J. N. J. J. Lammers
Shell Research B.V.
Amsterdam
Transportation of wet, CO2-containing gases in pipelines requires special measures to prevent corrosion of the pipe wall.
A relatively new option uses injection of glycol, the presence of which in the water phase mitigates corrosion.
A method has been developed for calculating the phase behavior of glycol in such wet CO2 pipelines. Composition of the glycol-water condensate in the pipeline is one essential in determination of the corrosion rate.
The effect of natural-gas condensate on the mass transfer of glycol and water between liquid glycol-water at the bottom of the pipeline and the gas phase is approached by means of two dispersion criteria.
The composition of glycol-water is shown for the various cases considered for the operation of the 66-km, 36-in. trunklines between the Troll platform off the Norwegian coast and the offshore processing plant on the Troll natural-gas pipeline system.
WET-GAS OPTION
Norske Shell recently completed the conceptual design of a twin, 66-km, 36-in. pipeline system for the transportation of natural gas from the offshore Troll field to the mainland.
Out of several possible transportation modes for the CO2-containing natural gas, the final choice was made in favor of the "wet gas transportation option."
In this option, water-saturated gas is allowed in the pipeline, and drying of the gas is carried out onshore.
Installation of the glycol-treating unit onshore instead of on the platform represents, apart from capital savings, improved operational flexibility.
Transportation of wet, CO2-containing natural gases in pipelines, however, requires special measures to prevent unacceptable corrosion of the steel pipewall. One of these measures is the injection of glycol into the pipeline.
In the presence of glycol, liquids condensing onto the cold pipewall will be relatively rich in glycol. Such water-glycol mixtures have been shown to exhibit reduced corrosion rates compared to pure water.1
The level of corrosion reduction depends on, among other factors, the composition of the liquid glycol-water mixture.
CALCULATION BASIS
Glycol, injected at the upstream end of the pipeline (usually on a platform), is assumed to form a liquid glycol-water phase in equilibrium with the wet natural gas at the pressure and temperature prevailing at the pipeline inlet.
Equilibrium will be promoted during the flow in the vertical riser. (The processes taking place in the riser will not be discussed any further here.)
In the subsequent, more or less horizontal part of the pipeline, the liquid glycol-water will flow along the bottom in case of stratified flow conditions (discussed presently). While flowing further downstream in the pipeline, the fluids will cool down, causing among other things condensation of glycol-water at the top of the pipeline.
It is conceivable that the top glycol-water condensate is somewhat richer in water than the glycol-water at the bottom of the pipeline. When some glycol-water condenses, the composition of the remaining gas will shift to a somewhat higher relative water content because the glycol-water condensate is relatively richer in glycol than the gas.
Because the mass transfer across the bottom liquid-gas interface has a finite rate, the resultant nonequilibrium condition cannot immediately be corrected by mass exchange with the bottom glycol-water.
Condensation somewhat further downstream against a cooler top of the pipeline then results in a glycol-water mixture that is somewhat richer in water than the bottom phase.
In the case of a liquid-hydrocarbon condensate layer covering the bottom liquid glycol-water, the mass transfer across the bottom liquid-gas interface would be further hampered.
Consequently, the water content of the glycol-water condensing against the top of the pipeline would increase as compared with the situation where no such layer is present.
This phenomenon will occur under stratified flow conditions.
In annular and slug flow regimes, no "top-of-the-line" condensation will occur. Moreover, in the slug-flow case, gas and liquid will resume equilibrium because of intense contact.
The assumption of stratified flow represents, therefore, a most severe situation from the point of view of top-of-the-line condensation.
On the basis of this concept, a proprietary computer program was written to calculate at specified positions along a natural-gas pipeline the composition of the bottom and top glycol-water condensates.
The calculation is performed for specified pipeline-operating conditions, using principles for phase equilibrium and mass-transfer calculations.
(See accompanying calculations box.)
Input data are the natural-gas flow rate, gas average mole weight, pipeline diameter and temperature, pressure, and hold-up profile (including glycol-water) along the pipeline as calculated by a multiphase-flow computer program.
The program also calculates the quantity of glycol needed to achieve a specified water content of the bottom glycol-water mixture at the downstream end of the pipeline given the water content of the inlet gas.
The model describing the condensation process is based on a number of simplifying assumptions:1
- The water-glycol mixture condensing at the top of the pipeline is in equilibrium with the vapor phase entering from the previous pipeline section at the local pipe-wall temperature.
This assumption is conservative with respect to predicted corrosion rates because it leads to underprediction of the glycol concentration of the top-of-the-line condensate.
- After condensation, the liquid runs down the wall to mix with the liquid accumulated at the bottom of the line without further exchange of water or glycol with the gas phase.
- The model assumes the rate of evaporation of glycol from the liquids at the pipe bottom into the bulk gas phase and the rate of absorption of water in the glycol mixture at the bottom of the pipe to be completely gas-phase limited.
