DRAG REDUCING ADDITIVES IMPROVE DRILLING FLUID HYDRAULICS

March 13, 1995
Simple pipe flow tests can provide a direct method for detecting and quantifying drag reduction activity (provided care is taken in selecting flow conditions). Drag reduction activity can be determined without introducing complicated Theological parameters. These pipe flow tests can be helpful in comparing different types of drag reducing additives. They are also valuable in analyzing the effectiveness of a drag reducer for a particular drilling application.
J.G. Savins
Baroid Drilling Fluids
Dallas

Simple pipe flow tests can provide a direct method for detecting and quantifying drag reduction activity (provided care is taken in selecting flow conditions).

Drag reduction activity can be determined without introducing complicated Theological parameters.

These pipe flow tests can be helpful in comparing different types of drag reducing additives. They are also valuable in analyzing the effectiveness of a drag reducer for a particular drilling application.

The use of drag reducing additives helps drilling fluids develop a lower pressure gradient at a constant flow rate under turbulent flow conditions. At a constant pressure gradient a treated mud will flow faster than an untreated mud because turbulence is suppressed.

Correlating friction velocity with drag reduction activity closely reflects additive characteristics and performance. Friction velocity provides a method for unifying turbulent drag reducing flow data obtained under different conditions.

BACKGROUND

Near the end of World War 11, K. Mysels disclosed in patent information prepared at the U.S. Army's Edgewood Arsenal that gasoline thickened with a low molecular weight paraffin chain soap flowed faster under turbulent conditions in 1/2-2 in. pipes than the unthickened gasoline at the same pressure drop.' In 1948, B. Toms, a research chemist in the U.K., reported a benzene derivative doped with a small quantity of a common polymeric-based plastic flowed faster in a capillary tube under turbulent conditions than the untreated solvent.2 Three aspects of this turbulence suppression are noteworthy:

  • The phenomenon occurred under turbulent flow conditions.

  • The materials were significantly different in composition and structure (that is, a derivative of a low molecular weight fatty acid and a high molecular weight linear polymer).

  • Rheological characteristics were significantly different (shear-thinning Napalm jelly and a Newtonian-like dilute polymer solution with viscosity about the same as that of the untreated solvent).

The term drag reduction is applied to this phenomenon." It has also been called friction reduction.

The phenomenon, however, is unrelated to the lubricity functions of a drilling fluid (for example, reducing drillstring torque and avoiding wall sticking and bit balling). Drag reduction cannot be predicted from the description of mud rheology obtained with conventional API test procedures, and the phenomenon is unrelated to shear thinning behavior. Flow properties predicted from the Bingham, power law, Herschel-Bulkley, or Casson models will not correlate with this phenomenon.

PRESSURE DROP

This phenomenon is best described in a plot of pressure drop vs. flow rate (Fig. 1). (20915 bytes) The data in Fig. 1 were recorded from hydraulics tests on aqueous solutions in a pipe flow loop. The flow rates used produced turbulent conditions while circulating the water. Polymer-based Systems A, B, and C are, respectively, solutions of an hydrolyzed polyacrylamide, and hydroxyethylcellulose, and Carbopol 941. System D is a micellar solution forulated with low molecular weight derivitive of oleic acid, and unsaturated 18-carbon fatty acid.

From a conventional viscometer test, A is Newtonian, and B, C, and D exhibit shear thinning behavior. System B is more shear thinning than D, which also exhibits highly viscoelastic behavior.

Drag reduction is indicated in the plot of flow rate vs. pressure drop by the data intersecting and eventually crossing the line representing the turbulent flow of water in the 1-in. tube. From the pipe flow test, A, B, and D exhibit drag reduction characteristics in turbulent flow, and C exhibits the turbulent flow behavior of a purely viscous and shear thinning non-Newtonian liquid.

Three drag reducing scenarios are illustrated:

  • A exhibits the onset condition below which critical flow rate hydraulics are identical to that of water flowing at the same rates.

  • B is thickened by the additive (viscosifier) resulting in a region of laminar flow preceding the transition to drag reduced turbulent flow.

