Sonar surveys used in gas-storage cavern analysis

This wellhead over Regina Cavern No. 5 is used for gas injection and brine removal.

TransGas Ltd., Regina, Sask., has successfully performed these operations on several of its deepest and highest pressurized caverns. The data can determine gas-in-place inventory and assess changes in spatial volumes.

These changes can result from cavern creep, shrinkage, or closure or from various downhole abnormalities such as fluid infill or collapse of the sidewall or roof.

TransGas Ltd. is the natural-gas transmission and storage side of the parent SaskEnergy Inc., the local distribution company in the Canadian province of Saskatchewan. TransGas has more than 35 years' experience developing caverns in bedded salts of the Prairie Evaporite formation within the Elk Point salt basin of western Canada.1

Cavern development occurs 3,000-5,000 ft below surface with 2,000-4,000 psig wellhead pressures. Usual salt thickness varies 300-500 ft. A single cavern is typically designed for a spatial volume of around 900,000 bbl or a usable gas volume of 800 MMcf with withdrawal rates of 40-200 MMcfd.

TransGas currently operates 24 bedded-salt caverns, of which 18 were solution-mined exclusively for sweet natural-gas storage while the remaining 6 were converted into bedded-salt LPG caverns (OGJ, Feb. 19, 1996, p. 39).

Confirming the interior condition of a solution-mined storage cavern in bedded salt is not performed regularly. Under special circumstances, a gas sonar would be used to check a suspected internal cavern problem.

Running a special gas sonar survey in an active pressurized storage cavern is similar to running wire line logging tools in a gas-filled cavern but more complex and considerably more costly.

The complexity relates to the sonar tool itself, the service conditions, and limitations under which the tool must be operated. Costs to run this service can be four to five times higher than those for running a conventional sonar-survey tool under a nonpressurized condition in a brine-filled cavern.

Cavern sonar-survey results can also be used to perform finite-element modeling which can then be used for predicting cavern performance behavior, stress levels, operating-pressure regimes, closure or creep rates, effects of cycling operations, rate of change of internal cavern pressure drops, along with other parameters.

TransGas has used this cavern modeling technique for increasing the maximum operating pressures in some of their caverns.

Conventional survey

This sonar is run under zero wellhead surface pressure with a hydrostatic brine column from surface to cavern bottom. This procedure is the same whether the final cavern is to be for natural gas or LPG or some other service.

A conventional sonar survey is usually required as part of the provincial (Canada) or federal government regulatory permitting process and is one method used to determine the spatial volume of the cavity. Canadian Standards Association Standard Z341-93 "Storage of Hydrocarbons in Underground Formations" sets out requirements for running sonar surveys.

A sonar survey will also provide a picture of the overall final shape of the internal cavern. Various displays are currently available, including two-dimensional (2D), isometric, and three-dimensional (3D) plots, as well as various options to manipulate data and schematics.

Normal cavern-logging tools will only investigate a small radius around the well bore. Sonar tools have a much greater range.

Usually a clear image of the shape of the roof can be obtained. Some masking may be present depending upon the amount of roof-control material (inert hydrocarbon) left in place floating on top of the cavern brine. The roof may be large and flat, tall and narrow, or arched in shape, information that is significant to the future operation of the cavern.

A sonar survey will also identify abnormalities in a cavern shape that may present future operating problems. Such abnormalities include ledge overhangs or protrusions into the cavern of unmined insoluble materials that may break off, fall, and damage the internal casing string as well as any areas in the sidewall structure where uncontrolled leaching (cavern enlargement in any one direction) may have been caused by the presence of soluble potassium salts.

It is only when a cavern is put into its intended service that future sonar surveys may become more challenging and more expensive to run.

Running surveys in pressurized caverns

A sonar run in a gaseous environment is different from that run in brine or liquid because of the type of fluid (gas or liquid) in the cavern, its densities, and the acoustic velocity of the signal in the medium.

