GAS-STORAGE CALCULATIONS YIELD ACCURATE CAVERN, INVENTORY DATA

R. G. Mason
Transcontinental Gas Pipeline Corp.
Houston

Determining gas-storage cavern size and inventory variance is now possible with calculations based on shut-in cavern surveys.

The method is the least expensive of three major methods and is quite accurate when recorded over a period of time.

MEASUREMENT METHODS

Several methods exist to determine gas-storage cavern size and inventory variance. Among these are direct measurement by sonar survey, direct measurement by water fill, and volume calculation from data furnished by shut-in surveys.

The method of direct measurement is the most accurate but is also expensive, requiring a sonar log or cavern degassing. Until recently, the cavern had to be filled with water to conduct the sonar survey.

The sonar device, which measures the time of return of sound waves reflected from the cavern wall, can also be used to determine void space behind the casing. This measurement is inaccurate if a cement sheath covers the casing, which is usually the case in wells completed with high-quality techniques.

The cavern size can also be directly measured by the cavern being degassed through means of a water fill. This method is not as accurate as the sonar because of a chance of measurement error between the amount of injected water and withdrawn gas.

The method may cost $100,000 to $1 million because most gas-storage caverns hold more than 1 million bbl in volume, and water fill may cost $0.10-1.00/bbl. Thus, water fill is usually reserved for well maintenance which can only be achieved by killing the well.

The final method involves the calculation of inventory and cavern size from subsurface data taken during shut-in surveys.

This method is the least expensive, involving a relatively small amount of down time and inexpensive logging costs. And it is quite accurate when recorded over a period of time.

INVENTORY VARIANCE

Many salt caverns in the U.S. that are used to store natural gas at high pressures experience unaccounted for gas losses (inventory variance) and salt creep (cavern shrinkage).

An inventory variance can be the result of gas migration or poor measurement practices. But, because salt caverns tend to be self-healing at the salt-casing contact as a result of the plasticity of salt at high temperatures and pressures, most inventory variances can be traced to poor measurement practices.

It should be noted, however, that gas migration caused by leaking tubulars above the salt stock has also been documented. This phenomenon occurs when the cement sheath is of poor quality and the sealing integrity of the collars has been compromised. Gas leaking in this manner may migrate to shallow low-pressure sands or escape at the surface.

Cavern shrinkage is a phenomenon caused by the plastic motion of the salt stock (Fig. 1). Such shrinkage is at a minimum when cavern pressure is at a maximum. Historical operating data have shown, however, that caverns can and will shrink at high storage pressures and, to a lesser degree, when filled with water.

The shrinkage occurs when the cavern pressure is less than that of the lithostatic pressures, the pressure caused by the weight of the salt and rock above the cavern and measured as slightly less than 1 psi/ft.

Cavern shrinkage can be monitored and to some extent limited by sound operating practices. General storage operations will cause cavern shrinkage due to the lower pressures associated with withdrawal activities. The maintenance of high cavern pressures, however, combined with shallow leaching depths and retention of cavern shape are factors used to limit shrinkage.

CAVERN SURVEY

Calculations for shrinkage and inventory, as demonstrated here, are arrived at in much the same manner as inventory verification in a porous reservoir. In addition, the same general shut-in procedure and material balance calculations are applied.

Cavern-sizing surveys are basic studies which produce information used to determine the volume of space in a salt cavern.

As previously noted, there are three methods used to discover cavern volume. Direct measurement requires the use of sonars introduced into the cavern by electric wire line, while degassing requires filling the cavern with water; both of these methods are expensive.

The third method, using subsurface data to calculate cavern volume, is the least expensive and very accurate when used with valid reservoir-engineering techniques.

Compilations and comparisons of survey results will allow the operating company time to make sound decisions concerning well bore and cavern maintenance with comparatively inexpensive logging instruments and wire line equipment. Cavern-sizing surveys that produce data used to calculate cavern size utilize temperature and pressure bombs introduced into the cavern on a slickline. These data and the resulting calculations are used to estimate inventory variance and cavern creep.

Cavern-sizing surveys involve two shut-ins and an injection or withdrawal of a finite volume of gas. The initial and final temperatures, pressures, Z factors, and injected/withdrawn volumes are data, which, when introduced into a material balance equation, produce a reasonable estimation of cavern size. This measurement and accompanying pressure and temperature data are then employed to estimate cavern inventory.

PROCEDURE

Essential to the subsurface method are consistent procedural practices.

This procedure compares results of calculations from data obtained during sizing surveys.

A condition for comparison is analogous data. Therefore, the shut-in survey should always follow the same operating cycle, injection or withdrawal.

In this way, the procedure will generate data which produce a reliable trend with reproducible results when each survey is graphed over an extended period of time.

