TEST IMPROVES MEASUREMENT OF CEMENT-SLURRY STABILITY

Feb. 12, 1990
Christopher Greaves, Ashley Hibbert BP International Ltd. Sunbury-on-Thames, U.K. New emphasis must be placed on a cement's settling stability. Many slurries have been found susceptible to settling if designed incorrectly. These slurries include those containing fluid-loss additives, dispersants, large quantities of retarder, combinations of these, and slurries designed specifically for turbulent flow.
Christopher Greaves, Ashley Hibbert
BP International Ltd.
Sunbury-on-Thames, U.K.

New emphasis must be placed on a cement's settling stability. Many slurries have been found susceptible to settling if designed incorrectly.

These slurries include those containing fluid-loss additives, dispersants, large quantities of retarder, combinations of these, and slurries designed specifically for turbulent flow.

BP International Ltd.'s Sunbury Research Centre developed a testing method that has had considerable impact on the cement-slurry designs used in North Sea. The test was first used to optimize the slurry design for a horizontal liner that was cemented subsequently with success.' BP is now using the method worldwide to measure settling stability in preference to the API operating free-water test.

NEED FOR THE TEST

It is well known that good displacement practices minimize the occurrence of uncemented channels which are produced when cement bypasses the mud during primary cementing. Perhaps not so well known is that uncemented channels can also be created after the cement has been displaced but while it is still fluid.

These channels arise from cement slurries that are unstable and produce large quantities of excess fluid when static. The excess fluid is normally apparent as free water.

This excess fluid will be particularly harmful in a deviated or horizontal well where it can collect along the high side of the annulus 2 3 and form a low density channel. This process can contribute to zonal communication and to gas migration.

Therefore, by minimizing the quantity of free water, the risk of annular communication and gas migration should be minimized .3 1

The API operating free water test' is designed to screen cement formulations and prevent those formulations exhibiting high free water from being used in the field.

Laboratory testing at BP's Sunbury Research Centre has shown, however, that many cement formulations, in particular those designed for cementing across production/test zones, exhibit very low free water according to the API test, but produce large volumes of "free fluid" when tested according to BP's settling stability test.

This free fluid is mainly water with a low concentration of fine cement solids and is as equally undesirable as free water since it does not set. It will remain as a noncementitious channel in a well bore. Therefore, many slurries that are acceptable according to the AP[ test have been shown to be unacceptable with the BP test.

The BP test provides a more appropriate method of measuring downhole cement stability. Some of these unstable formulations produce a volume of free fluid greater than 20% of the cement-slurry volume.

SETTLING AND FREE WATER

In dilute suspensions, the settlement of solid particles is described by Stokes' law. Individual particles settle at a rate determined by their size and density. In concentrated suspensions, such as Class G cement with 44% wt/wt of water, the particles generally settle in a hindered manner where all the particles, regardless of their size, settle at the same rate, therefore maintaining the same relative positions to each other .6-11

During hindered settlement, the settling rate of most particles is much less than the rate at which they would settle individually in a dilute suspension, hence the term "hindered settlement." Interparticle cohesion holds the cement particles together and keeps them in the same relative positions during settling.

The reduced settlement rate is therefore a result of a combination of the high solids concentration and the interparticle cohesion. Together this combination restricts the flow of free water upwards as the particles settle. When cement settles in this manner, a layer of clear free water will develop at the top of a column of the slurry. The more free water produced, the less stable the suspension.

For simple cement slurries, such as cement plus water only, this phenomenon is observed and measurement of the clear free water at the top of a cement column will give a relative measure of slurry stability. The greater the free water, the less stable the cement slurry. This is the basis of the API test where the cement slurry is poured into a 250 ml measuring cylinder and, after standing for 2 hr, the volume of free water at the top of the cylinder is measured.

More complex cement slurries, those containing additives such as retarders, dispersants, fluid-loss additives, and combinations of all three, can modify the cohesive forces which maintain the particles in the same relative positions during settling.

With most formulations containing a fluid-loss additive and dispersant and/or a retarder, these cohesive forces are often reduced. This usually coincides with reduced low-shear rheology, yieldpoint, and gel strength. In some instances the effect can be so dramatic that the cement particles no longer settle together and the coarser particles settle at a faster rate leaving the finer particles behind at the top of the cement column. This is called differential settlement.

The effect of this differential settlement is shown in Fig. 1 which shows a particle size distribution for unstable cement-slurry samples taken from the top and bottom of a 250 ml measuring cylinder during testing for free water.

It is clear that the cement has not settled in a hindered manner but in a differential manner where coarser particles have settled much more rapidly than finer particles. There is, therefore, a greater concentration of coarser particles at the bottom of the cylinder and a reduced concentration of coarser ones at the top.