Specialized sonar tools can view through one and two sets of inner casing in the open area of liquid-filled caverns (brine or brine/LPG) which allows the sonar to be run while the cavern is in its normal operating state and thus eliminates the expense of raising the inner string of casing before a sonar run.

Running a sonar survey in a gas-filled pressurized cavern is more complex.

First, the sonar cannot be run with inner casing throughout the open part of the cavern. At present, the sonar signal will not see through an existing string of pipe. Therefore, the inner casing string must either be raised, which can be an expensive snubbing job, or cut off around the roof level.

If the inner casing is to be cut off near the roof for cavern sonaring, then a chemical cutter is preferable to a jet cutter. A chemical cutter will give a clean, smooth cut to the pipe end, thereby eliminating casing splits or "fingers" (which occur with jet cutters) and avoiding damage to expensive sonar tools.

Chemical cutters are approximately ten times more expensive to run than jet cutters but could prevent loss of logging tools.

Another alternative is that a cavern may be emptied of all stored gas and filled with brine or water and a conventional sonar run with zero wellhead pressure. The brine is then removed and the cavern refilled with gas.

Few companies have sufficient time in their operating schedules or money in their operating budgets to consider this process.

This has necessitated the special gas survey route of sonaring the cavern while under normal operating-pressure conditions.

Special gas sonars work best at higher pressure when gas is denser and the transducer signal is more responsive and greater ranges (radius) of sonar measurements are achievable. Signal response depends on the speed of the sound wave through the denser medium. The attenuation of the signal is therefore higher at low pressure and lower at high pressure.

TransGas has run three special gas-sonar surveys in two caverns since 1989 at operating pressures of around 2,000 psig wellhead.

In the past, a sonar survey could only be run in a liquid-filled cavern, usually brine, under zero surface wellhead pressure.

The science of performing sonar surveys has advanced considerably since the 1960s when TransGas developed its first cavern. At that time, it was common to photograph each traverse or slice of cavern for further interpretation.

Today, with computerization-digital information such as 2D plus 3D isometric plots and 3D shade plots- images can be rotated and tilted through 360° vertically or horizontally. A cavern can even be turned inside out to reveal particular features.

Only since the late 1980s has the technology been available to the North American market to run sonar surveys in gas-filled caverns under pressurized conditions and to see through single or multiple casing strings for sonaring liquid-filled caverns.

Shortcomings; services

There are invariably areas of cavity such as ledges and irregularities that the sonar signal will be unable to penetrate. Even rotating the transducer/receiver head upwards or downwards to shoot the roof or floor areas will leave some areas undetected.

If the cavern has a second liquid (roof-control fluid; for example, diesel oil) floating by density difference on top of the brine, the complete shape of the roof will be masked and features critical for cavern operations may be concealed.

Early liquid-filled cavern sonars were as much as 8-10% low when compared to other methods of determining final spatial volume (OGJ, Mar. 3, 1997, p. 74; Mar. 10, 1997, p. 69). Technology development has improved the accuracy of sonar surveys so that current accuracy may be on the low side by 2-5%.

For mature and existing caverns filled with natural gas, the few special gas sonar surveys which TransGas has run were not to check on past cavern fluid-filled sonars but to respond to specific downhole issues in specific caverns.

This included confirmation of current cavern shape, identification of suspected problem areas, and the determination of a new spatial volume. It has not always been possible to run a complete sonar over the total cavern from top to bottom because of specific internal cavern problems. The roof area is still most critical for future operations and structural integrity, however.

TransGas has run three special gas sonar surveys (in two caverns only) and additional gas sonars are being considered only for specific caverns to check on suspected downhole problems.

There are several companies located in Canada and the United States that run the conventional sonar surveys in liquid-filled caverns. However, only two companies located in the U.S. (none in Canada) have the capability to run commercial sonar surveys in pressurized natural gas-filled caverns.

One, in Houston, was used by TransGas in 1989. Then, the company was called French Well Surveys Inc. (a joint venture between Prakla Seismos GmbH and PB-KBB). It has now split off and is called Cav-Tec Inc., Houston, and is affiliated with the German firm Socon Sonar Control that now has a Houston office called Socon Cavity Control Inc. (a Socon/Cav-Tech partnership).

