Hole cleaning 2

3. Filtration
When the hydrostatic pressure of the drilling fluid is greater than the pore pressure, drilling fluid invades the formation (spurt loss). Suspended solids attempt to flow in with the liquid fraction, but very quickly particles of the appropriate size (generally one-sixth to one-third the size of pore throats at the well bore) bridge the pores and begin to build a filter cake. In time, finer and finer particles fill the interstices left by the bridging particles and ultimately form such a tight web that only liquid (filtrate) is able to penetrate. Once this filter cake is established, the flow rate of fluid into the formation is dictated by the permeability of the cake. When mud is not being circulated, filter cake grows undisturbed (static fluid loss) and the rate of filtration after the cake is established is proportional to the square root of time. When the mud is being circulated, the filter cake grows to the point at which the shear stress exerted by the mud balances the shear strength of the filter cake (dynamic fluid loss). Under this condition, the cake has a limiting thickness and the rate of filtration after the cake is established is proportional to time. Often, spurt loss is greater under dynamic conditions. Whether static or dynamic, the particles that invaded the formation during the spurt-loss phase may or may not ultimately help to form an internal filter cake, too.
The API Fluid Loss Test (30 min, ΔP=100 psi through No. 50 Whatman filter paper, ambient temperature) is the standard static filtration test used in the industry; however, because it uses very fine mesh paper as the filter medium, all of the bridging particles are stopped at the surface of the paper and the spurt-loss phase is not simulated properly. Usually this leads to gross underestimates of the spurt loss. A better static filtration test is the PPT, or permeability plugging test, which uses a 1/4-inch-thick ceramic disk of known permeability [API 13B1/API 13B2]. Dynamic filtration, such as in the Fann 90 test, uses a core made of the same ceramic material and simulates shearing of the filter cake by the fluid in the annulus.
For a given pressure and temperature, cake thickness is related to the filtration rate and is a function of the concentration of solids, PSD, and the amount of water retained in the cake. Filtration rate decreases with increasing concentration of solids, but cake thickness increases. Permeability, on the other hand, does not change. Permeability is almost entirely dependent on the proportion and properties of the colloidal fraction (<2 mm diameter). Permeability decreases with increasing fraction of colloids and is affected strongly by particle size and shape. A broad distribution of particle sizes is important to attain low permeability. Particles that are flat (e.g., bentonite) can pack very tightly, in contrast to spherical, granular, or needle-shaped particles. On the other hand, some organic macromolecules, such as hydrolyzed starch, are highly deformable and appear to fit well in the interstices of most filter cakes. Similarly, polyelectrolytes like CMC (carboxymethyl cellulose) and PAC (polyanionic cellulose) are large enough to be trapped in the pores of filter cakes. In NAFs, colloidal control of filter cake permeability is achieved with surfactants and water, as well as organophilic clays.
Flocculation causes particles to join together to form a loose, open network. When a drilling fluid is flocculated (e.g., through the addition of salts), the filter cake that it generates at the well bore contains some of that flocculated character, and the rate of filtration increases. Conversely, thinners (deflocculants) like lignosulfonates disperse clay flocs, thereby decreasing cake permeability.
As important as it is to have a substantial colloidal fraction of solids with a broad PSD in the mud, it is equally important that it contain a substantial concentration of bridging particles with a broad PSD. Also critical is the maximum size of bridging particles. Particles about one-sixth to one-third the size of the maximum ‘‘pore throat’’ in a drilling interval suffice, but the fluid must maintain a significant concentration of those particles throughout the interval. The following can serve as a rough idea
of the required maximum bridging particle size [Glenn & Slusser]:

