The underground storage of hydrocarbon products has a history of several decades. The environments, operations, safety and security advantages of the underground oil and gas storage are well documented. In many cases, in particular for large volume storage, the underground storage has been proven to be cost-effective in comparison with the aboveground steel tanks. Other advantages include lower operation costs, less fire hazards, few surface land requirements and more constant storage temperature that may lead to less energy consumption.
Conventional methods for the underground storage of oil and gas include the uses of aquifers, depleted reservoirs in oil and gas fields and in rock salt caverns. Storing oil and gas in depleted oil and/or gas reservoirs has been the most frequently used method. In fact, up to 1950s, all gas storage facilities in USA were in depleted oil and gas reservoirs. By 2008 there were 480 gas storage facilities of this kind in the world, providing around 76% of gas storage volumes . It is commonly accepted that using depleted oil/gas reservoirs is the most widespread method of storing natural gas in large quantities and representing the most cost-effective storage option. The maximum storage gas pressure is the original reservoir pressure at the time of discovery, implying very little risk of gas leakage since oil or gas was there before extracted.
Storing gas in aquifers is more or less based on the same concept as in the depleted oil/gas fields. However, a great amount of site characterization is necessary to ensure suitable geological conditions for gas storage (reservoir, structures and caprock). The potential of the caprock failure and gas leakage from the storage formation resulting from the new pressure gradient needs to be carefully evaluated. As a result of more preliminary work to verify the capability to hold and contain gas under pressure, this option is more costly in comparison with the depleted oil/gas reservoir storage. 80 facilities of this type are existing worldwide , most of which are located in USA, the former Soviet Union and Western Europe.
It is a common practice to store compressed natural gas in caverns created in rock salt formation by leaching, which is carried out by injecting fresh water to dissolve the salt layers. One of the advantages for this type of storage is the high injection/withdrawal rate compared with the storage in depleted reservoirs and aquifers. The base gas (cushion) is also relatively low. However, the construction period is longer compared to other conventional storage methods. In some salt formations, the creep property may lead to a decrease in the storage volume with time. The gas storage in rock salt caverns can be used for both strategic and commercial purposes.
The so-called unconventional underground oil and gas storage methods include storages in unlined mined rock caverns, lined shallow caverns and abandoned mines. These methods were invented in recent decades, but have developed rapidly. This paper will focus on the rock mechanics, engineering geological and hydrogeological problems related to these storage approaches.
Gas and oil leakage control
The most essential issue associated with the underground oil and gas storage is the prevention of gas and oil from leaking out of the storage caverns. The leakage will create a loss of the stored product and contaminate the groundwater, thus leading to environmental hazards. Kjørholt  summarized the methods of gas leakage control, as given in Fig.1, which were also applicable to the oil leakage. There are two basic categories of leakage control alternatives, namely, by reducing the permeability of rock mass and by hydrodynamic containment.
The means of permeability control involves the engineering measures of reducing the permeability of rock mass if it is not sufficiently gas/oil tight. The commonly used measure is grouting. However, if the permeability cannot be reduced to the expected level, other methods, such as concrete or even steel lining, may be applied. Rock permeability can also be reduced by freezing, which may be used under special conditions.
As illustrated in Fig.2, Fig.3, Fig.4, the concept of hydrodynamic containment includes the following aspects:
(1)The hydrostatic pressure of groundwater is higher than the pressure of oil and gas inside the storage cavern.
(2)The groundwater pressure gradients around the cavern always act towards the cavern.
(3)The groundwater will flow into the cavern, but the stored oil and gas will not leak out of the cavern.
(4)The stored product must be lighter than water and insoluble in water.
(5)The leaked-in water will be collected in a sump at the bottom of the cavern and be pumped out to the ground surface.
This requires a stable groundwater regime with sufficient water supply. Furthermore, the regulation in Norway requires an additional safety reserve as specified below.
In the area where the groundwater level forms the barrier against leakage of the stored material, the groundwater level must correspond to the vapor pressure of the stored material, plus an extra 20 m water column as a safeguard against irregularities in the rock .
