Graeme King
SNC Partec Inc.
Calgary
Use of ultra-high pressure gas pipelines (2,170 psi) to transport intermediate weight hydrocarbons (C2, C3, and C4) from Canadian-Arctic oil and gas reservoirs will solve several problems inherent in conventional-pressure pipeline systems.
At normal pipeline pressures, such hydrocarbons cause vapor problems if they are included in the oil stream and liquids problems if they remain in the gas stream.
At ultra-high pressures, however, they can be commingled and transported successfully in a single dense phase. This eliminates the need for a separate gas-liquids pipeline or for re-injection facilities.
These pipeline systems also require smaller diameters than conventional gas pipelines carrying an equivalent amount of gas. This makes them more flexible and better able to conform elastically to differential movements from thaw settlement and frost heave.
They also have thicker walls which make them better able to resist ovalling and buckling under extreme bending stresses.
More importantly, however, at ultra-high pressures, natural gas flows at almost constant temperature and exhibits no marked Joule-Thomson cooling effect characteristic of conventional gas pipelines.
Ultra-high pressure gas pipelines can therefore be designed to flow at or near ground temperature as does the Norman Wells oil pipeline, and advantage can be taken of the experience gained from its successful operation in permafrost terrain.
Physical comparisons show an ultra-high pressure pipeline carrying gas with molecular weight in the range 20 to 40 would use a similar amount of steel as an equivalent conventional-pressure gas pipeline with a separate gas-liquids line.
With their better environmental acceptability and durability, ultra-high pressure pipelines deserve serious consideration for transporting natural gas from the Arctic.
The concept has been developed in an attempt to satisfy environmental, engineering, and economic constraints of Arctic regions. It arose out of earlier work on the dense-phase transmission of natural gas at conventional pressures and low temperatures near -80 C.1
CANADIAN ARCTIC
Exploratory drilling has found significant reserves of natural gas and gas liquids in the Canadian Arctic. The Geological Survey of Canada estimates that 330 billion cu m (11.7 tcf) of natural gas have been found so far in the Mackenzie Delta-Beaufort Sea area with the expectation that another 1.6 trillion cu m remain undiscovered.2
Gulf Canada Resources estimates that 30% of its natural-gas reserves is made up of associated and solution gases.2 These gases contain intermediate-weight hydrocarbons (ethane, propane, and butane) which are difficult, to carry in significant quantities at normal pipeline pressures because of vapor problems if they are included in the oil stream and liquids problems if they are included in the gas stream.
Pipeline proposals for transporting natural gas from the Canadian Arctic have envisioned operation at temperatures just below freezing to prevent high ice content permafrost, through which the pipelines must pass, from melting and causing surface erosion as well as differential movements that could cause pipeline failures. 3
Operation at sub-freezing temperatures and at conventional pipeline pressures requires relatively lean gases to prevent liquids from forming and causing troublesome two-phase flow in the pipeline.
A recent application to Canada's National Energy Board (NEB) for a natural gas export license, provides a lean-gas composition that would be suitable for transportation in a conventional high-pressure gas pipeline from the Canadian Arctic, along a corridor up the Mackenzie Valley to connect with the existing pipeline system in Alberta.
The lean gas composition (Mixture 1) is given on Table 1. It consists largely of methane with only small amounts of ethane and practically no quantities of propane or heavier components. A gas-liquids composition (Mixture 5) made up of ethane, propane, butane, and pentane which cannot be carried in the lean-gas pipeline or by crude-oil pipeline or tanker is also given on Table 1.
In determining the composition of Mixture 5, it was assumed that a cut would be made in the middle of the pentanes, with components heavier than pentane going to a crude oil pipeline or tanker and components lighter than pentane going to the ultra-high pressure gas pipeline.
Precise gas compositions will remain a matter for speculation until pipelines are actually built and reservoirs are fully explored, but at this time these two compositions represent hydrocarbon streams that would be available from the Arctic.
Mixtures of the two streams are also given in Table 1. The exact ratio of the two streams that will be obtained is unknown at this time and depends not only on the composition of gas and oil wells in the Arctic, some of which have not yet been discovered, but also on the ratio of natural gas to crude oil that ultimately will be produced.
PROPERTIES
Phase envelopes for the five mixtures are shown in Fig. 1. The familial behavior of light and intermediate-weight hydrocarbons is easily recognized. The cricondentherms of the mixtures range from -20 C. to 150 C., and the cricondenbars range from 4.5 MPa to 14 MPa (650-2,020 psi).
