Reference value developed for mechanical integrity of storage caverns

Fritz Crotogino
Kavernen Bau- & Betriebs-GmbH Hannover

The supporting research was funded and its results received technical review by the Solution Mining Research Institute, Deerfield, Ill., an association of the salt-cavern industry.

Salt caverns play important roles in large-scale storage of hydrocarbon gases and liquids. Required for safe and economical operation of these storage caverns is verification of the external mechanical integrity of the access (injection and withdrawal) wells.

Hydrocarbons leaking from the wells represent a safety risk on the surface, can contaminate groundwater, and result in substantial loss of product.

For many years, standard practice worldwide in construction of storage caverns has been to use a gas-interface test to verify the external integrity of the cavern wells (Fig. 1 [20624 bytes]). (External integrity refers to the tightness of the cavern, including the transition zone from the cased to the open well in the top section.)

A broad range of different procedures and particularly evaluation criteria exist, however.

The program, underwritten by SMRI, had the following goals:

• Provision of an overview of current practice

• Development of an SMRI reference for external well mechanical-integrity testing with respect to performance, data evaluation, and assessment.

The storage cavern operators expected to gain the following:

• Comparability between methods and assessments

• Aid in influencing the movement towards standardization by regulators

• A firm technical base for use in litigation between the operator and other parties.

The intention here was to provide test companies and operators with a platform from which they could perform and, in particular, compare their tests with international standard practice.

Selection of the actual test method, however, remains the individual operator's responsibility.

The investigations concentrated on those tests which quantify the tightness of the final casing and the cementation with the rock mass.

These interface tests are mainly characterized by injection of a limited volume of test medium into the annulus of the access well. This medium is lighter than the liquid in the cavern at the time of the test.

After a stabilization period, such parameters as pressure and interface depth vs. time are measured to indicate the integrity of the well.

Integrity testing

Mechanical integrity of a storage cavern can be defined as the technical tightness of the casing, tubing, and packer with no fluid movement either along the cementation or into formations adjacent to the well bore.

This condition protects the ecosystem from uncontrolled escape of product, assures safety of personnel and surface installations, and achieves loss-free storage.

As stated, the SMRI-sponsored study focused on verification of external integrity, especially of the access well in the particularly sensitive transition zone where the open hole meets the cemented cased well.

At this point, the risk of leakage is greatest because the different materials which interface at this point (rock salt, cement, steel) are all subject to the same mechanical stress levels.

Focusing the tests on the well as opposed to the entire cavern leads to substantial improvement in verification accuracy. The reason is that the volume of the well is generally less than that of the cavern by three or four orders of magnitude.

Current practice

A realistic assessment of the external integrity of an access well requires that the well's tightness be quantified. By comparison, the meaning of qualitative methods (for example, observing the pressure loss of a shut-in well) is limited.

In the past, the measure for quantitative assessment of tests was the accuracy of the pertinent test procedure (minimum detectable leak rate, MDLR): If the results were within measurement accuracy, then the well was considered technically tight.

As far as hazards above and below ground are concerned, however, it is not the minimum detectable leak rate which is important but the maximum admissible leak rate (MALR).

Therefore, the strategic objective of the SMRI research project was to develop a reference value for MALR.

All the test methods described presently have in common assumptions of a liquid-filled cavern, injection of a test medium lighter than and immiscible with the product, and a test pressure near the maximum operating pressure.

The differences are in area tested, equipment and procedure, test medium, and MDLR and MALR.

For comparison of the accuracy (MDLR) of the various methods, a model cavern was configured reflecting the layout of a typical gas cavern in a salt dome (Fig. 1).

• The in situ balance method is by far the most frequently used method and, in the U.S., is usually referred to as the nitrogen-interface test (Fig. 2 [21990 bytes]).

  The procedure for this method involves:

  • Injection of test gas in the annulus to below the final casing seat

  • Shutting in the gas

  • Periodic measurement of head pressure gas-product interface depth (and possibly temperature), and subsequent calculation of shut-in mass of the test gas

  • Evaluation of changes to the shut-in mass of the test gas.