When the natural groundwater condition does not meet this requirement, the water curtain may be a solution, which is a system of boreholes pressurized by water injection, creating an artificial and controllable new groundwater boundary condition. The boreholes are often drilled from small galleries located slightly above the cavern. The water pressure in the water curtain is controlled and maintained to meet the requirements.
Another condition for the successful implementation of the hydrodynamic containment is to keep the pores and joints in the rock mass around the cavern completely water saturated during the entire operation period, because even the local unsaturation may cause oil and gas leakage. In this regard, the water curtain can also help to establish and maintain the water saturation of rock mass.
For many underground oil and gas storage facilities, the permeability control and hydrodynamic containment are utilized jointly, the latter is to ensure the gas-tightness and the former is to reduce the amount of water flowing into the cavern for economic operations.
Storage of crude oil and LPG in unlined mined rock caverns
Crude oil storage
Due to the low pressure, the mined rock caverns for the storage of crude oil are commonly seated at a shallow depth with overburdens greater than 50 m. The stress-induced stability is normally not a critical issue because of the small overburden. The commonly adopted cross-section areas range from 400 to 550 m2. The caverns are unlined and excavated usually by the drill and blast method.
One of the principles for storing crude oil in mined rock caverns is the concept of hydrodynamic containment to ensure the oil/gas tightness. Water curtains are used in almost all storage facilities of this kind, which must be designed to meet the requirements for providing a continuous and fully saturated groundwater regime with sufficient water pressure. The rock conditions, such as the jointing system, must be considered in the water curtain design. Measures should be taken to prevent the water curtain boreholes from being clogged. Vertical water curtains may also be needed to avoid interference between neighboring caverns when different products are stored. Another function of the vertical water curtain is to prevent the rock mass in the pillars between caverns from desaturation.
In addition to the stored oil, there are gases above the oil surface, which are saturated with vapor from the oil. Below the oil there is a water bed formed by the water leaked into the cavern from surrounding rock. Very often a fixed-depth water bed is used, meaning that the height of interface between the oil and water is constant. The storage can be operated as a “closed” or an “open” system. For the closed system, the gas above the oil does not have direct communications with the atmosphere, and the gas pressure usually varies from 0.05 to 0.3 MPa, depending upon the oil level and so on, if the water bed is also fixed. The required groundwater pressure, as regulated by the water curtain, needs to take the gas pressure into account. For the open system, when the oil is pumped into the cavern, the gases will be released and very often be flared off outside the cavern.
The concept of hydrodynamic containment allows groundwater to flow into the cavern. However, the amount of the water ingress has to be controlled and minimized to reduce the operation cost resulting from pumping water out of the cavern and the economic operation of the water curtain system. Grouting is a common technique used for reducing the permeability of the rock mass surrounding the caverns. In Norwegian tunneling practices, probe holes are drilled ahead of the excavation face and pre-grouting is performed if the probe hole observation indicates such a necessity . Experiences have demonstrated that the pre-grouting is by far more effective and cost-saving in comparison with the post-grouting.
Monitoring of groundwater is essentially important that should be done, starting from the feasibility study or even pre-feasibility study phase, and be conducted throughout the entire operation period.
Figure 5 gives the sketches of the Sture crude oil storage caverns in western Norway , which consists of four storage caverns, providing a total of around 1 × 106 m3 storage volume.
LPG usually refers to propane or butane, or their mixture in the liquefied phase. The liquefaction may be realized by pressurization or cooling. The LPGs stored in mined rock caverns are all cooled and kept under around the atmospheric pressure, for instance, for propane the temperature is 40 °C–41 °C below zero. The reason that the petroleum gases are stored in liquid phase is simply due to its efficiency. The ratio of the volume of the vaporized gas to that of the liquefied gas is typically around 250:1, depending on composition, pressure and temperature.
In terms of leakage control, the storage of LPG is basically conducted following the same principles as for oil storage. The groundwater is used as a seal for tightening the rock masses. The difference is that, due to the low temperatures, the water will freeze to ice.
During operation, water ingress into the cavern is not acceptable. Actually, all water in the rock mass (primarily the water in joints and pores) in the close vicinity of the cavern will be frozen. Due to the low permeability of the frozen rock, the frozen layer may function as the second barrier against water migration.