Another company is Sonarwire Inc., Abita Springs, La. TransGas used Sonarwire to run two gas sonars on separate caverns in 1994 and 1996. Both companies have the capability to run conventional sonars and can run sonars in brine or oil-filled caverns through one or more inner strings of casing pipe.

Cavern No. 5

An investigation was initiated to determine the cause and significance of the salt fall plus cavern structural integrity.

After a further 2 yearly operating cycles, a second roof collapse occurred which went unnoticed until picked up during routine wire line logging. Subsequent analysis determined the cavern to be usable albeit at reduced operating pressure levels for stored product security.

In both instances, special gas sonar surveys were run to determine the consequences of the salt falls and integrity of the cavern roof area.

Initial development

The final conventional sonar spatial volume was 706,482 bbl or 3.9% less than the calculated mined volume. Total brine volume metered out of the cavern during gas displacement was 736,691 bbl (sonar volume 4.1% low).

The virgin salt bed temperature registered 125-130° F.; the cavern brine temperature before first gas injection was 82° F.

The cavern had a somewhat irregular shape: more than 300 ft wide at the bottom and 70-80 ft high before the cavern diameter came back to a more reasonable diameter of around 100 ft with the smallest section of around 20 ft at the final roof location. The cavern roof depth was 5,363 ft. The maximum cavern height was 180 ft.

The floor salt thickness was 102 ft, while the roof salt thickness was 50 ft. Several unmined insoluble ledges and fingers were left jutting into the cavern void space in the top half to one-third of the cavern.

The maximum allowable wellhead pressure was set at 3,000 psig with a usable gas storage volume of 900 MMcf. The minimum wellhead pressure was 600 psig (based on the transmission-line pressure). The normal cavern gas operating temperature range was 130-100° F. (1 cycle/year).

The final developed roof position was higher than originally planned and located in a structurally unstable area with many thin insoluble bands which threatened future stability. Fig. 1 [105,487 bytes] shows the final state of cavern development.

Roof fall

In September 1989, the cavern was gamma ray/neutron logged. An 8-16 ft roof fall was indicated at the well bore area, and 60 ft of new material lay in the cavern bottom, suggesting possible side-wall ledge falls.

Cavern cycling for 5 years between high and low pressures may have stressed or flexed the salt back immediately above the roof span, weakening it and causing it to fracture and fall.

Special logs were run to check for gas loss (accumulation) above the cavern roof; no presence of hydrocarbons was found. The 7-in. production-casing-cement bond appeared sound at the roof.

In November 1989, a special gas-sonar survey was run by French Well Surveys Inc., Houston, in a dry (gas-filled) cavern environment. The cavern shut-in wellhead pressure was 2,444 psig with a cavern gas temperature of 121° F. The sonar indicated that a massive roof fall of 27 ft had occurred over a radial distance of 90-100 ft and that a 65-80 ft pyramidal rubble pile lay on the cavern bottom.

The roof salt thickness was then only 31 ft. Fig. 2 [67,889 bytes] shows the top end of the cavern following the roof fall.)

The entire cavern could not be sonared because of the tall rubble pile and the fact that the rubble top coincided with the necking-in of the wide bottom of the cavern to the narrower portion of the upper cavern. Viewed from above, the rubble pile could be seen, but the buildup prevented the sonar from reading the shape in the largest and widest part of the cavern.

In October 1991 another salt collapse was identified, although there had been no noticeable cavern operating problems from 1989 to 1992. Logging at 2,300-2,400 psig found the rubble pile 60 ft higher than identified in 1989 which meant there was now a total of approximately 120 ft of rubble in the bottom of this cavern.

The cavern roof showed no change, still registering 30 ft of salt at the well bore. TransGas concluded that the roof fall likely occurred in early 1991 when the cavern was cycled down to low pressure and an abnormal pressure rise was observed, most likely the result of fluid re-entry.

