NOVEL PROCEDURES ACCURATELY MEASURE DRILLING MUD DYNAMIC FILTRATION

Martin E. Chenevert, Said Al-Abri
University of Texas at Austin
Austin

Liang Jin
Stim-Labs Inc.
Duncan, Okla.

New equipment and test procedures can determine dynamic mud cake properties such as equilibrium cake thickness, porosity, permeability, compressibility, and erosion resistance.

The following were developed to study dynamic filtration: a dynamic filtration cell; a recommended filtration medium; a mud cake thickness device; mud cake porosity determination method; calculation methods for shear rate determination beneath a rotating cone; determination of equilibrium cake thickness, erosion resistance, and compressibility; and preferred filtration display techniques.

Filter cakes deposited on the well bore wall play important roles in drilling. Filter cakes help control the amount of filtrate entering a formation, thus controlling formation damage. Additionally, inferior filter cakes are responsible for problems associated with the differential pressure sticking of drill pipe.

Evaluation of filter cakes has been limited in the past primarily to the study of static filtration. Results obtained under static filtration conditions are not the same as what occurs during dynamic filtration; however, few studies have been performed to date on dynamic filtration because of the lack of suitable test equipment.

Only recently has equipment been commercially available for the study of dynamic filtration. In most cases, technical detail for the complete analysis of the quality of the filter cake is not provided.

In particular, the shear rate across the filtration face is often not known nor easily controlled. In one study, the shear rate was controlled, but the mud cake was difficult to remove and analyze. 1 2 Shear rate and shear stress play a major role in the type of filter cake deposited. 3-6 Many researchers have pointed out the need for dynamic filtration information. 7-11 The accurate monitoring of dynamic filtration can lead to a model which predicts filtration into low permeability formations. Dynamic filtration results presented by Dewan were obtained using the experimental equipment and techniques described in this article. 12

The proper study of dynamic filtration requires the following:

A dynamic filtration cell
Mud flowing across a filtering surface at a known shear rate
Accurate monitoring of the fluid produced
Some method for determining the properties of the mud cake.

DYNAMIC FILTRATION CELL

The dynamic filtration cell developed consists of a rotating cone submersed within a pressured mud chamber located near the surface of the filtration medium (Fig. 1). The rotating motor, cone, and mud are housed within a chamber with nitrogen pressure applied and controlled during testing.

The motor speeds are monitored with a tachometer located within the pressure chamber, and the signal is recorded on a computer. Electrical conduits for monitoring the motor's speed and for driving the motor pass through a Conax electrical fitting located in the top of the pressure chamber.

A baffle, located near the mud/air surface, provides an upper barrier to prevent the formation of a cup as the cone rotates, thereby minimizing mud/nitrogen mixing. An O-ring seals the baffle at the walls of the chamber. A 1/8-in. space between the baffle and the shaft allows mud to pass during filtration. The filtration medium is located just below the rotating cone.

For the test, the filtration medium is saturated with water containing 3 wt % KCl. The mud chamber is then filled with drilling mud, the unit is assembled, the motor is set into constant rotation, and pressured nitrogen is supplied through the air inlet port. The constant nitrogen pressure exerted on the top of the mud attempts to force the mud through the filter medium. A portion of the solids associated with the mud filtrate is deposited on the surface of the filter medium, which thereby forms a dynamic mud filter cake.

The filtrate expressed during the test is collected in a beaker located on top of a Sartorius digital balance. The filtrate weight is continuously recorded, as a function of time, on the computer acquisition system.

FILTRATION MEDIUM

The filtration medium consists of a 1/4-in. thick, low-permeability (0.2-1.0 md) ceramic disk covered with a single sheet of Watmans No. 50 filter paper. Soilmoisture Equipment Corp. supplied the low permeability ceramic disks, which minimize rock/fluid reactivity. Disks of approximately 1.0-md permeability were used to facilitate modeling of filtration into low-permeability reservoirs.