The major challenge appears during the cooling period. When the temperature inside the cavern drops to the designed storage temperature, e.g. –40 °C for propane, the temperature in the surrounding rock may just start to drop. The thermal-induced tensile stress in the circumferential direction of the cavern will open the pre-existing joints in the rock, so that the enhanced groundwater may flow into the cavern along the open joints and freeze to ice inside the cavern, which is already chilled. If a large amount of ice is built up inside the cavern, it is extremely difficult to pump it out. A properly designed cooling-down process, therefore, needs to be followed to minimize the amount of water leaking into the cavern. Experiences indicate that the icing phenomenon may also be related to the cooling period, which is normally 60–150 days. Experiences also indicate that grouting has little effect in this situation. Numerical analyses are usually performed to predict the temperature development with time and the rock stresses (including the thermal stresses) during the cooling-down period as well as the operation time . An example of such analyses is shown in Figs.6 and 7. However, numerical prediction for the amount of ice, which will be accumulated inside the cavern, is a serious challenge.
An accident at an LPG (propane) storage cavern in Norway demonstrates how important roles the thermal stresses can play . The 126 m-long unlined cavern has varying cross-sections, of which the maximum is 21 m wide and 33 m high at the cavern’s end. The cavern was supposed to be cooled down to –40.5 °C at essentially atmospheric pressure within 90 days. However, a huge amount of water leaked into the cavern during the cooling period and froze. Finally, 40% of the storage volume of the cavern was filled with ice. It was believed that the excessive water inflow was caused by the openings of rock joints resulting from thermal stresses.
Another problem associated with the thermal stresses is the potential instability of the concrete plug, which is used to seal the cavern. When the temperature decreases, both the plug and the surrounding rock will shrink, leading to an opening of the interface between the plug and the rock. As a result, the shear strength of the interface will decrease, thus enhancing the potential of the plug slip. The concrete plug has to be designed to be sufficiently long and to have adequate capacity against shearing. Specially designed contact grouting and rock bolting may be utilized for this purpose.
Storage of compressed natural gas in deep rock caverns
Storing compressed high-pressure natural gas requires caverns with large overburden so that the in-situ rock stress can balance the gas pressure. From the point of view of the gas-tightness, it is also required that the groundwater pressure must be higher than the gas pressure in the close vicinity of the cavern periphery. If this condition is not met in nature, an artificial water curtain system is needed to prevent gas leakage. In addition, the capillary pressure provides an additional assistance on the gas-tightness, which is then sometimes taken as a safety reserve. Due to the higher cost resulting from the requirement of deep caverns, there is only one existing project of this type in the world, i.e. the Háje underground gas storage in the Czech Republic (Fig.8).
The storage facility is composed of some unlined tunnels with a total length of 45 km and cross-sections of 12–15 m2, offering a storage volume of 6.2×105 m3. The caverns are situated at the depth of 955–961 m below the ground surface and the groundwater table is 850 m above the caverns. There is no water curtain system in the facility for the purpose of gas tightness of the caverns.
The gas injection started in July 1998, and since then the operation gas pressure has fluctuated from 2.0 to 9.5 MPa with seasons. The highest gas pressure of 9.91 MPa was recorded in the winter between 2005 and 2006. The safe operation at a pressure higher than the hydrostatic pressure of the groundwater without a water curtain system is attributed to the extremely low permeability of the rock mass. The hydraulic conductivity of the host rock ranges from 10−10 to 10−12 m/s.
A risk analysis was performed on the feasibility of increasing the gas pressure to 12.5 MPa to meet the market demand of so-called super peak gas consumption . By increasing the gas pressure from 9.5 to 12.5 MPa, the stored gas volume will increase from 52 × 106 to 75 × 106 m3 and the daily output from 6 × 106 to 9 × 106 m3. The comprehensive risk analysis was supported by field tests and numerical analyses of cavern stability and gas migration. The conclusion of the study supported the plan of gradually increasing the gas pressure to 12.5 MPa and gave recommendations for enhanced monitoring systems.