  The advantage of this method is its straightforward and simple evaluation. The method's disadvantages are low accuracy at large open-hole diameters because of the interface depth determination by logging and the logging costs.

• The in situ compensation method was developed by UGS, a German subsidiary of KBB.

  This test uses special equipment which allows the test-gas volume to be reduced and, above all, the gas-product-interface to be fixed at a defined depth (Fig. 3 [22598 bytes]).

  Typically, the test gas is injected into the annulus until it reaches the weep-hole in the inner pipe. Minute amounts of gas rising to the surface within the drill pipe indicate that the interface position is at the weep-hole.

  If, after shut in, the gas pressure changes in the cavern or if there is a leak, the interface depth would therefore rise. The level is then reset by injection of additional gas which is measured with great precision at the surface.

  Advantages of this method are the following:

  • Substantial improvement in measurement accuracy which for the most part no longer depends on the diameter of the open hole

  • The upper section of the final casing being protected from pressure loads by the test gas

  • Calculation of loss rates being nearly continuous.

  This final aspect enables any trends in the development of a leak rate to be determined and hence, under certain circumstances, allows the test period to be curtailed.

  Disadvantages are primarily the high costs for installation of the test equipment and the need for a workover unit throughout the test period.

• The principle behind the aboveground balance method is comparison at the surface of the injected test fluid and the test-fluid volume recovered after the test (Fig. 4 [21684 bytes]).

The fact that the volume measurement can be carried out at the surface with great accuracy pre and post-test means there is no need to perform measurements below ground, and costs are therefore reduced.

In the mid-1980s, this method using nitrogen was patented in the U.S. In practice, however, gas oil is generally used today.

The advantages of this method are simple installation, no need for well logs, and straightforward test evaluation. Also, the use of gas oil, a liquid, helps reduce the pressure on the final casing in the upper section of the well.

The disadvantages of using gas oil are the reduced test accuracy (MDLR) and the possible source of error due to mixing with unrecovered blanket fluids.

Fig. 5 [16533 bytes] provides a comparison of the measurement accuracy of the three methods.

Bases for these values are data drawn from the model cavern (typical high-pressure gas cavern in a salt dome; Fig. 1 [20624 bytes]). Because the test accuracy of the various methods is initially time independent, an absolute minimum detectable leak is specified rather than a rate.

The MDLR is then derived by dividing by the test period (shut-in time). The comparison clearly highlights the high accuracy of the in situ compensation method (ISC) and the great dependency of the in situ balance method (ISB) on the diameter of the well.

SMRI reference value

As mentioned, a contradiction exists between the widespread use and the importance of the cavern integrity tests, on the one hand, and the range of different methods used in carrying out the tests, particularly evaluation and assessment, on the other.

As soon at it became clear that the setting of standards, guidelines, or even merely recommendations was unrealistic, the study was restricted to providing a reference with regard to performance, data evaluation, and assessment.

Reference means that both operators and contractors are able to assess the individual tests by comparing results with current standard engineering practice. Each remains free to select his preferred method.

In the definition of a reference, no preference was given to any specific method. Instead a list of basic requirements was established with respect to:

• Time of test (before start or after end of solution mining)

• Adequate pressure and temperature stabilization period before testing

• Variables to be evaluated (pressure drop, volume, or mass loss of test medium).

As for the important matter of determining the evaluation assessment criterion (MALR), two principally different approaches are available: the scientific and the pragmatic approach.

Determining the assessment criteria scientifically would involve establishing a quantitative relationship between leakage rate and effects on the hydrosphere and biosphere, safety and economic consequences, and the assessment of such interactions.

This approach would, therefore, require comprehensive modeling and assessment, and the results would thereby be site-specific and nontransferable. Because of the disproportionate amount of effort involved, this procedure was dropped.

The alternative is then the pragmatic approach. The basis for this approach is the positive experience gained in the performance of gas-interface tests over more than 10 years.