A single layer of filter paper was placed on top of the disk to serve two purposes:

To minimize the spurt loss by assisting in the filtration of the mud, thereby preventing many fine particles from invading the core
To allow for easy removal of the cake after the experiment, so that cake thickness and properties could be evaluated.

In these experiments, the objective was to understand the mud and filter cake characteristics that control filtration and not to study the role of rock permeability in filtration control. Therefore, the use of a constant-permeability disk with one layer of filter paper was acceptable.

Careful placing of the saturated disk in the pressure vessel was necessary to obtain an accurate measurement of spurt loss. To ensure that no air was present between the mud and the weight balance at the start of the test, the disk was first vacuum saturated in a desiccator, then placed with the wet filter paper in the filtration cell with 1 cm of 3% KCl water on top. The 3% KCl solution was then allowed to drain until only a thin film of liquid remained on the filter paper. The mud was then poured into the pressure vessel, and the test was begun.

Vacuum saturation of the core was very important. Several ceramic disks were kept vacuum saturated and submersed within a 3% KCl solution.

At the end of a test, the filter paper containing the mud cake was removed, and the ceramic disk was backflushed with 3% KCl water. Backflushing removed any fine particles that invaded the disk and also allowed for the measurement of the disk permeability. If the disk permeability was permanently reduced to less than 0.2 md, a new disk was used.

The 3% KCl solution, used for initial saturation and for backflushing, prevented swelling of any particles that became lodged in the ceramic disk.

In earlier work, various filtration tests were performed with low permeability disks made of natural sandstones. Most low permeability sandstones studied contain sufficient amounts of clay to cause irreversible permeability damage during the filtration experiment.

Such damage made testing of mud cake properties quite difficult. The ceramic disks were used to prevent this damage. Formation damage caused by mud filtrate into "dirty" sands is an important problem; however, this problem is not the subject of the current study. The equipment and techniques described here can be easily modified to study formation damage.

At the end of a filtration test, the filter paper containing the mud cake is removed, and its thickness and porosity are experimentally determined.

MUD CAKE THICKNESS

Many mud cakes are 85% water and are very fragile. Therefore, handling them and measuring thickness is problematic. Mud cake surfaces often contain a layer of mud (about 96% water), and the interface between the mud and the cake is very difficult to determine. Many investigators shun away from cake thickness measurements because of this difficulty. Consequently the determination of cake permeability is not possible.

A low stress penetrometer was developed to measure cake thickness accurately (Fig. 2). The filter cake is placed on the lab jack and positioned so that the foot rests on the edge of the filter paper only (next to the mud cake), and the Digimatic indicator is set to zero. The foot is then raised with the rod lifting cable. the filter cake is moved beneath the foot, and the foot is gently lowered until it contacts the cake.

The measurement indicates the thickness of the cake (which possibly contains some mud on top) at a force level of 135 g, which is the force exerted by the indicator spring. Weights, in 150-g increments, are then applied to the upper plate, and the vertical position of the foot is recorded. The cake is then repositioned under the foot for additional measurements of cake thickness.

Cake thickness is plotted against load applied (Fig. 3). The plot is extrapolated back to zero load to estimate the original cake thickness.

This device allows more accurate ( 0.0001 cm) determination of the surface between the mud cake and the mud. Standard API (American Petroleum Institute) testing procedures simply recommend the use of a ruler to measure this distance.

MUD CAKE POROSITY

After the determination of the cake thickness, the cake is gently removed from the filter paper with a spatula and placed in a weighing dish. The porosity is determined by a "gravinometric" technique. The mud cake is weighed in its wet condition. The cake is then dried overnight at 150 F. and its dry weight is measured. The porosity (water content) of the mud cake is then determined by Equations 1 and 2.

This wet/dry method for obtaining cake porosity assumes that there is no mud left on the mud cake and that the density of the solids in the cake is known. For muds containing solids of varying densities (such as those containing barite), this method does not work. Also, if the cake is very soft and fragile, it is difficult to handle, and large variations in the results are obtained.