There is an ongoing research program financed by the EEA Norway Grant and jointly performed by the Czech Technical University in Prague (CVUT), the Norwegian University of Science and Technology (NTNU) and SINTEF Building and Infrastructure. In the study, a series of field tests were carried out at the Josef Gallery, an underground research facility operated by CVUT, by injecting air into drilled boreholes. An attempt was made to investigate the correlation between the gas conductivity of the rock mass and the rock mass classification index. The preliminary outcome of the study indicated slightly higher gas conductivity than that predicted by an early empirical equation. The observations also confirmed that the gas injected into the rock formation would return to the borehole when the pressure dropped to zero with a certain time delay.
Storage of oil and gas in abandoned mines
The first storage for petroleum products in rock caverns converted from abandoned mines was found in Sweden in 1947–1950. Since then such storage technique was studied at several sites and only parts of them were materialized, of which very few were well documented. Peila and Pelissa  conducted a comprehensive survey in 1995 on the reuse of abandoned mines for a variety of purposes. Table 1 lists the oil/gas storage facilities in abandoned mines selected by the author. The Weeks Island salt mine for oil storage and the Leyden coal mine for gas storage will be taken as examples for detailed description.
The study on storing oil in the abandoned salt mine at the Weeks Island, Louisiana, USA, started in 1975 . Other sites considered in the same study include two limestone mines (Iconton mine in Ohio and Central Rock mine in Kentucky) and two salt mines (Kleer mine in Texas and Cote Blanche mine in Louisiana), as listed in Table 1. However, all other four candidates have logistical or operational problems, or are limited by volume, consequently storages in these mines are not realized. The Weeks Island salt mine was a two-level room and pillar mine in salt domes. Oil filling started in October 1980 and was completed in April 1982. No unusual occurrence or incident was reported during the first five years of operation until an accumulation of brine in the filling hole sump was noted in 1987. Concerns arose because the source of the brine was unknown. Some measures were taken at a surface sinkhole about 11 m across and 9 m deep over the edge of the storage facility, as observed in May 1992. The extensive diagnostics concluded that the sinkhole resulted from mine-induced fractures in the salt, which took many years to develop by creeping, eventually causing fresh water to leak into the storage chamber and dissolve the overlying salt, thus causing the overburden to collapse into the mined space. The decision on decommissioning the facility was made in 1994. Then the residual oil was removed, water inflow was prevented, the cavity was backfilled with brine and the facility was plugged and finally abandoned.
Leyden coal mine was located about 14 miles northwest of Denver, Colorado, USA, and the active mining operation was in 1903–1950 . The total production of the coal was 6 million tons from 2 seams about 365 m long and 245–260 m below the ground surface, as shown in Fig.9. There were three shafts for mining operations and one shaft for ventilation. The mining method was room and pillar. Extraction efficiency was about 35%, meaning that 65% (or about 11 million tons) of the original coal remained in place after mining ended, primarily in the pillars.
The water level at the shaft No.3 was measured in 1958 at an elevation of 1 550 m that was 213 m below the ground surface and 30–46 m above the mining areas, indicating that all the mining areas were filled with water. It was also found that the roof of the mine eventually collapsed, leaving only a few pillars standing. The mined areas were completely filled with broken rocks and rubbles. There was no evidence of ground subsidence. Below the coal seams there was an aquifer and above the coal seams there was a 21 m-thick impermeable claystone layer that provided a seal preventing gas from escaping from the facility.
Four shafts were sealed to ensure the gas-tightness. Figure 10 illustrates the sealing of the shaft No.3. When the sealing work started in January 1961, the mine was filled with water, so the first action was to pump out the 3.9×105 m3 water. The shaft was cleaned and the bottom was plugged with 7 m concrete, followed by filling of gravel to the final water level. Further above it was the main concrete plug. The upper 15 m of the shaft was completed with graded rock overlain by graded gravel, graded sand and 6 m compacted clay plus 1.5 m sand and bank-run gravel to ground surface. Then the drilling mud was injected to the space in the rockfill to act as the sealing agent. Sealing operations of other shafts were similar.
