DANISH PROJECTS REVEAL CHARACTERISTICS AND LIMITS OF GAS CAVERN STORAGE
Jesper Magtengaard
Dansk Olie OG Naturgas AS
Horsholm, Denmark
Danish projects have revealed the characteristics of underground gas storage caverns, their limits, and some uncertainties.
The subsurface part of such a storage facility takes the longest time to develop and accounts for most of the uncertainty and risk involved in an underground storage project.
These are some of the conclusions reached by Dansk Olie OG Naturgas AS (DONG) as a result of ongoing research into underground gas storage.
The research also revealed that the productivity of an aquifer gas-storage cavern is uncertain until late in the project when large-scale withdrawal tests have been conducted. And the most important limitation on initial capacity is the risk of plugging as a result of hydrate formation caused by cooling of the gas in the cavern during high rates of withdrawal.
Several factors limit aquifer storages, the most important being the maximum storage pressure and water production.
GAS STORAGE IN DENMARK
The Danish natural-gas transmission system (Fig. 1) was established between 1979 and 1984.1 Commercial and strategic considerations regarding gas-storage capacity in the system are described in Reference 2.
Currently, about 3 billion standard cu m/year (bscm; approximately 105.9 bscf) are transported through the system. About two thirds of the gas are delivered to the Danish market and the remaining is exported to Sweden and Germany.
Storage capacity for peak-shaving and emergency supply is currently provided by the Torup cavern storage, which is located in the northern part of Denmark (Fig. 1).
At Torup, the storage caverns have been leached in one of Denmark's 16 salt domes Denmark (Fig. 2). The salt was deposited during the Zechstein geological era.
In 1978, the Torup salt dome was selected for detailed storage investigations because of favorable geological conditions and because the nearby fjord could be used as a fresh water source as well as for brine disposal.
During 1978-81, eight wells were drilled by Dansk Olie OG Naturgas AS in order to evaluate the suitability of the dome for gas storage. The maximum depth of the wells was approximately 1,750 m (approximately 5,775 ft).
The wells indicated that suitable salt was present in the depth range of about 1,000-1,600 m.
The dome has an areal extent of approximately 24 sq km (approximately 15 sq miles), and the vertical extension is close to 4 km.
Based on the encouraging drilling results, it was decided in 1981 to develop gas storage in the Tostrup dome consisting of six caverns with a total working gas capacity of 300 million standard cu m (approximately 10.6 bscf).
Key data for the storage are given in Table 1.
All surface facilities have been constructed, and five of the caverns have been leached and gas filled. The last cavern has been leached and is currently being gas filled.
As part of a continuing exploration for underground gas storage, investigations were conducted between 1981 and 1985 on a geological structure near the town of Tonder in southern Denmark.
The potential storage structure is an anticline with a vertical closure of more than 300 m and an areal extent of about 60 sq km.
Two potential storage zones of Triassic geological age have been investigated: a nitrogen gas reservoir at 1,600 m and an aquifer at 1,800 m.
Based on the studies the following conclusions were made:
- Up to 230 million standard cu m (MMscm) of natural gas can be stored in the nitrogen reservoir. About 60% of this volume can be cycled every year without serious gas mixing problems.
- Several billion standard cu m of natural gas can be stored in the aquifer. About 50% of the gas is expected to be working gas.
The exploration and appraisal phases of the Tonder project have been successfully completed and the storage is ready for development, should the need arise.
Fig. 1 shows the location for the Stenlille aquifer storage project.
The potential storage structure is an anticline with a vertical closure of about 35 m and an areal extent of about 14 sq km (Fig. 3).
The potential storage zone is the Gassum formation, a Triassic sandstone aquifer. The depth of the formation is some 1,500 m; the thickness, 120-140 m.
Storage investigations at Stenlille were initiated in 1979 and completed in 1991. The studies were conducted between 1981 and 1987.
The investigation comprised seismic surveys, drilling and testing of seven wells, a long-term gas injection test, and a short-term withdrawal test.
The outcome of the project in Stenlille has been positive. The tightness of the caprock has been proved. The presence of a permeable and porous potential storage zone has been confirmed. The size and shape of the structure have been defined.
Gas injectivity is sufficiently high to allow the development of a large storage within a few years. The gas withdrawal test indicated high well productivity.
The objective of the first phase of the project is to establish storage with a peak production capacity of 300 Mscm/hr and a working gas volume of 300 MMscm.