maximum bridging particle size

A drilling fluid containing particles of sizes ranging up to the requisite maximum should be able to effectively bridge the formation and form filter cake. Above 10 D or in fractures, larger particles are required, and most likely the amounts needed to minimize spurt loss will also increase with the size of the opening. Generally, with increasing concentration of bridging particles, bridging occurs faster and spurt loss declines. For consolidated rock with permeability in the range 100 to 1000 mD, only
1 lb/bbl of 10-mm particles is necessary to prevent mud spurt from invading farther than 1 inch into the rock. On unconsolidated sands of that same permeability, 5 to 30 times that amount may be required. Reservoir drilling fluids typically contain as much as 30 lb/bbl total of acid-soluble bridging materials (usually CaCO3), sized to provide a broad size distribution for all solids in the fluid.
For nonreservoir applications, enough particles of the required size range are usually present in most drilling fluids after cutting just a few feet of rock. However,extensive use of desanders and desilters when drilling unconsolidated sands may deplete these particles, and some bridging material may have to be added back. Likewise, when no drilling is involved (e.g., production repair jobs), bridging particles will need to be added to the fluid.
4.Rate of Penetration
With higher ROP, both rig time and cost of bits are greatly reduced, and the total drilling operation is less costly, as shown in Figure 2.11.