In May 1993, the cavern spatial volume was recalculated and found to be reduced to 400,000 bbl or more, or a 45% loss of space from the original volume.

In June 1993, a second special dry-gas sonar was run at 2,250 psig wellhead pressure by Sonarwire to identify the new roof shape.

The sonar tool looked down along the slope of the rubble pile to the bottom. The height of the rubble pile was confirmed at 120 ft. The cavern floor may have come up by 40 ft from the original floor position.

No major change in roof position was recorded. Although the roof had sagged 10 ft since 1989, there still should have been a minimum of 31 ft of salt above the cavern roof.

The southern half of the roof area was not well defined on the sonar. It could not be determined, therefore, whether any salt was left there or if the roof breached up into the overlying formation.

At present, it appears the entire upper portion of the cavern roof and sidewall areas have changed significantly from the original brine sonar in 1984 and the gas-sonar survey in 1989. The cavern roof area is fairly flat and approximately 230 ft wide.

Fig. 3 [113,671 bytes] shows the new shape of the top end of the cavern following the second roof fall with the rubble pile still preventing the lower larger part of the cavern from being sonared. Fig. 4 [62,257 bytes] shows the same new shape of the roof but is a 3D shade plot of the upper section of the cavern.

Since 1989, this cavern has been used only sparingly and is still classified as "emergency use only." The exact location of the cavern leak point is still unknown. The maximum operating pressure was further reduced to the 2,250 psig level to curtail the risk of additional possible gas loss.

Cavern No. 4

Investigations revealed these rock layers to be structurally tight, and no cavern integrity was compromised, although the operating pressure was reduced to ensure stored product security. A special gas sonar survey was run to determine the consequences of the salt fall and integrity of the roof area.

Initial development

In August 1977, the cavern was first converted to initial gas storage. That year a conventional sonar survey on the brine-filled cavern indicated a spatial volume of 637,431 bbl. A revised corrected sonar interpretation yielded a value of 705,429 bbl (or a 9.6% difference from the original interpretation).

The final calculated mined spatial volume for the total development phase was 711,345 bbl or 0.8% greater than the revised sonar volume and 10% greater than the original sonar volume. (Fig. 5 [98,170 bytes] shows the initial shape of this cavern sonar.)

Following cavern development in 1979, the final spatial volume was calculated at 916,906 bbl, based on brine-displacement metering during gas injection. The calculated cavern development spatial volume was 960,000 bbl. The final volume based on brine measurement proved to be the most accurate. No final sonar survey was run.

The widest section of the cavern was 208 ft and the maximum cavern height was 197 ft. The cavern roof depth was 5,349 ft with the floor at 5,546 ft. The roof salt back thickness was 15-32 ft (gamma ray-neutron log vs. sonar survey at well bore location), while the floor salt thickness was 102 ft.

At 20 ft below the roof, the cavern walls necked-in to about 62 ft in diameter because of an insoluble layer.

The maximum allowable wellhead pressure was set at 3,200 psig with a usable gas-storage volume of 1,100 MMcf. Based on the transmission pressure, the minimum wellhead pressure was 600 psig. The normal cavern operating temperature range was 130-100° F. (1 cycle/year).

Damaged string

A gamma-ray/neutron log confirmed the roof fall and indicated that the cavern roof position may have risen by 23 ft or a 23 ft roof fall had occurred. This meant that now no salt remained above the roof of the cavern.

The cavern roof was up by 6 ft into the overlying carbonates and anhydrites of the Red Beds formation and just below the Dawson Bay limestone and dolomites formation. TransGas believed the Red Beds would provide an effective seal to natural gas if it remained undisturbed.

Drillstem tests on the Dawson Bay formation indicated it to have low permeability. The low cavern pressure experienced was greater than the minimum established pressure for cavern stability concerns (0.15 psi/1 ft of depth to the top of the cavern roof).

A review of the cavern salt core data from the Regina caverns indicated that the current cavern roof position was located in competent rock that is neither porous nor fluid-bearing. There should be no concern of gas loss through the cavern roof in the cavern's current status.