Later, the storage may be extended considerably to satisfy a growing demand for storage capacity if needed. The total potential for gas storage in the Gassum formation of the Stenlille structure is estimated at 3 bscm of which 1-1.5 bscm is expected to be working gas.
A potential for storage of additional gas volumes exists in layers above and below the Gassum formation.
The three storage projects described, Torup, Tonder, and Stenlille are not the only possibilities for underground storages in Denmark.
Geophysical surveys and well drilling carried out in connection with exploration for oil and gas have indicated the existence of a number of salt domes and anticlinal structures which hold potential for large underground storage of gas.
OPERATIONAL LIMITATIONS
The theoretical maximum working volume of a gas cavern is determined by the maximum and minimum allowed pressures in the cavern.
The maximum pressure is dictated by cavern integrity constraints and, therefore, cannot exceed the overburden stress less some safety factor. The minimum cavern pressure is dictated by rock mechanical constraints in order to eliminate excessive creep.
At Torup, a minimum pressure of 80 bar (approximately 1,160 psi) has been established as a "safe" value.
The pressure reduction necessary to withdraw gas from the caverns is associated with a temperature drop caused by gas expansion. Depending on the content of water in the gas, hydrates may start to form when the gas temperature approaches the water dewpoint temperature (Fig. 4).
Inhibitor injection at the wellheads is available to prevent hydrate formation in the surface installations, but it cannot prevent hydrate formation in the well bores. For new wet caverns such as the Torup caverns, hydrate formation is the primary limitation to long-term gas availability.
Because the temperature drop per unit pressure decrease is lower for lower withdrawal rates due to the heat influx from the surrounding formations, it is possible to withdraw greater volumes of gas at lower withdrawal rates.
The gas withdrawal rate is primarily limited by the surface processing system. The Torup plant is designed to deliver 7.2 MMscmd, comparable with the capacity of the pipeline from the plant.
In addition to these constraints, the gas velocity is limited to 40 m/sec. This, however, does not seriously affect deliverability.
The main operational limitations for the Stenlille aquifer storage are listed in Table 2.
The storage pressure is very important for both the deliverability and the working gas volume. A high pressure is favorable, but in practice the storage will have to be operated between a minimum and a maximum pressure.
The minimum pressure is determined by the required inlet pressure to the pipeline system. The maximum pressure is governed by the reservoir conditions, for instance the strength of the caprock.
As illustrated in Fig. 5, the capacity of the process facilities (Fig. 6) is the limiting factor for the storage deliverability during the first part of the withdrawal period. During the last part of the period, the well capacity is the limiting factor.
The gas-flow velocities during withdrawal will be kept to less than 15 m/sec in order to avoid sand erosion problems. This sets a maximum limit to the withdrawal rate from each well.
Hydrates are not expected to be a problem for the Stenlille storage because of the relatively large depth and high temperature (50 C.) of the storage reservoir. Due to the large heat capacity of the reservoir, the temperature of the gas in the reservoir will stay constant during pressure depletion.
The gas temperature at the bottom of the wells is therefore expected to be constant and close to the reservoir temperature.
Sand production may set an upper limit to the withdrawal rate from the Stenlille wells. However, if sand production proves to be severe, the wells can be gravel packed which will probably solve the problem.
At the end of the withdrawal period, water production will most likely limit the well deliverability and, consequently, also the working gas volume. Water production has several negative effects, including the following:
- Large pressure drop in the reservoir and in the well tubing because of two-phase flow
- Low well productivity
- Problems with disposal of salty formation water.
Water production rates are difficult to predict. Reservoir simulation models can be used to compute predictions, but actual injection/withdrawal history is needed to calibrate the models and verify the predictions.
STORAGE-CAPACITY TESTS
Since 1989, DONG has carried out extensive programs to gain understanding of the hydrate-formation problem at the Torup cavern storage facility.
DONG is currently using the data and knowledge from these programs to limit hydrate formation by operational means and by reducing the amount of water in the caverns.
In 1989, DONG designed a test to produce a cavern for the entire working volume at high rate. During this withdrawal test, pressures and temperatures at several locations including the cavern itself were monitored and retrieved for later analysis.
In addition, the water-vapor content of the gas was monitored to enable an analysis of any hydrate formation which might occur during the withdrawal period because it was believed that hydrates would form when the gas reached its dewpoint at the prevailing pressure.
Inhibitor injection was omitted to ensure that any hydrate formation would occur in the surface installations and not in the wellbore.