Effect of Solids on Drilling Rate

Drilling-fluid parameters that can affect ROP include:
. Density
. Solids content/solids control
. Filtration
. Rheological profile
. Coefficient of friction/lubricity
. Shale inhibition
Generally, any process that leaves only the desirable solids, lubricants, and rheology modifiers in a drilling fluid will enhance ROP.
The most important mud property affecting ROP is density. ROP decreases as the pressure differential (well-bore pressure minus pore pressure) across the rock face increases. Accordingly, a drilling fluid should be as light as possible while still able to maintain well-bore stability. For a given fluid density, use of weighting materials with a high SG (e.g., hematite or ilmenite instead of barite) can increase ROP,because the volume of solids required to generate that fluid density is less and high-shear-rate viscosity is lower. If well-bore stability is not a factor, a gaseous fluid or one containing entrained gas should be considered first, followed by fluids containing hollow crush-proof beads, then oil or emulsion fluids, followed by invert-emulsion muds, freshwater muds, and brines. On the other hand, special equipment is required to drill with gases or gaseous fluids, and safety risks increase with increasing pore pressure of the formation (see Chapter 19 on Underbalanced Drilling).
Solids Content
The relationship between solids content and drilling rate has been known for many years. Broadly speaking, a low concentration of solids leads to a high ROP, as shown in Figure 2.11. The effect is most pronounced with low-solids drilling fluids, such as clear brine fluids and low-solids nondispersed muds. PSD also affects ROP, indirectly. A wide distribution of particle sizes is required to achieve adequate filtration control, which is necessary for most drilling operations. However, if the concentration of noncolloidal drilled solids can be kept below 4% by weight, ROP can be maintained at a high level [Darley]. When these particles are at such a low concentration, they are not able to form an internal filter cake below the chip (cutting) between successive tooth or cutter strikes, that is, the spurt loss is high, and the pressure differential across the chip remains high (see the following section).
Colloidal solids, which fill the interstices of a filter cake formed by the noncolloidal particles, also reduce ROP, but in a different way. With increasing fraction or total concentration of colloidal solids, the external filter cake on a chip forms more quickly and is less permeable, again reducing the probability of being able to form an internal filter cake. The result is that ROP decreases as dynamic fluid loss decreases (more on this in the following section). For clear-brine polymer drilling fluids, very
high ROP is achievable only by removing essentially all of the drilled solids. If a clear brine is infused with enough fine solids to be opaque, ROP will decrease by more than 50%. Nevertheless, in hard rock formations, desanders/desilters or mud cleaners may be able to keep the brine clear, but it is unlikely in younger formations.
Both concentration and PSD of solids also affect the performance of solids-control equipment. For optimal removal of cuttings at the shakers, controlled drilling—limiting the ROP—may be necessary so as not to exceed the operating limits of the pumps and shakers.
Historically, the significant reduction in ROP observed during displacing of clear water with clay-based drilling fluid was attributed to chip hold-down pressure (CHDP) [Garnier & van Lingen]. As a tooth from a tricone bit creates a crack in a rock, a vacuum is created under the chip (cutting) unless enough liquid can rush in to fill this incipient crack. Better penetration of the fluid into this crack reduces the pressure drop holding the chip in place, thus facilitating its removal and enabling the
tooth to engage fresh rock. For a permeable formation, the fluid to fill the crack can come from within the formation; this is one of the reasons that sandstones generally drill faster than shales. A somewhat different scenario is postulated for PDC bits. Here the argument is that the differential pressure acting across the chip opposes its initial dislodgement.
In keeping with CHDP theory, several years of study indicate that ROP increases as density decreases and filtration control is relaxed, regardless of the type of bit. However, ROP does not appear to correlate with static fluid loss, such as is measured with the API Fluid Loss Test. On the other hand, ROP appears to correlate very well with dynamic fluid loss. These are tests designed to simulate downhole flow of fluid across the face of the filter cake, leading to continual erosion and production of a constant thickness cake. Thus, as dynamic fluid loss increases, so does ROP. It is essential in these tests to use core from the area to be drilled.
As might be expected on the basis of CHDP, if the rock being drilled has very low permeability (in the extreme case, shales with no microfractures), dynamic fluid loss measurements will show very low fluid loss, and consequently ROP will be lower than in permeable rock.
Fluids with low viscosity at high shear rates effectively overcome the chip hold-down effect and sweep the hole clean of drilled solids quickly, thereby minimizing regrinding. Often, though not necessarily, PV is also low; the relevant viscosity, however, is the viscosity under the bit, that is, the Fann Reading at high Fann Speed. Generally a formation drills faster the more shear-thinning and flatter the rheological profile of the mud. This again reinforces the advantages of a ‘‘clean’’ drilling fluid.
The ability to turn the drill string, log the well, and run casing in highly convoluted well bores is considered desirable. High lubricity (low coefficient of friction) of the drilling fluid, whether attained mechanically with the addition of glass or polystyrene beads or chemically with the addition of oils or surfactants, enables more accurate control of weighton-bit and drill string rotation, thereby enhancing ROP.
Shale Inhibition
‘‘Balling’’ occurs when drilled cuttings are not removed from beneath the bit and they collect between the bit and the true hole bottom (bottom balling) or in the cutters or teeth of the drill bit (bit balling). This effect is most pronounced with hydratable shales (e.g., gumbo) in WBMs. Static and dynamic CHDP conspire with the hydrational and adhesive forces to make the cuttings very soft and sticky. This effect can be reduced by making the mud either more inhibitive or less inhibitive so as to reduce the hydrational and adhesive forces. More on this may be found in the following section.
5.Shale Inhibition Potential/Wetting Characteristics
Because the continuous phase in NAFs is nonaqueous, cuttings drilled with NAFs do not hydrate, and they are left oil wet and nearly intact. Invert-emulsion NAFs can actually increase cuttings’ hardness by osmotically removing water from the cuttings. WBMs, on the other hand, generally are not very efficient at removing water from the cuttings; indeed they may only slow hydration, so that cuttings will still tend to imbibe water, swell, soften, and even disperse. The same phenomenon occurs in the well bore, so that both well-bore stability and cuttings integrity suffer with increased residence time of the mud downhole.
Highly inhibitive WBMs, such as PHPA/glycol in a 20 to 25% solution of NaCl, can remove water from the cuttings, but the cuttings may actually get stickier, depending on how wet they were when generated. The Atterberg limits give a qualitative picture of the effect of removal or imbibition of water on plasticity, or stickiness, of cuttings [Bowles], which may promote bit balling (see Fig. 2.12). Shale-laden cuttings from young formations, such as gumbo-like argillaceous formations of the Gulf of Mexico, tend to be very wet and on the downslope right side of the Atterberg curve. Exposure to highly inhibitive WBMs may remove some, but not enough, water and cause the shale to travel left back up the curve to a more sticky condition. Cuttings generated using this kind of mud tend to be stickier than those generated with a less inhibitive mud [Friedheim et al.], so that blinding of shaker screens is a common occurrence. Replacing the fluid with NAF can remedy this problem, but treatment of the WBM with a drilling enhancer (or ROP enhancer) may be more economical. Although WBM treated with a drilling enhancer presents more risk than NAF, it can reduce bitballing tendency significantly, as well as blinding of screens and other solids-control problems. Most drilling enhancers possess the added virtue of imparting additional lubricity to the fluid and reducing abrasiveness of the cuttings.