The interfacial area is calculated from the liquid holdup, assuming stratified smooth flow.
- Vapor-liquid equilibrium calculations for glycol (ethylene glycol, diethylene glycol, or triethylene glycol) and water are based on the principle of equal fugacities of each component in the gas and liquid phases. The calculation of the gas fugacity coefficients has been generalized via the gas density to be applicable to different types of natural gases.
EFFECT ON MASS TRANSFER
When hydrocarbon condensate flows as a stratified layer on top of liquid glycol-water along the bottom of the pipeline, the mass transfer between the glycol-water liquid phase and the gas phase will be hampered because of the very low solubility of water as well as glycol in liquid hydrocarbons.
In the case of a stagnant layer of condensate, the mass transfer will be fully governed by diffusion and therefore become so low that no mass transfer takes place effectively.
Under turbulent conditions in the condensate layer, the mass transfer will be increased, of course, but only to a small extent, again because of the very low solubility of glycol and water.
Only under such conditions of flow that the glycol-water liquid starts to disperse in the condensate can direct exposure of glycol-water droplets to the gas phase occur. The mass transfer would then again become gas-phase limited as in our standard case of full exposure (calculations box).
To estimate the effect of the degree of exposure of bottom liquid glycol-water to the gas phase, we have introduced an "exposure coefficient." This coefficient represents the fraction of the liquid-phase surface area which is occupied by glycol-water and thus available for mass transfer to and from the gas phase.
In order to yield an estimate for the value of the exposure coefficient, the extent to which dispersion might be expected to occur has been investigated.
Three assumptions have been made to estimate the effect of dispersion on exposure:
- In case of full dispersion, exposure of the liquid glycol-water will occur in the volumetric ratio of glycol-water: (condensate + glycol-water); this implies an exposure coefficient equal to that ratio.
- In case of no dispersion, the exposure coefficient will be zero.
- In the intermediate case, exposure will depend on the fraction of glycol-water droplets taken up by the condensate, transported to the liquid-gas interface, and exposed for a certain period of time.
The exposure coefficient from No. 1 should then be multiplied by a reduction factor. It is not the purpose here to present a method which determines the value of such a factor for the intermediate case.
DISPERSION CRITERIA
Criteria for dispersion have been obtained following two different routes: one concerns the breaking-up of two liquid layers into a dispersion; the second route concerns the dispersion of already formed liquid droplets into a second liquid.
TWO LAYERS
The breaking-up of two liquid layers has been approached via a criterion which was developed by Ishii and Grolmes2 for the entrainment of liquid in a gas flow.
When applied to our case, this criterion implies that entrainment occurs when the drag force (Fd) exerted by the continuous liquid (hydrocarbon condensate) on the liquid to be dispersed (glycol-water) exceeds the retaining force of the surface tension (Fs).
For large-film Reynolds numbers and viscosity number Nm < 1/15, the criterion reads as shown in Equation 1 (see equations box).
The viscosity number is defined as shown in Equation 2.
The Reynolds number of the film glycol-water is defined as shown in Equation 3. And the wetted perimeter is shown in Equation 4.
The angle (a) is calculated with Equation A2 (accompanying calculations box) using the glycol-water hold-up. (a is defined in Fig. A1.)
The latter has been obtained from the total hold-up (from multiphase-flow calculations) and the assumption that the ratio of the glycol-water hold-up to the total liquid hold-up equals the ratio of the flows (from multiphase flow calculations and calculation of the bottom glycol-water flow).
The glycol-water hold-up is calculated to be a quarter of the total hold-up.
The value of this factor appears to vary only by 0.05 and is for our approach assumed constant all along the pipeline for the four cases considered.
The value for s has been assumed 0.02 Newtons/m. In Table 1, four different cases have been considered.
SECOND LIQUID
The dispersion of droplets into a second liquid has been approached via a criterion developed by Davies3 for dispersion of solid particles in liquid pipeline flow.
A similar approach was followed by Duckler and Taitel4 in gas-liquid pipe flow for the transition between intermittent flow and dispersed flow of gas bubbles.
Dispersion occurs when the eddy fluctuation force exerted by the liquid exceeds the sedimentation force of the particles.
The sedimentation force (FS) is shown in Equation 5; the eddy fluctuation force (FE) in Equation 6.
The eddy fluctuation velocity (v') is given by Equation 7.
Extension of this theory to liquid dispersions requires determination of the dispersed-phase droplet diameter.
The droplet diameter has been calculated according to Hinze5 (Equation 8).
The value of dmax has been used for the droplet diameter in the various equations. This renders the dispersion criteria conservative.
In our case of three-phase flow, P denotes the power dissipated in the two liquids flowing along the bottom of the pipeline as a result of friction with the pipewall. Equation 9 can be derived.
The wall shear stress tw may be calculated from Equation 10.