When differential settlement occurs, the development of free water is dependent on settlement of the dilute suspension of the fine particles at the top of the cement-column tube. The finer particles can remain in suspension for a long time, creating "free fluid," but clear free water may not even develop at all. Therefore, when differential settlement occurs, measurement of free water will give no measure of the slurry stability.

SETTLING TEST

As discussed above, analysis of the mechanisms of settlement indicated that the API test could not quantify the stability of an unstable slurry. Free water could not be measured at the appropriate downhole temperature if it exceeded 85 C. because the test is done at atmospheric pressure. Also the free-water tube cooled rapidly at the beginning of the test so that when the free water was measured, the cement was much cooler than the application temperature.

This is significant because many liner formulations are thicker and have a more pronounced gel structure at lower temperatures than at the bottom hole circulating temperature (BHCT) for which they are designed. The development of gel strength is a reflection of the cohesive strength between the cement particles which resist settlement. Therefore, reducing the test temperature will generally result in reduced settlement.

The BP settling test was developed to perform relative measurements of cement formulation stability and to do this at elevated temperatures and pressures. In the settling test, the cement slurry is allowed to set hard, permitting both free fluid and free water to be distinguished and measured. Therefore, slurries which experience hindered settlement or differential settlement can be measured for their relative stability using this test method.

A cement slurry is heated in a pressurized consistometer to the test temperature (BHCT), poured into a settling tube, and then cured overnight for 16 hr at BHCT. If the BHCT is greater than 85 C. the slurry is heated to 85 C. and poured into the settling tube, which has been preheated to 85 C. The tube is then heated to the BHCT in the curing chamber. After the cement has set, the stability is determined by:

  • Measuring the height of the set column of cement and comparing it with the original height (tube length). This measures the amount of settlement.

  • Cutting the cement column into sections and determining the density along its length.

The BHCT is used as the test temperature since settlement can only occur while the cement is fluid, typically up to 6 hr. The cement will spend most of its time at or around the BHCT during this time. If the slurry is designed for a long liner then it is appropriate to test at the previous casing-shoe temperature if this is lower than the BHCT.

The settlement tube, illustrated in Figs. 2 and 3 consists of a brass tube (height 203 mm, ID 25 mm) that is split into two halves so that the set cement column can be removed. The halves of the tube are clamped together and screwed into a brass base. All surfaces are coated with a thin layer of grease for a watertight seal and to ease disassembly.

A brass lid (not shown) is placed on top of the tube to prevent the cement-column top from being disturbed while allowing the water in the curing cell to communicate hydrostatic pressure to the cement via a groove on the underside of the lid.

The dimensions of the settling tube were chosen so that the test could be performed with convenience in standard cementing laboratory equipment and also to be of sufficient size to allow clear distinctions between stable and unstable cementing formulations. It is important that the dimensions of the tube used in the setting test are recorded and that only tests performed with tubes of the same dimensions (diameter and length) are compared. This is because the rate of settlement and amount of total settlement will be different for settling tubes of different dimensions.6 7

Aeration of the cement slurry must be minimized. Entrained air in the cement slurry will be compressed when the cement is pressurized in the curing chamber. This will cause the cement column to shrink (apparently settling) because the entrained air is compressed. In some tests performed without removing the air from the slurry, up to 10 mm of settlement has occurred. To prevent this, the slurry should be placed under a partial vacuum for 1 min using a waterdriven vacuum pump.

STABLE AND UNSTABLE SLURRIES

Results of settling tests are illustrated in Fig. 4. The amount of settlement and the density along the column of the set cement has been determined.

Slurries 1 and 2 are relatively stable slurries that have settled in a hindered manner. This is evident from the presence of clear free water between the top of the cement and top of the tube, the relatively low amount of total settlement, and the constant density along the column. There is a small zone of increased density at the bottom of the tubes. These observations are similar to those reported by Powers 6 who made the following observations on hindered settling of cement pastes:

  • A zone of clear free water at the top.

  • A zone of uniform density, equal to the original density.

  • A compressed zone at the bottom of the column with a density gradient increasing towards the bottom of the column

The densities we have observed along the column are typically a little greater than the original density. We have also observed that the density of the set column increases for higher levels of free water. This is not entirely in agreement with the observations of Powers but may be explained partly by the compression zone rising to the cement-column top because of the relatively small length of our settling tube and partly by an increase in density caused by absorption of water during the set. Hydrating cement increases the mass without increasing the bulk volume.

Slurries 3 and 4 are examples of unstable slurries. The instability is shown by the high settlement of the column; the presence of a sludge of water and fine cement particles (free fluid) at the top of the set column; and a density gradient along the entire column. This is the result of differential settlement where the coarser particles have settled at a much faster rate than the finer particles. The finer particles have formed a sludge of noncementitious free fluid at the top of the set cement. Slurry 4 is less stable than Slurry 3 because of the larger settlement and greater density variation along the column.