Effect of Water Content of Clays and Shales on Stickiness.6.Lubricity
A drilling-fluid coefficient of friction that is low (0.1 or less) is generally advantageous, inasmuch as it helps the cuttings to travel as discrete particles over shaker screens. Most mud lubricants will also tend to adsorb onto almost any surface, including the exposed surfaces of the solids-control equipment. A thin film or coating of mud lubricant on those surfaces can help to protect them from corrosion and mitigate adhesion of sticky solids.
To minimize corrosion of steel tubulars and solids-control devices, control of the responsible agents is a necessity. NAFs effectively prevent corrosion because they are nonconductive and oil-wet the steel surfaces. WBMs, on the other hand, can contain dissolved materials that set up electrochemical cells that ultimately lead to loss of iron from the steel surfaces in contact with the drilling fluid [Bush]. Dissolved O2 forms rust and pits on the steel surface, and is best controlled by minimizing air
entrainment: use only submerged guns in the mud pits; rig all return lines from desanders, etc., to discharge below mud level; and minimize use of the hopper. Keeping the mud at a pH between 9 and 10—with Ca(OH)2 (lime), NaOH (caustic), or MgO—helps greatly to keep the rate of corrosion at an acceptable level. A higher pH is not recommended, particularly in high-temperature wells, because under those conditions the hydroxyl ion becomes very reactive toward clays and polymers. If too
much corrosion still occurs, O2 scavengers such as sodium sulfate (Na2SO3) and triazine can be very effective. Less common but also very effective are corrosion inhibitors, such as amines and amine salts, which produce an oily barrier to O2. The other two primary agents of corrosion are carbon dioxide (CO2) and hydrogen sulfide (H2S). Both of these form acids in aqueous drilling fluids. H2S in particular is a cause for concern because of its high toxicity and its ability to cause hydrogen stress cracking that can lead to fatigue failure of tubulars and solids-control equipment. Again, a high pH can serve as the front line of defense. For high levels of H2S, though, zinc carbonate, zinc chelate, powdered iron, or magnetite may also be necessary. A mixture recommended by the API for polymer-based WBMs to minimize both corrosion and degradation of polymers by O2, CO2, and H2S consists of MgO, Na2SO3 or triazine, and triethanolamine (to sequester iron and remove H2S/CO2) [API RP 13C].
It should be noted that dissolved CO2, O2, and salts can all accelerate stress cracking and failure of steel hardware, though the effect is most pronounced with H2S.
Finally, microbes can form corrosive agents, particularly H2S, via degradation of mud components in the drilling fluid like lignosulfonate or biopolymers. The most effective ways to control microbial corrosion are through use of clean make-up water and a biocide, such as glutaraldehyde or bleach.
8.Drilling-Fluid Stability and Maintenance
Maintaining the drilling fluid in good condition is essential not only for controlling the mud properties but also to ensure proper operation of solids-control equipment. Vigilance against the effects of contamination and elevated temperatures is particularly important. Invasion of foreign materials, such as water and oils, and thermal degradation of polymers can affect viscosity and filtration properties radically and compromise the performance of some solids-control equipment. Elevated temperatures can also destroy direct and invert-emulsion systems and can cause gelation in clay-based muds, either of which can negatively affect equipment performance. Keeping the mud properties within the design parameters is critical, which requires maintaining the concentrations of mud products and drilled solids at appropriate levels.

Hole cleaning

Good solids control begins with good hole cleaning. One of the primary functions of the drilling fluid is to bring drilled cuttings to the surface in a state that enables the drilling-fluid processing equipment to remove them with ease. To achieve this end, quick and efficient removal of cuttings is essential.
In aqueous-based fluids, when drilled solids become too small to be removed by the solids-control equipment, they are recirculated downhole and dispersed further by a combination of high-pressure shear from the mud pumps, passing through the bit, and the additional exposure to the drilling fluid. The particles become so small that they must be removed via the centrifuge overflow (which discards mud, too) and/or a combination of dilution and chemical treatment. Thus, to minimize mud losses, drilled solids must be removed as early as possible. Figure 2.9 shows a decision tree that can be useful in identifying and solving hole-cleaning problems.