  • D exhibits the stress-controlled condition, which is a reversible behavior characterized by a reversion to the hydraulics of the carrier phase when a critical shear stress is exceeded.6

When a drag reducer is added to a drilling fluid, the following are expected on hydraulics in the turbulent regime:

  • At a constant pressure gradient, the treated fluid flows faster than the untreated fluid.

  • At a constant flow rate, the pressure gradient developed by the treated fluid is less than that developed by the untreated fluid. This response is opposite to drag augmentation that would be experienced if the flow regime were laminar, as any increase in viscosity from the additive would increase resistance to flow. Drag reduction is also manifested as extended laminarization, that is, a delay in transition from laminar to turbulent flow.

APPLICATIONS

Applications for drag reduction in drilling were recognized in the early 1960s. For example, the National Science Foundation and industry were jointly funding a program to develop additives that would reduce circulating system pressure losses. This study was in support of Project Mohole, a deep sea drilling venture involving a 35,000-ft drillstring. During the same period, flow tests at Socony Mobil Oil Co.'s field research laboratory demonstrated significant drag reduction in tubing/annulus sizes encountered with mining rigs and similar micro-slimhole drilling operations.

In 1962, the addition of high molecular weight polymers to low-solids muds significantly improved turbulent hydraulics and reduced costs in drilling operations in Wood County, Tex. In 1982, the dynamic injection of a drag-reducing calcium chloride brine contributed to the successful kill of ARCO's CO2 Well 4-15-H in the Sheep Mountain Unit in Colorado.8 In 1986, a rig hydraulics program differentiated between the turbulent friction factors of waterbased and polymer-based muds in the calculation of downhole-motor pressure losses.9

Drag-reducing mud systems have been developed recently with improved hydraulics with reduced circulating system pressure losses.

Recently, muds with drag-reducing additives have been used to reduce pressure drop in downhole motors. Also, on a recent horizontal well in Ector County, Tex., a mud formulated with 1 lb/bbl Barazan (drag reducer) and 2% KCI provided hole cleaning and developed sufficient drag reduction activity (turbulence suppression) to reduce the tubing pressure loss by 1,000 psi. The drag reducer relieved the pressure increase at the mud motor and the rate of penetration increased.

Pressure loss reductions of 20-30% are possible under the turbulent conditions experienced in conventional drill pipe sizes at typical circulating rates. Greater reductions are possible at the high levels of turbulence experienced in smaller diameter drill pipe and coiled tubing. When turbulent conditions are experienced in the drilling of small-annulus wells, there is the potential for reducing the equivalent circulating density (ECD) when turbulent conditions occur in the annulus.

DRAG REDUCTION

The following steps can be used to quantify drag reduction directly:

  • Calculate drag ratio

  • Calculate drag reduction activity

  • Test for diameter effect

  • Scale-up with friction velocity.

DRAG RATIO

The first step in this direct method is to calculate the drag ratio, Jp/Jo. This parameter is used to evaluate turbulent drag-reducing pipe flow when the same flow rate and pipe size are used. jp is the pressure gradient for a fluid treated with a drag reducer, and Jo is the pressure gradient for the untreated fluid at the same flow rate and in a tube of the same diameter.

Fig. 2 (22283 bytes) shows the results for Systems A through D. These drag ratios were computed from the data in Fig. 1 (20915 bytes) The results can be separated into two parts:

  • Drag ratios above Jp/Jo 1 correspond to nondrag reducing conditions

  • Drag ratios below Jp/jo

Because the drag ratio compares pressure drops before and after treatment of the fluid at the same flow rate, the initial effect with an additive that also viscosifies is usually a marked increase in the region of unfavorable hydraulics, corresponding to jp/jo 1. In this regard, a feature shared by all shear thinning systems is a tendency towards higher drag ratios at low flow rates where laminar conditions prevail.

Trends in drag ratios typically reflect differences in rheology or flow conditions. The trends in Fig. 2 (22283 bytes) indicate the following:

  • Turbulent flow with transition to Jp/jo

  • Laminar flow is well developed from high viscosity and precedes transition to fully developed flow and jp/jo

  • Turbulent flow signature typical of a purely viscous or inelastic-type or nondrag reducing non-Newtonian liquid is indicated when all drag ratios exhibit Jp /Jo 1 (System C).