SHEAR RATE CONTROL

The novelty of the filtration device is the use of a cone to achieve relatively constant shear rates (of a known value) above the filtration medium. By keeping the cone at a constant rotating speed using a Compumotor electronic control device, a relatively constant shear rate across the surface of the mud cake could be maintained during the filtration test.

The concept of a constant shear rate cone-and-plate device is not new. Cones with angles of less than 3 are often used in viscometers for achieving constant shear rates.

In this study, such low-angle cones were not feasible because of the development of a mud cake which closes the space between the cone and filtering surface. It was therefore necessary to use larger angles (15-30) and to develop equations which describe the shear rates below such cones as the mud cakes build up. The two configurations studied had the cone nearly touching the filter paper (no-offset case) and the cone raised 0.1 cm above the filter paper (offset case).

No-offset case

For cakes expected to have a final thickness of less than 0.1 cm, the no-offset arrangement was used, and a small circular section of the paper (0.5 sq cm) was sealed in the center with impermeable tape. This taped area provides a tough surface to prevent the tip of the cone from wearing a hole through the soft filter paper. The tape reduces the total filtration area by only 2%.

As the cake grows in thickness, the gap between the cone and the cake decreases, and the shear rate across the mud cake increases (Fig. 4).

The shear rate at a Point A on the surface of a mud cake is determined by Equation 3. Using the geometrical configuration shown in Fig. 4, the area-weighted average shear rate across the filtered surface can be determined with Equation 4.

At the start of an experiment when there is no filter cake (Tc = 0 cm), Equation 4 reduces to the standard cone-and-plate equation (Equation 5).

Fig. 5 shows calculated shear rate results using Equation 4 and a 15 cone for cases of no cake (Tc = 0 cm) and for a 0.1-cm cake (Tc = 0.1 cm). As shown in Fig. 5 for Tc = 0 cm, a constant shear rate exists over the entire filter paper. As the cake builds up, a distortion takes place.

For Tc = 0.1 cm, the high shear rates at low radiuses are not excessively disruptive; only a small percentage of the total filtration area experiences the very high shear rates. On an area-weighted basis, the Tc = 0.1 cm case has an average shear rate of 502 sec-1, which is a 25% increase above the 400 sec-1 case for Tc = 0 cm.

Offset case

For muds with high fluid loss characteristics, excessively thick filter cakes develop. Therefore, the cone should be offset at the start of the test to give the cake more room in which to build up (Fig. 6).

Using the same approach for deriving the average shear rate across the filtered surface as in the no-offset case, the derived governing equation is given by Equation 6.

Fig. 7 shows the calculated results using Equation 6 and a 15 cone with a 0.1 cm offset (h0 = 0.1 cm). For the no-cake case (Tc = 0 cm), the 0.1-cm offset causes the shear rate to drop off completely near the center of the cone. On an area-weighted basis, however, the average shear rate for this case is 340 sec-1, a 15% decrease in shear rate compared to the no-offset case of 400 sec-1.

The curve labeled Tc = 0.1 cm in Fig. 7 represents the case in which the cake has built up by 0.1 cm. The gap has been essentially eliminated, and the situation reduces to the no-offset case with no filter cake as described by Equation 5.

TYPICAL FILTRATION RESULTS

A filtration test was performed on a freshwater bentonite/lignosulfonate mud with 3 lb/bbl of starch. During the test, four filtration steps were run to determine the mud cake's equilibrium cake thickness, erosion resistance (or lack thereof), and compressibility (Table 1).

Fig. 8 shows the results in a typical plot of cumulative filtrate volume vs. the square root of time. The test produced a good definition of cumulative filtration volume vs. time.

Such typical plotting methods are very useful with static filtration tests. For analyzing dynamic tests, such plots are not very useful because of the difficulty in observing the time at which a mud cake has reached an equilibrium thickness.