Hole-Cleaning Flow Chart.

Figure 2-9. Hole-Cleaning Flow Chart.

1. Detection of Hole-Cleaning Problems
Historically, the combination of the necessity to pump or backream out of the hole and a notable absence of cuttings coming over the shale shaker prior to pulling out of the hole has been a reliable indicator of poor hole cleaning. When some cuttings are observed, however, the quantity of cuttings itself does not adequately reflect hole-cleaning efficiency. The nature of those cuttings, on the other hand, provides good clues: Good cuttings transport is indicated by sharp edges on the cuttings, whereas smooth and/or small cuttings can indicate poor hole cleaning and/or poor inhibition. With the advent of PWD (pressure while drilling) tools and accurate flow modeling, a number of other indicators have come to light that foreshadow poor hole cleaning and its attendant consequences. Among these are:
. Fluctuating torque
. Tight hole
. Increasing drag on connections
. Increased ECD when initiating drill string rotation
2. Drilling Elements That Affect Hole Cleaning
Critical elements that can affect hole cleaning include the following:
. Hole angle of the interval
. Flow rate/annular velocity
. Drilling-fluid rheology
. Drilling-fluid density
. Cutting size, shape, density, and integrity
. Drill string rotational rate
. Drill string eccentricity
For a given drilling-fluid density, which is generally determined by well bore stability requirements, the hole may be classified into three hole cleaning “zones”according to hole angle:

relationship of Hole Angle and Critical Parameters

Generally, in near vertical and moderately inclined hole intervals, annular velocity (AV) has the largest impact upon whether a hole can be cleaned of cuttings. However, in extended-reach, high-angle wells (Zone III), AV places third in critical importance, though there is a critical velocity below which a cuttings bed will not form [Gavignet & Sobey].
In practice, the optimum theoretical flow rate may vary from the achievable flow rate. The achievable flow rate is restricted by surface pressure constraints, nozzle selection, use of MWD (measurement while drilling) tools, and allowable ECD. On the other hand, little is gained from very high AVs. Indeed, above 200 ft/min, little improvement in hole cleaning is usually observed, and the primary effect of increasing AV above this level is to increase ECD. In Zone III applications, lowviscosity
sweeps—so low that the flow regime in the annulus changes from laminar to turbulent—can be effective. Unfortunately, the volume of fluid required to reach critical velocity for turbulent flow is frequently outside the achievable flow rate for hole sizes larger than 8(1/2)-inch and is frequently limited by maximum allowable ECD and/or hole erosion concerns.
Another way to increase AV is to reduce the planned size of the annulus by using larger-OD drill pipe. Not only does a larger pipe generate a smaller annular gap, thereby increasing fluid velocity, it also increases the effect of pipe rotation on hole cleaning. Thus, increasing the OD of drill pipe to 6(5/8)inches with 8-inch tool joints has proven to be effective in aiding the cleaning of 81 2-inch well bores. A caveat: Although reducing the annular gap can greatly improve hole cleaning, it also makes fishing more difficult; indeed, it violates the rule of thumb that stipulates a 1-inch annular gap for washover shoes.
In a vertical hole (Zone I), laminar flow with low PV and elevated YP or low n-value and high K-value (from the Power Law model) will produce a flat viscosity profile and efficiently carry cuttings out of the hole [Walker]. Viscous sweeps and fibrous pills are effective in moving cuttings out of a vertical hole.
In a deviated hole (Zones II and III), cuttings have to travel but a few millimeters before they pile up along the low side of the hole. Consequently, not only do cuttings have to be removed from the well bore, they also have to be prevented from forming beds. Frequently a stabilized cuttings bed is not discovered until resistance is encountered while attempting to pull the drill string out of the hole. Close monitoring of pressure drops within the annulus using PWD tools can provide warning of less than optimal hole cleaning. Increased AV coupled with low PV, elevated low-shear-rate viscosity, and high drill string rpm will generally tend to minimize formation of a cuttings bed. To remove a cuttings bed once it has formed, high-density sweeps of low-viscosity fluid at both high and low shear rates, coupled with pipe rotation, are
sometimes effective in cleaning the hole. Viscous sweeps and fibrous pills tend to channel across the top of the drill pipe, which is usually assumed to be lying on the lower side of the hole.
For extended-reach drilling programs, flow loop modeling has generated several rules of thumb for low-shear-rate viscosity to avoid cuttings bed formation. The most popular is the rule that for vertical holes the 6-rpm Fann Reading should be 1.5 to 2.0 times the open-hole diameter [O’Brien and Dobson]. Another rule of thumb specifies a 3-rpm or 6-rpm Fann Reading of at least 10, though 15 to 20 is preferable. However, each drilling fluid has its own rheological characteristics, and these rules of thumb do not guarantee good hole cleaning. If the well to be drilled is considered critical, hole-cleaning modeling by the drilling-fluid service company is a necessity.
NAFs generally provide excellent cuttings integrity and a low coefficient of friction. The latter allows easier rotation and, in extended-reach drilling, more flow around the bottom side of the drill string. As the drill string is rotated faster, it pulls a layer of drilling fluid with it, which in turn disturbs any cuttings on the low side and tends to move them up the hole.
Optimizing the solids-control equipment so as to keep a fluid’s drilledsolids content low tends to produce a low PV and a flat rheological profile, thereby improving the ability of the fluid to clean a hole, particularly in extended-reach wells. The fluid is more easily placed into turbulent flow and can access the bottom side of the hole under the drill pipe more easily. In the Herschel-Bulkley model, a moderate K, a low n (highly shear-thinning), and a high o are considered optimal for good hole cleaning.
Carrying Capacity
Only three drilling-fluid parameters are controllable to enhance moving drilled solids from the well bore: AV, density (mud weight [MW]), and viscosity. Examining cuttings discarded from shale shakers in vertical and near-vertical wells during a 10-year period, it was learned that sharp edges on the cuttings resulted when the product of those three parameters was about 400,000 or higher [Robinson]. AV was measured in ft/min, MW in lb/gal, and viscosity (the consistency, K, in the Power Law model) in cP.
When the product of these three parameters was around 200,000, the cuttings were well rounded, indicating grinding during the transport up the well bore. When the product of these parameters was 100,000 or less, the cuttings were small, almost grain sized.
Thus, the term carrying capacity index (CCI) was created by dividing the product of these three parameters by 400,000:
CCI =(AV)×(MW)×(K)/400,000.
To ensure good hole cleaning, CCI should be 1 or greater. This equation applies to well bores up to an angle of 35°, just below the 45° angle of repose of cuttings. The AV chosen for the calculation should be the lowest value encountered (e.g., for offshore operations, probably in the riser).
If the calculation shows that the CCI is too low for adequate cleaning, the equation can be rearranged (assuming CCI¼1) to predict the change in consistency, K, required to bring most of the cuttings to the surface:
K= 400,000/(MW)×(AV)
Since mud reports still describe the rheology of the drilling fluid in terms of the Bingham Plastic model, a method is needed to readily convert K into PV and YP. The chart given in Figure 2.10 serves well for this purpose. Generally, YP may be adjusted with appropriate additives without changing PV significantly.