For System D, the abrupt drop in drag reduction activity followed by an eventual reversion to non-drag-reducing behavior reflects the reversible stress-controlled effect characteristic of micellar systems.

This loss of activity is 100% recoverable, reappearing when the flow rate is throttled back such that the wall shear stress is reduced to a value less than the critical stress level for the system. A similar, but irreversible, signature is produced by a polymeric drag-reducing system experiencing shear degradation.

DRAG REDUCTION ACTIVITY

Systems A, B, and D illustrate the next step, which is to convert the drag ratios to drag reduction activity using Equation 1. (13518 bytes)

Fig. 3 (22527 bytes) is a graph of the drag reduction activity vs. flow rate. System D exhibits the highest level of activity between 5 and 100 gpm and reaches a maximum of about 78% between 60 and 80 gpm. Above 100 gpm, System D gradually becomes ineffective, whereas Systems A and B exhibit maximum activities of about 50%.

The maximum activity realized in a given process fluid is determined by flow conditions, concentration, and certain qualities and structural features of the additive, including sensitivity to mechanical shear degradation. An additive-independent limitation to activity, the dashed line labeled limiting asymptote, shows an upper limit to drag reduction activity of about 85%. Typically, maximum activity is about 80%. In Fig. 3 (22527 bytes), the dotted line labeled 100% laminar flow corresponds to the situation in which an additive would convert the flow regime to a fully laminar regime.

This correlation with flow rate is commonly used to illustrate drag reduction activity. Its principal advantage is that by inspection one can easily identify the flow rate range over which optimum activity occurs in the tube diameter used in the hydraulics tests.

DIAMETER EFFECT

These examples demonstrate that drag reduction activity can be quantified, and additive X from one company can be compared to additive Z from another in a few steps.

Can the correlations illustrated in Fig. 3 (22527 bytes) be applied to predict drag reduction activity in a pipe of a different diameter? Fig. 4 (20171 bytes) shows that the answer to this question is no.

When the same procedure is applied to another set of turbulent flow data for System B but from tests in a tube of different diameter (1.3-in. tube), the correlation is different (Fig. 4). (20171 bytes) Activity vs. flow rate cannot be used in a scale-up because two different correlations of activity result for the same composition.

This complication is the diameter effect. Although trends in the correlations are similar, pipe size has a significant affect on activity. One additional step is therefore required in the direct method of calculating drag reduction.

FRICTION VELOCITY

The last step is a fine tuning of the direct method to minimize the diameter effect when the intent is to incorporate drag reduction activity into a hydraulics scale-up involving other pipe diameters. The remedy is to express activity as a function of the friction velocity for the untreated fluid. 10

Friction velocity is defined by Equation 2. (13518 bytes) The term in parentheses is the shear stress at the pipe wall, MW is fluid density, and the subscript o indicates the shear stress is defined for the additive-free or untreated process fluid.

The dimensional conversion factor gc (lbf/lbm) X fpS2 indicates shear stress and density are expressed here in units of lbf/ft2 and lbm/ft3, respectively.

Applying this expression to both data sets and graphing drag reduction activity vs. friction velocity yields the results illustrated in Fig. 5. (21295 bytes)

Replacing flow rate with the friction velocity for the untreated fluid collapses the two diameter-dependent correlations (Fig. 4) (20171 bytes) to a single function that more closely reflects the performance of System B.

The direct method of drag reduction scale-up is useful when the hydraulics data are collected under test conditions approximating the actual application. Inspection of Fig. 3 (22527 bytes) shows drag-reducing conditions in the 1-in. and 1.3-in. tubes correspond to Uo* values in the range of 0.4-1.7 fps. It is easy to confirm that these levels are in reasonable agreement with Uo* values typically generated in drilling operations.

Note that Uo* was selected because it is easy to determine. Once pipe diameter, flow rate, and pipe roughness are defined, this parameter is calculated from standard correlations found in mud handbooks. Values for Uo* can also be determined directly from measurements of flow rate vs. pressure drop on the untreated system.