A preferred way to display dynamic filtration results is with a "slowness" plot (Fig. 9). 13 Slowness is defined as the reciprocal of the flux of the filtrate flow through a unit area (Equation 7). The volume is the cumulative filtrate produced in cubic centimeters, the time is in seconds, and the filtrate area is the flow area of the ceramic disk in square centimeters. The term has the following physical meaning: Slowness is the time for the filtrate to move 1 cm through the filtering medium (mud cake or ceramic disk).

Fig. 9 shows that for Step 1, after a filtration value of 0.36 cm occurs (volume/unit area) at a shear rate of 400 sec-1, an equilibrium cake has been established, as shown by the constant slowness value of about 18,000 sec/cm. Step 2 achieves equilibrium at a filtration value of about 0.61 cm, after the filtration rate has decreased to a slowness value of about 46,000 sec/cm. Such a decrease results from the deposition of more filter cake at the lower shear rate of 100 sec-1.

Increasing the shear rate back to 400 sec-1 in Step 3 erodes the cake, and the filtration rates therefore increase to a slowness value of 32,000 sec/cm. Finally, increasing the differential pressure to 500 psi in Step 4 initially causes more filtrate to be expressed as the cake is dewatered, but then a new slowness rate is established at 23,000 sec/cm.

Detailed analysis of this test, along with the filter cake removed after Step 4, can produce numerical values for cake thickness, porosity, permeability, compressibility and erosion resistance.

ACKNOWLEDGMENT

The authors thank the Gas Research Institute for its support (contract 5089-2601861). The authors also thank A.W. Gordy, J.T. Dewan, and S.A. Holditch for their assistance.

REFERENCES

Fisk, J.V., and Jamison, D.E., "Physical Properties of Drilling Fluids at High Temperatures and Pressures," SPE Drilling Engineering, December 1989, pp. 341-46.
Fisk, J.V., and Shaffer, S.S., "The Use of Filtration Theory in Developing a Mechanism for Filter-Cake Deposition by Drilling Fluids in Laminar Flow," SPE Drilling Engineering, September 1991, pp. 196-202.
Fordham, E.J., Ladva, H.K.J., and Hall, C., "Dynamic Filtration of Bentonite Muds Under Different Flow Conditions," SPE paper 18038, 1983.
Fordham, E.J., and Ladva, H.K.J., "Cross-Flow Filtration of Bentonite Suspensions," Physicochemical Hydronames, 1989, Vol. 11, No. 4, pp. 411-39.
Fordham, E.J., and Ladva, H.K.J., "Cross-Flow Filtration of Bentonite Suspensions II," journal of Colloid and Interface Science, January 1992, Vol. 148, No. 1, pp. 29-34.
Chenevert, M.E., and Huycke, J., "Filter Cake Structure Analysis Using the Scanning Electron Microscope," SPE paper 22208, 1990.
Prokop, C.L., "Radial Filtration of Drilling Mud," AIME transactions, 1952, Vol. 195, pp. 5-10.
Ferguson, C.K., and Klotz, J.A., "Filtration from Mud During Drilling," American Institute of Mining Engineers, Petroleum Transactions, 1954, Vol. 201.
Krueger, R.F., "Evaluation of Drilling Fluid Filter-Loss Additives Under Dynamic Conditions," Journal of Petroleum Technology, January 1963, pp. 90-98.
Bezemer, C., and Havenaar, I., "Filtration Behavior of Circulating Drilling Fluids," SPE journal, December 1966, pp. 292-98.
Wyant, R.E., Reed, R.L., Sifferman, T.R., and Wooten, S.O., "Dynamic Fluid-Loss Measurement of Oil Mud Additives," SPE paper 14246, 1985.
Dewan, J.T., and Chenevert, M.E., "Mudcake Buildup and Invasion in Low Permeability Formations: Application to Permeability Determination by Measurement While Drilling," presented at the SPWLA/CWLS Symposium in Calgary, June 13-16, 1993.
Fordham, E.J., Allen, D.F., Ladva, H.K.J., and Alderman, N.J., "The Principle of a Critical Invasion Rate and its Implications for Log Interpretation," SPE paper 22539, 1991.