At conventional pipeline pressures in the range 6 to 10 MPa and at sub-freezing temperatures, Mixture 1 is transported as a gas and Mixture 5 is transported as a liquid. At 0 C., Mixtures 2, 3, and 4 cannot be carried at conventional pressures because of liquids problems.
At 0 C., Mixtures 2 and 3 can be successfully transported in a single phase only at pressures greater than 14 MPa. Mixture 4 can be transported in a single phase at pressures greater than 9 MPa.
An ultra-high pressure natural gas pipeline can therefore provide a means of transporting practically almost any quantity of ethane, propane, and butane in a single-dense phase.
The five hydrocarbon mixtures range from a lean hydrocarbon gas with a molecular weight of 17.6 (Mixture 1) through progressively richer gases containing increasing quantities of gas liquids, to a pure gas-liquids stream with a molecular weight of 57.0 (Mixture 5). Prediction of pipeline performance with these mixtures therefore requires knowledge of the behavior of mixtures of gases and liquids.
One of the most accurate techniques for predicting properties of gases and liquids over a wide range of pressures and temperatures is the Benedict-Webb-Rubin-Starling (BWRS) equation of state using the generalized correlation parameters developed by Starling and Han.5
Thermodynamically consistent values for density, Joule-Thomson coefficient, and isentropic temperature exponent at 0 C. for the five mixtures were developed with the BWRS equation of state, and Starling and Han's generalized correlation are shown in Figs. 2 and 3, respectively.
The densities of Fig. 2 show that Mixture 1 behaves as a gas. Its density at 10 MPa is approximately 100 kg/cu m; at 20 MPa, approximately 200 kg/cu m.
Nature 5 behaves as a liquid at pipeline pressures. At 0 C., its density remains constant at approximately 600 kg/cu m. The other three mixtures exhibit a uniform transition in behavior from gas to liquid.
The temperature of Arctic pipelines is important because of frost heave and thaw-settlement problems associated with permafrost terrain. Pipeline temperature profiles and temperature elevations in pumps and compressors need to be predicted accurately, and refrigeration units need to be sized properly to return the gas to the pipeline at subfreezing temperatures.
For those reasons, temperature-related properties such as Joule-Thomson (J-T) coefficient and isentropic temperature exponent need to be calculated accurately and consistently. The J-T coefficient is defined thermodynamically by Equation 1 (see equations box).
The J-T coefficients shown in Fig. 3 reveal the pressure-sensitive temperature behavior of lean gases such as Mixture 1 at 0 C. and conventional pipeline pressures. In the range 6 to 10 MPa, its J-T coefficient is approximately 8 K./MPa as a result of expansion cooling effects.
A typical system with a pressure loss of 2 MPa between suction and discharge conditions could experience a temperature drop as high as 16 C. between stations.
Mixture 5 has a J-T coefficient that is typical of a liquid. It heats slightly at 0.51 K./MPa because of frictional dissipation of pressure energy as it flows along the pipeline. Its flowing temperature therefore can conform closely to ground temperature.
Mixtures 3 and 4 have J-T coefficients close to zero and can flow at almost constant temperature.
ENVIRONMENT
The approximate extent of permafrost zones in Canada is shown in Fig. 4. Pipelines from the Mackenzie Delta and Beaufort Sea in the Canadian Arctic must traverse more than 1,500 km (930 miles) of continuous and discontinuous permafrost before connecting with existing pipeline systems in Alberta.
Permafrost terrain provides unusual engineering challenges such as thaw settlement and frost heave which can cause large differential movements over short distances and can overstrain buried pipelines.
Furthermore, Arctic environments tend to be fragile. Clearing ground cover to bury pipelines can have far-reaching effects on the thermal regime of the subsoil, leading to permafrost regression, and severe surface erosion unless the right-of-way is carefully revegetated.
Buried-pipeline designs for Arctic regions therefore must incorporate concepts which can cope with the difficulties encountered along the route and which can reduce environmental disturbances.
The large J-T coefficient of lean gases such as Mixture 1 at 0 C. causes problems for buried gas pipelines in permafrost terrain. If station discharge temperatures are maintained at just below freezing to preserve the permafrost, station suction temperatures at the next downstream station can fall to -15 C.
This can cause ice lensing and frost heave problems, particularly at the thousands of interfaces between frozen and unfrozen ground that are encountered along pipeline routes from the Canadian Arctic. Fig. 5 shows how frost heave can oval and buckle the pipe at these interfaces.