DIRECT METHOD

The following sequence of steps can be used to develop a data base of turbulent drag reducing flow data, to predict drag reducing activity, or to rank additive performance:

  1. Select a minimum of two pipe diameters, and measure pressure gradients over a range of flow rates in the turbulent regime for the additive-free fluid (water, brine, mud). Flow rates should span the anticipated range of friction velocities which can vary significantly depending on the type of pipe (coiled tubing, tubing, or drill pipe). For each test section diameter, calculate the friction velocities corresponding to each flow rate.

  2. Repeat for the treated fluid. Measure pressure gradients over the same range of flow rates used in each diameter.

  3. For each tube diameter, create a data base containing the following information at each flow rate: friction velocity for the additive-free fluid, friction velocity for the treated fluid, and drag reduction activity, together with the derived master correlation of drag reduction activity vs. friction velocity for the additive-free fluid.

  4. Select the pipe diameter and range of flow rates for the application, apply the applicable resistance law, calculate pressure gradients predicted for the untreated fluid at each flow rate, and convert each pressure gradient to friction velocity.

  5. Using the master correlation, find the drag reduction activity corresponding to each friction velocity calculated in Step 4, and convert these results to pressure gradients for the treated fluid.

It has been demonstrated these correlations apply to data from pipe and annular flow hydraulics tests.

MECHANISMS

At some concentration, the Theological response of a solution containing a drag reducer displays one or more of these indicators of viscoelasticity:

  • Recoil of suspended air bubbles when an oscillatory shear is applied

  • Surface rigidity or "ringing" when the container is lightly tapped

  • Pulling away or elongation of a liquid thread when a stirring rod is rapidly withdrawn

  • Flow radially against the action of centrifugal forces and upwards against gravity forces as the solution climbs a rotating impeller

  • Swelling when expelled from a nozzle, orifice, or capillary tube

  • Significantly slower flow under a fixed head around a single column of spheres or through a finely woven screen, even though additive is present at a few parts per million and viscosities of solution and untreated liquid are identical.

For certain polymers, 1 ppm will reduce turbulent friction losses 50%, and the solution will appear to be a simple Newtonian liquid. More typically, significant shear thinning is evident at the concentration where this level of drag reduction activity is displayed.

TURBULENCE

A Newtonian liquid in turbulent flow without a drag reducing additive can be described as follows:

  • Gross fluctuations in the velocity and frequency of turbulent motion are accompanied by deformation of "lumps" of fluid by flexing, stretching, and shrinking.

  • A thin laminar boundary or sublayer surrounds the turbulent core or mean flow.

  • Turbulent nuclei form in the wall region, and these vortices transfer energy at a high rate to the mean flow.

A dilute or moderately concentrated low-solids-content liquid in turbulent flow with drag reducing additive present can be described as follows:

  • The relaxation times of the bonds holding the liquid together increase significantly.

  • Recovery of a fluid "lump" from the deformations characterizing turbulent motion is delayed.

  • The laminar boundary layer thickens.

  • Production of dissipative vortices near the wall region is decreased.

The first three effects are characteristic of fully developed turbulent flow, but energy spectra measurements indicate some turbulence intensities increase under drag-reducing conditions.

During the past 45 years, experimental and theoretical research have developed the effective slip model, the vis-coelastic model, the elastic sublayer mean flow model, the modified thickened laminar sublayer model, the length-scale hypothesis, and the time-scale hypothesis.'

None of these models approximates a universal mechanism applicable to all drag reducers in the presence of dilute solutions and gels and with solids present in thin suspensions and slurries flowing under turbulent conditions inside a pipe. The models, however, do contain elements supporting evidence of a direct association between drag reduction and elasticity. At some concentration, any drag reducer is manifestly viscoelastic (that is, components of viscous and elastic properties are displayed), and this combination of Theological properties produces a less dissipative turbulent flow.

Structure, or morphology, is important for both polymer-type and micellartype drag reducers. 11-12 For example, the onset stress behavior of polymer System A correlates with intrinsic viscosity, a measure of size and conformation of polymer molecules in solution, and concentration.

TEST METHODS

Indirect methods of identifying drag-reducing materials are the indicators of elasticity, mentioned above. Direct methods involve the measurement of pressure losses under turbulent flow conditions, over the same range of flow rates with and without a drag reducer and using the identical pipe. Optimum control of flow conditions mandates either a closed-loop recirculating pipe flow or a single-pass or blowdown pipe flow.