With a wall friction factor calculated according to the Blasius equation:
fw = 0.08ReL-0.25
Equation 11 follows from Equations 9 and 10 and in which the wetted-wall area per unit pipe length (aL 1/2da), a being defined and calculated as shown in the calculations box.
The Reynolds number of the liquid layer is defined according to Equation 12 and the wetted perimeter (PL) according to Equation 13.
Results of calculations are presented in Table 2. The value for c is, as before, 0.25. The internal pipeline diameter is 0.854 m.
Both criteria described indicate the likelihood of glycol-water to be dispersed in the condensate.
The exposure coefficient is then determined, according to the first assumption previously described, by the phase ratio, which as mentioned before is approximately 0.25.
It should be noted that the two criteria applied do not account for a homogeneous dispersion. The concentration of glycol-water droplets will therefore probably be greater in the lower part than in the upper part of the condensate layer.
The exposure coefficient according to the first assumption may therefore be somewhat optimistic, especially in case these criteria are only just fulfilled.
On the other hand, our approach assumes a flat interface between gas and liquid. This is a conservative assumption because in reality the interface will be wavy. Moreover, especially in the case of a flow regime close to the stratified-annular flow boundary, the interface will be curved.
The actual interfacial area will therefore be appreciably larger than the flat interface assumed in this approach.
PRODUCTION SCENARIOS
Several gas production scenarios were considered by Norske Shell. A limited number were selected for the present calculations (Table 3).
The pipeline temperature and pressure profiles have been calculated with a multiphase-flow computer program.
Table 3 shows the pipeline inlet conditions for calculation of the water content of the natural gas, the water content of the aqueous condensate at the end of the pipeline as used in the present calculations, and the calculated amount of glycol injected (DEG, 90 wt %, per pipeline).
Results of computer simulations are shown in Figs. 1 and 2 for Cases 1 and 3 with a value of 0.25 for the exposure coefficient.
The results show a gradual increase in water content of the bulk glycol-water up to the specified water content at the end of the pipeline (Table 3), as expected.
The composition of the "top-of-the-line" condensate, however, shows a remarkable course, especially between approximately km 8 and km 22 for both Cases 1 and 3 in which the water content exceeds 90 wt %.
This must be attributed to the reduced mass transfer between the bulk glycol-water at the bottom of the pipeline and the gas phase due to the presence of a layer of condensate.
The effect of coverage of glycol-water by condensate is demonstrated in Fig. 3 on the basis of Case 3 by changing the value of the exposure coefficient. Four different values (0-1, 0.25, 0.5, and 1) of the exposure coefficient (e) have been used.
For e - 1.0, as expected, the compositions are nearly identical to the "bottom-of-the-line" compositions (compare Fig. 2). For e - 0.01 over quite a stretch of pipeline, nearly 100% water condenses at the top of the pipeline.
Further sensitivities have been demonstrated in the following:
- Lowering of the pipeline inlet temperature from 55 to 45 C.
The major effect of this change is to be expected from a lower water content of the inlet gas. Injection of the same amount of glycol as in base Case 3 (1.8 kg/sec) indeed results in a much more favorable composition of the glycol-water condensate (Fig. 4).
- Increase of the amount of injected glycol by 50% (taking 55 C. for inlet temperature as in base Case 3). This also leads to a favorable change in glycol-water "top-of-the-line" condensate composition (Fig. 5), although less than in the case of lower inlet temperature.
REFERENCES
- van Bodegom, L., van Gelder, K., Spaninks, J. A. M., and Thomas, M.J.J. Simon, Corrosion 88, Mar. 21-25, 1988. St. Louis.
- Ishii, M., and Grolmes, M. A., AIChE Journal, Vol. 21, No. 308 (1975).
- Davies, J. T., Chem. Eng. Sci., Vol. 42 (1987), No. 1667.
- Duckler, A. E., and Taitel, Y., in Multiphase Science and Technology, Vol. 2 led. G. F. Hewitt, et al.), Springer, 1986.
- J.O. Hinze, AIChE J. 1, 289 (1955).
- Perry, R. H., and Chilton, C. H., Chemical Engineers Handbook, 5th ed., McGraw-Hill, New York 1973.
- Ambrose, D., and Hall, D.J., J. Chem. Thermodyn., Vol. 13, No. 61 (1981).
- Hales, J. L. et al, J. Chem. Thermodyn., Vol. 13, No. 591 (1981).
- Villamana, M.A., Gonzales, C., and van Ness, H. C., J. Chem. Eng. Data, Vol. 29, No. 427 (1984).
- van Ness, H. C., and Abbott, M. M., Classical Thermodynamics of Non-Electrolyte Solutions, McGraw-Hill, New York 1982, pp. 220-231.
- Gas Conditioning Fact Book, Dow Chemical Co., Midland, Mich., 1962.
- API Technical Data Book, Petrol. Ref. Vol. 2, 4th ed., 1983.
Copyright 1991 Oil & Gas Journal. All Rights Reserved.
- Lowering of the pipeline inlet temperature from 55 to 45 C.
