPE 26568, SPE Annual Technical Conference and Exhibition, Houston, Oct. 3-6, 1993.

The Authors

Anis A. Inayat-Hussain is advisor for development engineering with Petronas Research and an adjunct professor in chemical and natural resources engineering at Universiti Teknologi Malaysia. He is currently on secondment from BHP Research where he was head of the Applied Mechanics Group.

Inayat-Hussain holds a BS with joint first class honors in physics and mathematics and a PhD in theoretical physics from the University of Western Australia, Perth.

Mohd Sapihie Ayob is senior research engineer with the development engineering group at Petronas Research. He holds BS and MS degrees in continuum mechanics from Michigan Technological University, Houghton.
Ahmad Bazlee Mat Zain is senior research engineer with the development engineering group at Petronas Research. He holds a BS in chemical engineering from the University of Toledo, Toledo, Ohio, and is currently on study leave to pursue an MS in petroleum engineering at the University of Tulsa.

Copyright 1996 Oil & Gas Journal. All Rights Reserved.