Absolutely smooth flow surfaces are not required but a rough wall may aggravate premature degradation of shear-sensitive materials. In this regard, the single-pass configuration tends to prolong the performance of shear-sensitive additives.

If the objective is to identify the most promising drag reducer out of a group, measurements at one or more flow rates in one tube diameter can provide a simple index of performance in terms of the drag ratio parameter. If the objective also includes predicting drag reduction activity over a range of flow conditions and pipe sizes, a minimum of two pipe diameters is required.

High-shear turbulent flow in a capillary can also help identify and rank additive performance provided viscosities are low, solids are eliminated, and shear degradation does not occur. Fully developed turbulent flow in a capillary, however, produces high friction velocities. In other words, a drag reduction measured from the capillary test may be overly optimistic. The friction velocities should be typical of flow conditions with drill pipe and average circulation rates.

DRAG REDUCING MATERIALS

Among the synthetic polymers, spectacular levels of drag reduction have been recorded with high molecular weight grades of polyethylene oxide at concentrations of 10 ppm or less. Aqueous solutions of high molecular weight polyacrylamide, its partially hydrolyzed derivatives, and acrylamide-acrylate copolymers are characterized by significant levels of drag reduction activity.

A variety of natural gums, including guar, locust bean, tragacanth, karaya, and okra, exhibit drag reduction activity in freshwater and brine.

Synthetic polymers, natural gums, and derivatives of these materials impart drag reduction activity to hydraulic fracturing (frac) fluids. Anionic polyacrylamides are used in waterbased frac systems, and cationic polyacrylamides are used in frac systems formulated with acids and brines. Guar gum and hydroxypropyl guar are used in cross-linked water-based frac fluids.

Microorganisms (bacteria, common freshwater and marine algae, and molds) synthesize products with drag-reducing qualities. For example, Baroid Drilling Fluid's Barazan mud additive is based on the xanthan product synthesized by the bacterium Xanthamonas campestris.

MICELLAR SYSTEMS

Derivatives of certain low molecular weight soaps or surfactants form aggregates of micellar structures with viscoelastic and drag reducing characteristics similar to those produced in concentrated solutions of high molecular weight polymers.

Examples include the derivatives generic to micellar System D, in which drag reduction activity is restricted to carbon chain lengths in the C12-C18 range.

HYDROCARBON-BASED FORMULATIONS

The earliest hydrocarbon-based formulations involved gasoline gelled by the metallic salt of a fatty acid (Napalm) and monochlorobenzene doped with polymethylmethacrylate. 1-2

Among the synthetic polymers, the higher molecular weight grades of polystyrene, polyisobutylene, and polydimethylsiloxane impart drag reduction activity to a variety of hydrocarbon solvents and petroleum products. Low molecular weight nonpolar materials (for example, aluminum disoaps) show drag-reduction activity in certain hydrocarbon liquids. A variety of proprietary drag-reducer formulations are marketed as frac additives and pipeline flow improvers.

PERFORMANCE

Factors critical to the level of drag reduction activity realized with a particular combination of additive and fluid or suspension include additive molecular characteristics, degradation, adsorption, and solute/solvent compatibility." 12

Additive molecular characteristics include molecular weight, molecular rigidity, and entanglement capacity. Typically, the rigidity of a polymer comprised of linear and flexible chains will be lower than that of a polymer comprised of a rigid backbone with stiff side chains and pendant groups.

The data base compiled by Liaw, et al., on biopolymers, natural gums, and cellulose derivatives suggests that the correlation of drag reduction activity with entanglement capacity is better than a correlation with chain rigidity."

For specified molecular weight and concentration, a synthetic polymer molecule comprised of fairly flexible chains and few side groups should yield more drag reduction activity in deionized water than a molecule comprised of a rigid backbone with steric hindrance from bulky side chains or pendant groups.

Compared to partially hydrolyzed polyacrylamides (PHPA), materials like xanthan and other biopolymers, gums, and cellulose derivatives are less susceptible to the shear induced by mud mixers and the drill bit but are probably more prone to thermal degradation and fermentation.