The sealing work finished in September 1961, followed by immediate gas injection. There were 14 injection/withdrawal wells. By November 30, 1961, a total of 34 × 106 m3 of gas at a pressure of about 1.45 MPa was contained in storage. A minor leakage was detected by the monitoring system and remedial work was performed. Ever since the facility has worked successfully at the maximum gas pressure of 1.72 MPa, the total capacity and maximum working capacity is approximately 85 × 106 and 62 × 106 m3, respectively. The annual withdrawal period is over 100 days.
It was observed that the portion of the gas that migrated into the surrounding rock at high pressures was recoverable when the pressure of the mined area was lowered during gas withdrawal. Water was also produced with gas when the cavern pressure was low and the withdrawal rate was very high. So gas was produced into a gas-water separator first, where most of the water was removed prior to the gas entering the gathering system.
The total investment for the facility was 18 million US dollars and the annual operation cost was around 8 × 105 US dollars. The major benefit is to enable the owner to balance their gas supply by purchasing gas at the off-peak prices, so that the company can save about 14 million US dollars per year. The successful construction and operation of the Leyden facility demonstrate that converting abandoned mines to gas storages is both technically feasible and economically profitable.
However, according to literatures, there is only a handful abandoned mines that have been successfully converted to the gas storage facilities. One of the reasons might be the strict demand on the hydro-geological conditions associated with the gas-tightness requirement, which is not so easy to be met for abandoned mines. For storing oil in abandoned mines, the requirements on geological and hydrogeological conditions are lower.
Discussions and concluding remarks
Hundreds of successful projects worldwide have demonstrated that storing hydrocarbon products in mined rock caverns is a proven technique. In addition to the common requirements for cavern stability, the main challenge comes from the prevention of oil/gas leakage. Both the permeability control and the hydrodynamic containment are widely used for this purpose. When the natural groundwater regime does not meet the requirements, a man-made water curtain system can help to establish the hydraulic boundary condition. For the cool storage of LPG, the design should take into account the effect of thermal-induced stresses and the opening of pre-existing joints in the surrounding rocks. In addition to weakening the cavern stability, such thermal-induced joint opening may create channels for water inflow into the chilled cavern, thus leading to ice accumulation inside the cavern. A properly designed cooling-down process, assisted by numerical simulation of the transient temperature development, can certainly prevent the disaster from happening.
Depth is a crucial parameter for oil/gas storage caverns. In addition to the geomechanical requirements for the cavern stability, the hydrodynamic criterion described in Section 2 must be met. The permeability of the rock surrounding the caverns must be sufficiently low so that the water inflow into the cavern is under the designed level, which is related to the capacity of pumping and the volume of the sump located at the cavern bottom for collecting water.
It should be noted that one of the important features of the oil/gas storage caverns in comparison with the caverns for other purposes is that there is almost no possibility for any remedial works inside the storage space after the facility is put into operation. Therefore, the long-term structural stability of the caverns has to be ensured for the whole life of the facility. Adequate rock support design by empirical means and numerical simulations based on reliable geomechanical characteristics of the rock mass and allowable operating conditions is essential for the lifetime safety of the facility.
After construction, air pressure tests shall be carried out by filling the caverns with compressed air to a pressure exceeding the maximum operation pressure to demonstrate the gas-tightness. When several caverns are involved, each cavern shall be pressurized in turn while other caverns are kept at an atmospheric pressure. The first filling shall not be conducted until the air tests have been completed successfully.
Both positive and negative experiences have been gained in converting abandoned mines to oil and gas storage facilities. Some people describe it as a proven technique, while others on the contrary say the test has failed. The author believes that an abandoned mine could be converted to a crude oil or natural gas storage if the existing natural hydrogeological conditions meet the gas-tightness requirement and the cavern stability is ensured. However, such sites can not be easily found.
Monitoring is an essential part for all types of underground oil and gas storage facilities. The groundwater regime has to be permanently monitored by a network of observation wells (piezometers), ensuring the integrity of the groundwater table in the vicinity of the caverns.
The storage facilities shall be designed, constructed, operated and maintained so as to present no inadmissible risk on the safety of the staff and the public. Risk analysis should be performed to evaluate the likelihood of risks and consequences. Prevention and mitigation measures have to be taken beforehand in particular for occurrence of blow-out and leakage.
Source: Rock engineering problems related to underground hydrocarbon storage
Author: Ming Lu