Hydrates were in fact found to block the filters and the pressure reduction valve after 13 days of withdrawal, at which time approximately 50% of the expected working volume had been withdrawn.
After installation of new filter elements, withdrawal was continued until severe hydrate problems stopped the test. At this point, 70% of the expected working volume had been withdrawn.
Subsequent analysis showed that an abnormal pressure buildup in the surface flow lines began after 9 days of production. The water dewpoint of the gas was calculated from published correlations based on the water vapor content and the gas pressure.
The gas never reached the dewpoint according to these calculations (that is, it was never saturated with water vapor). But after 9 days, the dewpoint temperature stabilized at 5-6 C. below the continually decreasing gas temperature.
Sloan revealed that hydrates may form directly from the water vapor phase at low water contents .4
This is in fact what is assumed to have happened during the 1989 test. The temperature drop, due to gas expansion, during withdrawal from a cavern is a function of the withdrawal rate.
Recordings from the 1990 withdrawal test (Fig. 4) indicate three distinct periods. A to B is the initial heating of the gas from ambient temperature to flowing gas temperature.
B to C is the pressure/temperature relation for a withdrawal rate of 50 Mscm/hr. This curve is markedly steeper, showing a lower temperature drop per unit pressure decrease than the C to D part, which corresponds to a withdrawal rate of 100 Mscm/hr.
This relationship shows that more gas can be withdrawn from a wet cavern at low rates than at high rates. In a dry cavern, where the dewpoint is never reached, the working volume is practically independent of withdrawal rate.
Two other caverns have also been long-term tested. The results of these tests confirm these findings, showing that wet gas in a newly leached cold cavern will form hydrates which, in turn, constrain the volume that can be produced from the cavern.
TESTS AT STENLILLE AQUIFER
The design capacity curve for Stenlille is shown in Fig. 5. This curve represents the ultimate Phase 1 capacity of the storage and is expected to be reached about the year 2000.
Between 1991 and 1994, the permanent injection-withdrawal facilities will be designed and constructed, and gas will be injected into the storage by a temporary compressor plant.
From 1994 to 1997, gas will be withdrawn during the winter, but more gas will be injected during the summer to bring gas inventory up to the design volume.
After 1997, gas cycling will continue as needed.
The following tests of well-withdrawal capacity are planned:
- Short-term pressure falloff tests during injection to evaluate gas permeability and skin of each well.
- Short-term withdrawal tests of selected wells to evaluate productivity and possible problems with sand or water production. These tests will be carried out in 1991-94, that is, before the permanent withdrawal facilities are operational, which severely limits the duration of the test.
- Long-term withdrawal tests between 1994 and 2000 of selected wells and the entire field to evaluate water production and overall field performance. These tests will be carried out during the winter with the permanent withdrawal facilities.
It is planned to cycle 1020% of the inventory every year.
This should be compared with the ultimate goal which is to be able to withdraw 40% of gas in-place in case of an emergency.
Between 1991 and 2000, a reservoir simulation model will be used to predict the deliverability and working gas volume of the storage. The model will be updated continually and refined as more test data become available.
Reliability of model predictions is expected to increase with time as shown in Fig. 5.
The short-term tests between 1991 and 1994 will only give exact information about the first few days of withdrawal.
The model predictions of long-term performance of the storage wells will be rather uncertain.
Cycling 10-20% of the inventory between 1994 and 2000 will provide exact information about a much larger part of the capacity curve. Furthermore, the practical experience gained during the cycling operations will be used to verify the reservoir simulation model.
Subsequently, the model is expected to be able to predict the rest of the well capacity curve with reasonable accuracy.
REFERENCES
- Dreyer, Erik, "Operational considerations in gas storage,2 presented at the Norwegian Petroleum Society Gas Transport Symposium, Haugesund, Jan. 30-31, 1989.
- Asserhoj, Povl, "Kommercielle og strategiske overvejelser ved etablering at naturgas lager," presented at the Norwegian Petroleum Society Conference, "Gas i Norden," Trondheim, Nov. 12-13, 1990.
- Oebro, Hans, "Study of underground gas storage using nitrogen as cushion gas," presented at the 1986 International Gas Research Conference, Toronto, Sept. 8-11, 1986.
- Sloan, E. Dendy, Jr., Clathrate Hydrates of Natural Gases, Marcel Dekker, New York, 1989.
Copyright 1991 Oil & Gas Journal. All Rights Reserved.