Conversion of Bingham Plastic Yield Point to Power Law K

Example: A vertical well is being drilled with a 9.0 lb/gal drilling fluid circulating at an AV, with PV=15 cP and YP=5 lb/100 ft2. From Figure 2.10, K=66 cP, and from the CCI equation, CCI=0.07. Clearly, the hole is not being cleaned adequately. Cuttings discarded at the shale shaker would be very small, probably grain size. For a mud of such low density, PV appears to be much too high, very likely the result of comminution of the drilled solids. Solving the equation for the K value
needed to give CCI=1 generates K=890 cP. From Figure 2.10, YP needs to be increased to 22 lb/100 ft2 if PV remains the same (15 cP). If the drilled solids are not removed, PV will continue to increase as drilled solids are ground into smaller particles. When PV reaches 20 cP, YP will need to be raised to 26 lb/100 ft2. As PV increases and YP remains constant, K decreases. It is easier to clean the borehole (or transport solids) if PV is low. Low PV can be achieved if drilled solids are removed at the surface.
Cuttings Characteristics
The drier, firmer, and smaller the cuttings, the easier they are to remove from the hole. Small polycrystalline diamond compact (PDC) bit, small cutters on the bit generate small cuttings, which settle out more slowly than large cuttings and are more easily entrained in the annular column of drilling fluid by drill string rotation. As per Stokes’ law (see Chapter 13 on Centrifuges), large cuttings will fall out of suspension more rapidly than smaller cuttings, but in high-angle holes, even smaller cuttings may settle and form a cuttings bed. Rounded or agglomerated cuttings are indicative of an extended period of time in the hole and poor hole cleaning.
Rate of Penetration
Preventing cuttings beds in deviated wells is far easier than removing them. Controlling instantaneous ROP is one way to avoid overloading the annulus with cuttings. ROP should always be controlled so as to give the fluid enough time to remove the cuttings intact from the bottom of the hole and minimize spiking of the fluid density in the annulus. The treatment for poor hole cleaning is to reduce ROP, circulate the hole clean, and take steps to optimize hole cleaning. Additional information is provided in the next section.
Pipe Rotation
As pipe rotation rate increases, the pipe drags more fluid with it. In deviated wells, this layer of drilling fluid disrupts cuttings beds that have formed around the pipe while lying on the low side of the hole. Step changes appear to be the norm, occurring in most cases at around 85, 120, and 180 rpm. There is some evidence that above 180 rpm, turbulent flow ensues for many fluids. At these high levels, there seems to be little additional benefit to hole cleaning from increasing pipe rotation any
further; most likely this is because cuttings beds cannot form in turbulent flow. During sliding, hole cleaning is minimal and cuttings beds are likely to form. Thus, sliding should be kept to a minimum during any drilling operation. Indeed, this is one of the reasons that rotary steerable tools have become popular.
Drill String Eccentricity
In high-angle wells, the drill string does not remain stable on the bottom of the hole while rotating. The drill string tends to climb the wall of the well bore and fall back, providing additional agitation—though also additional cuttings degradation—while aiding in the removal of cuttings beds on the low side of the hole.

Properties of drilling fluids

Just as the nature of drilling-fluid solids affects the efficiency of solids control equipment, the nature of the solids also plays an integral role in the properties of drilling fluids, which in turn affect the properties of the solids and the performance of the equipment. This intricate and very complex dynamic relationship among the solids, drilling fluid, and solids-control equipment is represented in Figure 2.7. Any change made to one of these affects the other two, and those in turn affect all three,and so on. To optimize a drilling operation, it is important to understand how the solids affect bulk mud properties, particularly rheology, hole cleaning, filtration, drilling rate (rate of penetration [ROP]), along with surface properties such as shale inhibition potential, lubricity, and wetting characteristics.