Insulation has been proposed to reduce the severity of frost-heave problems at interfaces between frozen and unfrozen ground, but it only delays the onset of differential heave movements. 7 After differential movement begins, the insulation can be crushed at the interfaces, reducing its effectiveness and increasing the rate of frost-heave.
The long-term effectiveness of insulation in preventing potentially harmful frost heave is therefore uncertain, but high-strength insulation systems which are tough enough to resist the forces associated with differential movements along the pipeline could be developed. Heat tracing has also been proposed to reduce the severity of frost heave around buried pipelines operating at sub-freezing temperatures. 7 The cost of high-strength insulation and heat tracing has not been considered in the comparisons that are made later in this work.
Fig. 3 shows that the value of the J-T coefficient decreases at higher pressures and for richer gases, For example, Mixture 1 has a J-T coefficient of 2.5 K./MPa at 20 MPa compared with 8 K./MPa at 8 MPa, and Mixture 3 has a value of only 0.5 K./MPa at 20 MPa.
Ultra-high pressure pipelines carrying rich gases therefore flow at temperatures that conform more closely to ground temperature and are less likely to cause frost heave or thaw settlement problems than conventional pipelines carrying lean gas. They would cause less environmental disturbance on this account and, because they can carry gas liquids, would eliminate the need for a separate gas-liquids pipeline, further reducing environmental disturbance.
DESIGN
Pipeline design needs to be on common bases to permit fair comparisons between alternatives.
The formulas and assumptions shown in the accompanying box provide a physically realistic description of pipeline and compressor performance and were used consistently for all systems. Table 2 shows the parameters and values for the compared systems.
Pressure loss along a section of pipe was calculated from Equation 3.
Velocity (v) in Equation 3 can be replaced by mass flow rate using Equation 4.
All systems that were analyzed had high Reynolds numbers so that frictional effects followed the rough-pipe law evident in Equation 5.
Power requirements of pumps and compressors were calculated from Equation 6.
As a rough rule of thumb for a wide variety of systems, optimum power per unit length of a pipeline is approximately proportional to mass flow rate (Equation 7).
The constant of proportionality (Cf), termed here the "pipeline power factor," has the dimensions of force per unit mass and provides a measure of the force applied by pumps or compressors to propel fluid along the pipeline.
Its optimum value is slightly affected by density, mass flow rate, and pipe diameter as well as by relative costs of fuel, station equipment, and pipe. For scoping studies of this kind, it is convenient to fix the pipeline power factor at a constant value for all cases.
For this article, a value of 0.25 N/kg was chosen based on results of detailed optimization of large-diameter Arctic natural-gas pipelines 3 and on experience with the optimization of liquid pipelines.
The pressure differential (DELTA P) between suction pressure and discharge pressure can be obtained by integrating Equation 3 and adding station losses (Equation 8).
Pipeline design pressure can then be found from the average pressure (P) with Equation 9.
Because diameter (D) in these equations is the inside diameter, the Barlow formula for wall thickness needs to be modified slightly, as shown in Equation 10.
Finally, pipe mass per unit length can be found from Equation 1.1.
Equations 3 through 11 were solved for pipe inside diameter (D), design pressure (PD), wall thickness (t), and pipe mass (M) for a range of values of average pressure (P) and mass flow rate (m) with values for other parameters specified in Table 2.
The relationship between pipe inside diameter and design pressure of systems carrying 500 kg/sec for the five different mixtures set out on Table 1 are shown in Fig. 6.
As anticipated, pipe diameters for Mixture 1 decrease with increasing pressure because Mixture 1 behaves as a gas behaves at pipeline pressures, Pipe diameters for Mixture 5 are unaffected by pressure because Mixture 5 behaves as a liquid behaves.
Pipe diameters for Mixtures 2, 3, and 4 lie between Mixtures 1 and 5.
Wall thicknesses for systems carrying 500 and 1,000 kg/sec are given in Fig. 7, respectively. Because pipe diameters for Mixture 1 decrease with increasing pressure, wall thicknesses for Mixture 1 do not increase in direct proportion to design pressure.
For Mixture 5, pipe diameter remains constant for any pressure, and wall thickness increases in direct proportion to the pressure. Wall thicknesses for Mixtures 2, 3, and 4 follow a smooth gradation between Mixture 1 and Mixture 5.
COMPARISONS
It has already been established by a study of thermodynamic properties that the temperature of ultra-high pressure pipelines carrying rich gases can conform more closely to ground temperature, reducing frost heave and thaw-settlement problems.
Additionally, higher fluid densities and pressures lead to smaller diameter pipes with thicker walls which are better able to handle differential frost heave and thaw-settlement movements.