As a basis for reference MDLR, the majority of interface tests carried out are based on the in situ balance method. Using standard instruments and evaluation methods, a time-independent MDL (minimum detectable leak) for the model cavern (Fig. 1) is calculated to be MDL = 37 kg (nitrogen) 40 kg (nitrogen).

The mass of 40 kg nitrogen corresponds to 0.2 cu m geometrical volume at 170 bar and 300° K. In the cases of more viscous test fluids like LPG or gas oil, the MDL increases by a factor of approximately 2-3.

Typical 24 hr and 48-hr test periods would yield the following:

MDLR = 40 kg/d (N2) @ Dt for 24-hr test duration; MDLR = 20 kg/d (N2) @ Dt for 48-hr test duration.

This value for the MDL of 40 kg (N2) is based on the evaluation of questionnaires sent to storage-cavern operators worldwide: By maintaining this accuracy, all caverns deemed to be technically tight based on the results of one of the three test methods discussed (in accordance with current engineering practice) were able to fulfill the following conditions for cavern well integrity:

• No detectable contamination of the hydro or biosphere

• Protection of subsurface installations

• Safety on surface

• No detectable loss of product.

Objective; performance

The objective of the reference value is verification of the external mechanical integrity of the cavern well. The area covered is determined by final casing (complete or partial), cement bond in casing shoe area, and uncased well.

The interface test is to be performed at completion of solution mining and before storage operations begin. (Repeating the test during operations is up to the operator or regulator.)

With respect to the requirements on method, reference is only made here to the use of appropriate pressure sensors as essential for quantitative evaluation in terms of mass vs. time, or at least volume vs. time.

Indirect methods, without running interface logs and based only on monitoring of the pressure drop over a time period, are not recommended.

During the test, geometrical data of the uncased well should be acquired, preferably with sonar or caliper logs. Also possible is indirect determination by injection of the test medium into the uncased well.

The test medium should be injected close to the average well temperature, and the test consists of at least three independent measurements.

Normally, in measurement of technology, the admissible value is first defined; then, based on this value, the necessary accuracy of the method is defined. In order to determine a sensible limiting value, it is necessary to ensure that the value for the accuracy is at least a factor of three smaller than the criterion itself.

In this case, the problem lies in the fact that previous statements were almost exclusively restricted to the accuracy of the method but not to the limiting value or did not appreciate the essential difference between accuracy and limiting value.

It was thus necessary to reverse the normal order when establishing the reference values for MDLR and MALR in the following steps:

1. Determination of the reference MDLR based on calculations and practical experience

2. Determination of the MALR in relation to the reference MDLR.

• MDLR. The positive experience with the widely used in situ balance method was described previously.

  With parameters from the model cavern (Fig. 1), a theoretical MDLR of 40 kg (nitrogen)/day was calculated. From the underlying principles in its determination, the following SMRI reference MDLR is proposed as the value: reference MDLR = 50 kg/d, based on nitrogen as the test medium.

• MALR. The recommended reference MDLR of 50 kg/d and the constraint that the smallest possible reliably determinable measured value must be at least a factor of three larger than the measuring accuracy yields the following value for the MALR:

  3 x MDLR = 3 x 50 kg/d = 150 kg/d, based on nitrogen as the test medium.

  The mass of 150 kg corresponds to 0.8 cu m geometrical volume at 170 bar and 300° K.

  These statements on MDLR and MALR were based on nitrogen as test medium because of its low viscosity and, thus, small MDLR which is equivalent to high sensitivity.

  The possible hazard resulting from a leak must reflect the leakage rate of the specific product and not the test medium. In other words, the admissible value must be converted to the rate for the product itself.

  The reference MALR of 150 kg/d refers to nitrogen gas as the test medium. The actual MALR for more viscous storage products (LPG, gas oil, crude oil) is reduced by the following factors depending on the product: LPG, 2; gas oil, 3; and crude oil, 10.

Based on a presentation to the Spring 1996 Meeting of the Solution Mining Research Institute, Houston, Apr. 14-17.

The Author