Another consideration for consistent practices is the amount of shut-in time before the down-hole surveys are conducted. A certain amount of time must be allotted for the cavern pressure to stabilize.

The stabilization period varies due to the size, shape, and depth of the caverns. Generally, the time required to allow the cavern pressure to stabilize may be predicted by a graph of the daily wellhead pressures and observation of an asymptotic curve trend.

Fig. 2 shows examples of near stabilized shut-in periods. It is worth noting, however, that the cavern is usually shut-in for a period of time that depends upon pipeline operating conditions. An urgent need for gas may dictate an early end to field shut-in time.

As with all surveys utilizing measured quantities from use of mechanical instruments, some measurement error will exist. Errors introduced into the formula due to data gathering will not allow direct comparisons with measured volumes.

The data will always have some margin of error due to the operating conditions of the caverns, with the greatest data errors doubtlessly existing in the temperature recordings. With the best of log data, the accuracy of the calculations will be no better than approximately 5% and with obvious errors in data, the inaccuracy will range to 100% or more. This range of error percentage produces clear evidence that sound engineering procedures must be practiced.

APPLICATION

Preplanning is essential for a successful shut-in survey because cooperation from gas control, field operations, and storage engineering departments is integral. Operations must be coordinated to prepare the caverns for shut-in.

The cavern may then be shut-in and allowed to stabilize before downhole temperature and pressure surveys are conducted. This is necessary to obtain the most accurate data possible.

Gas is then injected or withdrawn, depending on the inventory status. It is not necessary completely to withdraw to base gas or fill the cavern. An injection or withdrawal of a volume of gas approximately equal to 15% of the total cavern volume is, however, suggested to produce acceptable results.

The cavern is then shut-in a second time and allowed to stabilize.

A final bottomhole survey must now be conducted to obtain the final data necessary for inclusion in the material balance equation. The result yielded by the equation is the pore volume of the cavern.

Some manipulation of the material balance equation is involved to obtain cavern size and inventory. It is helpful to point out that the caverns are not identical in size at every survey because some shrinkage always occurs. Pore volume must, therefore, always be recalculated for each shut-in period.

This is slightly different from a depleted gas-storage field, with no water drive or other externally induced drive force, in that pore volume of the depleted field may be assumed constant. The initial calculation as discussed will reestablish pore volume for every survey.

Inventory verification may then be established at current conditions for comparison with measured volumes to estimate possible variance. For estimation of cavern shrinkage, cavern volume must be established at original maximum operating temperature and pressure for comparison with past cavern sizes. Each formula as discussed will determine the necessary criteria successfully to complete the sizing surveys.

The pore volume obtained from the calculations may be utilized to calculate an estimate of the gas in place at current operating conditions to reveal inventory variance and at original maximum conditions to reveal cavern shrinkage.

DATA GATHERING

Primary data are collected at time zero or initial shut-in. Wellhead pressure is recorded at regular time intervals to ascertain cavern stabilization because wellhead pressures depend on cavern pressures and respond to downhole stabilization. Wellhead temperatures, on the other hand, follow ambient surface temperatures and tend to respond to weather patterns.

Stabilization time for cavern temperature is much longer than that for cavern pressure. Due to the volume of the cavern envelope and relatively small amount of contact area at the cavern walls, the temperature of the gas may take several months to reach that of the surrounding rock.

Hence, bottomhole temperature data cannot be calculated with surface readings. Since cavern temperature at a given depth does not parallel gradients, surface readings prove inaccurate. As such, calculations with gradient equations and which use surface readings are useless.

Consider a comparison with a depleted gas-storage reservoir. The gas touches an infinite area of the host rock through contact with the interstices of the sandstone or vugular spaces of carbonate reservoirs.

The contact area tends to reduce the time necessary for the gas to reach a temperature corresponding to a regional gradient. Because there are no interstices' contact points in a salt cavern, the stabilization time for temperature is too long to conform to a reasonable shut-in period.

The temperature of the stored gas, therefore, is a function of the operation, injection or withdrawal, not time intervals functionally tied to the geology (Fig. 3). Operating history has shown that temperature is a separate function for the injection or withdrawal operation at near-time" intervals and is directly related to the ongoing or recently completed storage operation.

Temperatures following an injection cycle are warmer, and temperatures following a withdrawal cycle are cooler than gradient temperatures. These temperatures will stabilize to gradient temperature after a long period of time, much longer than practical for a shut-in survey. Therefore, temperature data for calculations are said to be "near-time" dependent, governed by the storage operation.

Cavern shape functionally affects the temperature of the gas in the cavern (Fig. 3). Spherically shaped caverns yield temperatures which tend to cool as depths increase after an injection period. Cylindrically shaped caverns, on the other hand, yield temperatures which remain relatively constant over the entire depth interval.