The question to be resolved at this stage is whether gas pipelines designed to operate at ultra-high pressures can compete economically with pipelines designed to operate at conventional pressures.
Distinguishing between similar systems in the early stages of concept development requires comparing them physically. For this work, all systems have the same distance between stations and the same power requirements per unit mass of gas delivered. This means that systems delivering the same quantity of gas can be compared in terms of number of pipes, pipe diameter, wall thickness, and pipe mass.
The rationale for comparing systems on a mass-flow basis is that the constituent components of oil and gas have similar specific heating values (Table 3). Therefore, systems which have the same mass flow rate are comparable in terms of the amount of energy they deliver.
Although some components of natural gas are more valuable than others to the petrochemical industry, such considerations are beyond the scope of this study.
Ultra-high pressure pipelines carrying 500 kg/sec were compared with conventional-pressure systems carrying the same quantity of each hydrocarbon component.
A design pressure of 20 MPa was chosen for the ultra-high pressure pipelines to provide a sufficient margin against forming liquids in the two-phase region. A design pressure of 10 MPa was chosen for the conventional pressure systems, requiring two pipelines to avoid two-phase flow-a gas pipeline for Mixture 1 and a liquids pipeline for Mixture 5.
Details of the ultra-high pressure pipelines and each separate pipeline of the two-pipe conventional-pressure systems are given in Table 4. Comparisons of pipe mass required for the ultra-high pressure pipelines and the conventional pressure two-pipe systems are shown in Fig. 8.
Wall thicknesses for conventional-pressure pipelines given in Table 4 and used as the basis for comparison in Fig. 8 are too small to handle the severe differential movements that could occur in the Arctic.
In practice, wall thicknesses would be increased to reduce the pipe diameter-W.T. ratio. A lower strength steel could then be used but pipe mass and pipe costs for conventional-pressure systems would increase.
The comparisons on the basis of pipe mass here, therefore, favor conventional-pressure systems.
A number of observations can still be made from these comparisons.
As an example of their economy, first of all, an ultra-high pressure pipeline carrying Mixture 3 would use 30% less pipe steel than a single conventional-pressure pipeline carrying an equivalent amount of Mixture 1.
Secondly, an ultra-high pressure pipeline carrying either Mixture 2, 3, or 4 would use a similar amount of pipe steel as a two-pipe, conventional-pressure system carrying Mixture 1 and Mixture 3 in the same proportion as the single ultrahigh pressure pipeline.
The conventional-pressure system with two pipes instead of one would also have an additional pump and driver at each station. It would therefore cost more to construct and cause greater environmental disturbance even though it used a similar amount of pipe steel and station power.
Furthermore, the conventional-pressure system would have an exaggerated J-T effect which would aggravate frost heave and thaw-settlement problems and would result in higher costs for mitigative measures.
Finally, in the unlikely event that only lean gas is available and no advantage can be taken of the volumetric shrinkage effect of commingling lean gas with gas liquids, an ultra-high pressure lean-gas pipeline would use no more than 20% more steel than a conventional pressure lean-gas pipeline.
In this case a decision would have to be made whether the advantages of having a more robust pipeline with lower environmental impact would counterbalance the increased cost of pipe steel,
ACKNOWLEDGMENTS
The author would like to acknowledge the assistance of SNC Partec Inc., Tensor Engineering Ltd., and Alex Huddleston in developing this article.
REFERENCES
- Katz, D. L., and King, G. G., "Dense Phase Transmission of Natural Gas," Energy Process/Canada, December 1973.
- Gilbertson, G. "Mackenzie Delta/Beaufort Sea," Oilweek, Vol. 40, No. 38, Oct. 30, 1989.
- King, G. G., "Cooling Arctic Pipelines Can Increase Flow, Avoid Thaw," OGJ, Aug. 15, 1977, p. 38.
- Esso Resources Canada Ltd., NEB Application for a License to Export Natural Gas from the Mackenzie Delta, September 1988.
- Starling, K. E., Fluid Thermodynamic Properties for Light Hydrocarbon S),stems; Gulf, Houston, 1973.
- Brown, R. J. E., Permafrost in Canada, National Research Council (Toronto, University Press, 1970).
- King, G. G., "Permafrost, Pipeline Design Considerations," Encyclopedia of Chemical Processing and Design, Vol. 34 (New York: Marcel Decker, 1990).
- Gas Processors Association (Eds.), SI Engineering Data Book, GPSA: Tulsa, 1990.
Copyright 1992 Oil & Gas Journal. All Rights Reserved.