Salt creep or cavern closure has not been a significant factor for caverns developed in the Prairie Evaporite salt bed in Saskatchewan.

A decision was made to run a special gas sonar survey to determine the condition of the cavern roof and general cavern condition to verify the future suitability of this cavern to hold and store gas.

In May 1993, the inner casing was cut off in mid-cavern with a chemical cutter. The cavern floor was also tagged and subsequently indicated that the cavern bottom had 40 ft of additional material. As a result, TransGas is now unable to log a cavern brine-pond fluid level. There was also no indication of a change in the roof position.

In June 1993, an attempt was made to run a special gas sonar survey using Sonarwire. The inner casing was still bent somewhat, however, and would not allow the sonar tool to pass.

In July 1993, before the running of the special Sonarwire gas sonar tool, the inner casing was again cut with a chemical cutter inside the outer production casing. On running the gas sonar, tool problems developed in that the tool was damaged by the outer production casing which may have been bent by the past roof fall.

Further attempts at running the special gas sonar survey were deferred to 1995.

Definitive survey

In May 1996, Sonarwire successfully ran a special gas sonar survey tool under wellhead pressure conditions of 2,080 psig. There was approximately 3 ft of viscous mud and approximately 7.5-8.5 ft of fluid on the cavern bottom. There appeared to be very little change to the cavern bottom.

The gas sonar gave a very good picture of the new cavern roof position. There was a bit of a rise in the roof surface towards the south side which was higher than at the well bore center line.

The 1996 sonar indicated that there must have been 31-40 ft of roof salt fall which occurred in 1993 and possibly up to 50 ft of roof fall on the south side. This was more than originally determined from well bore logging.

The new cavern roof was actually up into the overlying rock of the Red Beds by approximately 14 ft and up approximately 24 ft on the south side. This configuration could not have been determined from normal well bore logging.

Historic geologic studies indicated the new cavern roof position was located in competent rock that was neither porous nor fluid bearing.

Pressure-vs.-time monitoring since 1993 has revealed no obvious problems with this cavern. The width of the new roof fall area is approximately 120 ft.

The new cavern sonar spatial volume is 807,452 bbl. This is deemed accurate to within 2% on the low side and 5% in published literature.

The new sonar volume is 13-21% greater than the original 1977 sonar volume and 14% smaller than the final 1979 spatial volume based on brine removal. This new sonar volume is also just 0.4% greater than the calculated spatial volume from a 1993 gas-injection test.

For several years, the original 1979 spatial volume figure of 916,906 bbl had been suspect as being too high based on the annual gas-in-storage audit analysis. The cavern height is now 191 ft and the maximum width is now 223 ft. (Fig. 6 [51,662 bytes] shows the sonared shape of the cavern. Fig. 7 [102,944 bytes] is a 3D wireframe plot of the same cavern.)

Running a special gas sonar survey on this cavern was the most cost-effective method for verifying the new shape of the cavern and the current spatial volume. It went a long way toward strengthening our belief in the future usability of this cavern.

Future cavern logging and pressure-vs.-time monitoring will help to verify the downhole conditions, determine if the brine-pond is changing significantly, and help to assess the further usability of this cavern for yearly cycling operations.

Reference

1. Crossley, N. Graeme, "Gas Storage in Saskatchewan Bedded Salt," presented to Solution Mining Research Institute, Houston, Oct. 18-21, 1992.

The Author

N. Graeme Crossley is manager of process and storage for TransGas Ltd. He holds a BS (1970) in civil engineering from the University of Saskatchewan, Saskatoon, and is a registered professional engineer in Saskatchewan. He holds membership in the Solution Mining Research Institute.

CORRECTION

An error appeared in the article, "Global refining addresses increased oil demands, new challenges," by Tim W. Martin (OGJ, Mar. 16, 1998, p. 51). In Fig. 1, the values in parentheses are imports, and the values without parentheses are exports, contrary to what is stated.

Copyright 1998 Oil & Gas Journal. All Rights Reserved.