Note that these materials are also viscosifiers, and the concentrations required to produce drag reduction activity comparable to PHPA may cause excessive apparent viscosities, pressure surges, and solids removal problems.

Unlike solids-free systems, depletion of drag additives from adsorption on the high surface area solids suspended in a mud system is important. Shale stabilization involves adsorption of polymer on drill solids.

By contrast, because a drag reducer may also be adsorbed on the solid phase, an excess must be present in the free stream or liquid phase in an amount sufficient to yield the target level of drag reduction activity.

Compatibility with the chemical environment of the liquid phase is a factor in screening drag reducers. For example, divalent ions have a more adverse effect on the rheology of a polyacrylamide solution than on a solution of xanthan or hydroxyethylcellulose.

REFERENCES

  1. Mysels, K.J., Drag Reduction, Savins, f.G., and Virk, P.S., editors, Chemical Engineering Progress Symposium Series, Vol. 67, No. 11, 1971, pp. 45-49.

  2. Toms, B.A., "On the early experiments on drag reduction by polymers," in Structure of Turbulence and Drag Reduction, The Physics of Fluids, Vol. 20, No. 10, Part 11, October 1977, pp. S3-S5.

  3. Savins, J.G., "Drag Reduction Characteristics of Solutions of Macromolecules in Turbulent Pipe Flow," Society of Petroleum Engineers Journal, Vol. 1, September 1964, pp. 203-214.

  4. Crawford, H.R., and Pruitt, G.T., paper presented at technical session No. 40, Symposium an Non-Newtonian Fluid Mechanics, 56th annual meeting, American Institute of Chemical Engineers, Houston, Dec. 5, 1963.

  5. Hoyt, J.W., Drag Reduction, Invited Lecture 1, presented at Third International Conference on Drag Reduction, University of Bristol, July 2-5, 1984.

  6. Savins, I.G., "A Stress-Controlled Drag Reduction Phenomenon," Rheologica Acta, Vol. 6, No. 4,1967, pp. 323-330.

  7. Lummus, J.L., Anderson, D.B., and Fox, J.E. Jr., article in World Oil, February 1962, pp. 68-73.

  8. Lynch, R.D., McBride, E.J., Perkins, T.K., and Wiley, M.E., "Dynamic Kill of an Uncontrolled CO2 Well," journal of Petroleum Technology, July 1985, pp. 1267-1275.

  9. Matthews, J.C., and Matthews, W.R., "Master rig hydraulics program adjusts for downhole temperatures, motor use," OGJ, Feb. 17,1986, pp. 62-66.

  10. Savins, J.G., and SeVer, F.A., "Drag reduction scale-up criteria," in Structure of Turbulence and Drag Reduction, The Physics of Fluids, Vol. 20, No. 10, Part 11, October 1977, pp. S78-S84.

  11. Liaw, Gin-Chain, Zakin, J.L., and Patterson, G.K., "Effects of Molecular Characteristics on Drag Reduction," American Institute of Chemical Engineers Journal, Vol. 17, No. 2, pp. 391-397.

  12. Hand, J.H., and Williams, M.C., Drag Reduction, Savins, J.G., and Virk, P.S., editors, Chemical Engineering Progress Symposium Series, Vol. 67, No. 11, 1971, pp. 6-9.

THE AUTHOR

J.G. Savins is a consultant in applications for rheology and rheologically complex materials/fluid mechanics. He is currently a consultant scientist in the research and engineering department of Baroid Drilling Fluids.

Savins is retired from Mobil research in Dallas where he worked on projects relating to viscometer design, drilling and fracturing fluids, cementing, drag reduction, underground coal gasification, uranium and enhanced oil recovery, algae biotechnology, pipeline transport, and gas-liquid separators.

Savins graduated from Texas A&M University with a BS in chemistry. He has written 36 technical articles and has 30 U.S. patents. He served as assistant editor of the Journal of Rheology and is a member of the Society of Rheology, British Society of Rheology, Society of Petroleum Engineers, and the American Institute of Physics.

Copyright 1995 Oil & Gas Journal. All Rights Reserved.