LINER SLURRIES

The BP settling tests and API tests performed on slurries designed for field use are listed in Table 1. Representations of three of the cemented columns removed from the tubes are shown in Fig. 5. Details of the formulations are given in Table 1.

Formulation 5 is clearly unstable. There is a large amount of settlement some sludge at the top of the tube, and a density gradient along the column. On the other hand, the API test on this slurry gave a result of 0.0 ml which, according to this test method, means that the slurry is stable.

There are two reasons, which have been discussed previously, for the discrepancy between the results of the BP and the API test:

  1. The different test conditions

  2. The inability of the API to detect or measure stability if differential settlement occurs.

A similar comparison of the BP and API tests was obtained for Slurry 6; however, the discrepancy between the two tests is even greater. Of the five formulations tested only one gave any indication of instability from the API test. Slurry 7 produced dark bands of settled solids at the bottom of the free-water tube and a light band of a fine-cement solids in suspension at the top of the free-water tube.

Slurries 5 and 6 were improved by redesigning the cement formulations, to 5a and 6a. Slurry 5 was improved by adding sodium silicate (cement extender) which improved the gel strength of the slurry at BHCT. Slurry 6 was improved by increasing the fluid-loss additive concentration and removing the retarder. In this instance, the combination of the high-temperature retarder and fluid-loss additive dispersed the slurry to the extent that it was highly unstable.

Removing the retarder is not always possible or desirable, but reducing the retarder concentration with an appropriate increase in fluid-loss additive concentration to maintain the thickening time can improve stability of some cement formulations.

The above are just two examples where liner-slurry formulations used in the field have been modified to improve settling stability. The use of additional chemicals to counter the detrimental effects of others, such as adding cement extender to Slurry 5 to obtain Slurry 5a, is not recommended and should only be used as a last resort.

This method may encourage the unnecessary use of other chemicals such as dispersants. Other methods that have improved stability include alternative fluid-loss additives, slurries designed for stability in preference to turbulence, higher-fluid loss (100-150 ml), slurries, dispersant reduction or elimination, and higher slurry density.

Based on field experience, a settling test result of 5 mm total settlement or less is considered to be a good design figure for a liner cement slurry. It may often be difficult to achieve this without adversely affecting other properties; however, it is a good target figure. More importantly, differential settlement must be avoided because of the relatively large volume of noncementitious fluid created when cement settles in this manner.

OTHER APPLICATIONS

The results previously discussed are for settlement of liner formulations designed with fluid-loss additives. In our experience, cement formulations designed for most types of cement jobs are susceptible to settling because of bad slurry design. This often goes unnoticed because the API tests frequently provide misleading results.

Other formulations that are stable according to free water tests but unstable in the settling test are slurries containing dispersants, slurries designed for turbulent flow, and slurries containing relatively large quantities of retarders.

For example, the use of unnecessary quantities of dispersants can result in very high settlements This is illustrated by the results in Table 2 where the effect of adding excess dispersant increases the settlement to very high and unacceptable levels.

The effect of using the same quantity of dispersant with a cement that is more sensitive to additions of that dispersant can also be seen in this table. As a further example, Slurry 8 in Table 1 is a cement-plug slurry designed for field use where the API test was used to measure stability.

Because the test involves sectioning the cemented column, it is a particularly useful method for examining stability of formulations containing weighing agents such as hematite. While a cement formulation may be stable enough to prevent any development of free water, it may not be able to prevent settlement of weighing agents.

Density measurements along the column length will determine to what extent the weighing agent has settled.

ACKNOWLEDGMENT

The authors would like to thank British Petroleum for giving permission to publish this work.

REFERENCES

  1. Buchan, R., and Little, M.T.S., "innovative techniques to improve liner cementation in North Sea wells an operators experience, SPE 15896, 1986.

  2. Coulson, J.M., and Richardson, J.F., Chemical Engineering, Vol 2, 3rd edition, 1980, p. 178.

  3. Webster, W.W., and Eikerts, J.V., "Flow after cementing - a field and laboratory investigation," SPE 8259, 1979.

  4. Davies, D.R., Hartog, J.J., and Stewart, R.B., "An integrated approach for successful primary cementations," SPE 9599, 1981.

  5. "API Specification for Materials and Testing for Well Cements," Specification 1 0, 4th edition, 1988.

  6. Powers, T.C., Properties of fresh concrete, Wiley, London, 1968.

  7. Wheatland, A., "Sedimentation," Chemistry and Industry, Feb. 6, 1982, pp. 81-86.

  8. Raffle, J.F., "Pressure variations within concentrated settling suspensions," J.Phys. D: Appl. Phys, 9, 1976.

  9. Michaux M., and Defosse, C., "Oil well cement slurries 1. Microstructural approach of their rheology," Cement and Concrete Research, 16, 1986, pp. 23-30.

Copyright 1990 Oil & Gas Journal. All Rights Reserved.