All cavern temperatures tend to remain constant, however, after a withdrawal exercise over the entire depth range.

Temperature and pressure data should be taken at several intervals to obtain an average over the entire cavern interval. An arithmetic average may be used for cylindrically shaped caverns. But a geometrical average is used for spherically shaped caverns because of the changes in diameter.

An interesting condition exists near the bottom of all caverns surveyed. The remaining water pool at the cavern bottom has a temperature which is usually cooler than the gas temperature. As shown in Fig. 4, the temperature drop begins below 5,750 ft for spherically shaped caverns and below 6,100 ft for cylindrically shaped caverns.

Continued recordings of temperature and pressure while the instruments are in contact with the water will affect the average temperature and pressure of the cavern. Thus, the operator must not use temperature and pressure data obtained while the instruments are in contact with the water.

This situation is avoidable because the interface can be identified by the sudden temperature change.

SURVEY DATA; CALCULATIONS

The equations presented calculate the gas in place (GIP) for inventory verification and the maximum volume to determine cavern shrinkage. The data obtained in the cavern surveys are given in Table 1; the results of the calculations are presented in Table 2.

These results (Figs. 5 and 6) indicate the trend in inventory and cavern size necessary to determine variance and shrinkage.

Graphical results depict surveys which have been conducted over a period of several years. As seen in Fig. 5, the results do not produce a smooth line graph because of changing cavern conditions.

Several states can affect the results of the shut-in survey, including salt spalling from the cavern roof, beginning and ending shut-in pressures and temperatures, cavern shrinkage, and gas migration.

Salt spalling is a serious condition which can have deleterious effects on the cavern. Certainly one problem is damage to the well tubulars. Minor damage to the well bore casing may prevent the introduction of temperature and pressure tools to the cavern, while major damage may restrict or prevent gas flow into and out of the cavern.

Beginning and ending pressures generally have the most effect on the appearance of the graphs. Data errors caused by a short stabilization period cause a severe sawtooth-shaped graph.

This shape is the reason several surveys are graphed and comparisons made. The graph can be "smoothed" over a period of time, for instance, by a least-squares fit of the curve.

Major discrepancies in the graph caused by shrinkage and migration are the focus of the studies. The curves tend to decline with time, indicating shrinkage and variance. This decline is used to develop conclusions concerning the caverns.

Obviously, a severe declination would indicate that immediate remedial action be taken to correct an existing problem. In most cases, however, concerning inventory variance, a periodic write-off of a small amount of gas may be all that is necessary (Fig. 5).

A small amount of declination can be expected in observations of cavern-shrinkage graphs (Fig. 6). A management decision must be made when caverns have become too small and releaching is required.

Care must be taken to exercise patience when evaluating a new leaching program because shrinkage is far worse during initial operation. Older caverns tend to shrink less, probably due to the drying effect of the gas on the walls of the cavern.

Reducing the water content of the salt may induce some brittleness. Consequently, the plasticity of the salt may be reduced, but this does not mean that salt creep will be completely arrested. The oldest salt-dome caverns in the U.S. continue to shrink, although at a slower pace.

Data presented in Table 1 and the resulting calculated information depict a typical cavern-sizing survey. The cavern was shut-in after a withdrawal operation and allowed to stabilize. Every other survey was conducted in this same manner (i.e., the initial shut-in followed a withdrawal operation).

Bottomhole pressure and temperature data were recorded, and the cavern was filled with gas. The cavern was again shut-in and allowed to stabilize (Fig. 2). Subsurface readings were again taken, and the resulting data were used to calculate volumes according to the material-balance equations (Equations 1, 2, and 3 in accompanying equations box).

The form of the material-balance equation used to calculate the cavern pore volume is shown in Equation 1. The equation is modified to obtain GIP (Equation 2).

Further modification yields maximum possible GIP for cavern-shrinkage estimates (Equation 3).

Initial calculations from Equation 1 produce a pore volume in cubic feet occupied by the cavern; barrels of space available are calculated by simple conversion.

Cavern shrinkage can be estimated from this result. For reporting purposes, however, barrels of space are converted to maximum possible gas in place and depicted in cubic feet. The volume of natural gas stored should be reported in cubic feet.

It is important to remember, however, that the result of the calculation of the material-balance equation does not compare exactly with measured volumes of inventory or capacity. Data errors will cause alarming dissimilarities.

As previously mentioned, however, the errors tend to smooth when graphed trends are developed (Fig. 5). Inventory variance and cavern shrinkage may be estimated by use of these graphical trends. Precise adherence to a sound procedure and the use of quality tools for measurement should yield data which in turn produce results with an acceptable margin of error.