Mud Processing Circle

Rheology is the study of the deformation and flow of matter. Viscosity is a measure of the resistance offered by that matter to a deforming force. Shear dominates most of the viscosity-related aspects of drilling operations. Because of that, shear viscosity (or simply, ‘‘viscosity’’) of drilling fluids is the property that is most commonly monitored and controlled. Retention of drilling fluid on cuttings is thought to be primarily a function of the viscosity of the mud and its wetting characteristics. Drilling fluids with elevated viscosity at high shear rates tend to exhibit greater retention of mud on cuttings and reduce the efficiency of high-shear devices like shale shakers [Lundie]. Conversely, elevated viscosity at low shear rates reduces the efficiency of low-shear devices like centrifuges, inasmuch as particle settling velocity and separation efficiency are inversely proportional to viscosity. Water or thinners will reduce both of these effects. Also, during procedures that generate large quantities of drilled solids (e.g., reaming), it is important to increase circulation rate and/or reduce drilling rate.
Other rheological properties can also affect how much drilling fluid is retained on cuttings and the interaction of cuttings with each other. Some drilling fluids can exhibit elasticity as well as viscosity. These viscoelastic fluids possess some solid-like qualities (elasticity), particularly at low shear rates, along with the usual liquid-like qualities (viscosity). Shear-thinning drilling fluids, such as xanthan gum–based fluids, tend to be viscoelastic and can lower efficiency of low-shear-rate devices like static separation tanks and centrifuges.
Viscoelasticity as discussed above is based on flow in shear. There is another kind of viscoelasticity, however, that is just now receiving some attention: extensional viscoelasticity. As the term implies, this property pertains to extensional or elongational flow and has been known to be important in industries in which processing involves squeezing a fluid through an orifice. This property may be important at high fluid flow rates, including flow through the drill bit and possibly in highthroughput
solids-control devices. High-molecular-weight (HMW) surface-active polymers, such as PHPA and 2-acrylamido-2-methylpropane sulfonic acid (AMPS)–acrylamide copolymers, which are used as shale encapsulators, produce high extensional viscosity. Muds with extensional viscosity—especially new muds will tend to ‘‘walk off ’’ the shakers. Addition of fine or ultra-fine solids, such as barite or bentonite, will minimize this effect.
Rheology Models
Shear viscosity is defined by the ratio of shear stress (ζ) to shear rate (γ): μ= ζ/γ
The traditional unit for viscosity is the Poise (P), or 0.1 Pa-sec (also1 dyne-sec/cm2), where Pa¼Pascal. Drilling fluids typically have viscosities that are fractions of a Poise, so that the derived unit, the centipoise (cP), is normally used, where 1 cP=0.01 P=1 mPa-sec. For Newtonian fluids, such as pure water or oil, viscosity is independent of shear rate. Thus, when the velocity of a Newtonian fluid in a pipe or annulus is increased, there is a corresponding increase in shear stress at the wall, and the effective viscosity is constant and simply called the viscosity [Barnes et al.]. Rearranging the viscosity equation gives
and plotting ζ versus γ will produce a straight line with a slope of μ that intersects the ordinate at zero. Drilling fluids are non-Newtonian, so that viscosity is not independent of shear rate. By convention, the expression used to designate the viscosity of non-Newtonian fluids is
μe= ζ/γ
where μe is called the ‘‘effective’’ viscosity, to emphasize that the shear rate at which the viscosity is measured needs to be stipulated. All commonly used drilling fluids are ‘‘shear-thinning,’’ that is, viscosity decreases with increasing shear rate. Various models are used to describe the shear-stress versus the shear-rate behavior of drilling fluids. The most popular are the Bingham Plastic, Power Law, and Herschel- Bulkley. The Bingham Plastic model is the simplest. It introduces a nonzero shear stress at zero shear rate:
μe= ζ/γ=μp+ζ0/γ
where μp is dubbed the plastic viscosity and ζ0 the yield stress, that is, the stress required to initiate flow. μp, is analogous to μ in the Newtonian equation. Thus, with increasing shear rate, ζ0/γ approaches zero and μe approaches μp. If ζ is plotted versus γ, μp is the slope and ζ0 is the ordinate intercept. The Bingham Plastic model is the standard viscosity model used throughout the industry, and it can be made to fit highshear- rate viscosity data reasonably well. μp (or its oilfield variant PV) is generally associated with the viscosity of the base fluid and the number, size, and shape of solids in the slurry, while yield stress is associated with the tendency of components to build a shear-resistant.
When fitted to high-shear-rate viscosity measurements (the usual procedure), the Bingham Plastic model overestimates the low-shear-rate viscosity of most drilling fluids. The Power Law model (also called the Ostwald de Waele model) goes to the other extreme. The Power Law model can be expressed as follows:


where K is dubbed the consistency and n the flow behavior index. The Power Law model underestimates the low-shear-rate viscosity. Indeed, in this model, the value of ζ at zero shear rate is always zero. To alleviate this problem at low shear rates, the Herschel-Bulkley model was invented. It may be thought of as a hybrid between the Power Law and Bingham Plastic models and is essentially the Power Law model with a yield stress [Cheremisinoff]:


Portraits of the three rheology models are shown in Figure 2.8. For NAFs and clay-based WBMs, the Herschel-Bulkley model works much better than the Bingham Plastic model. For polymer-based WBMs, the Power Law model appears to provide the best fit of the three models; better yet is the Dual Power Law model: one for a low-shear-rate flow regimen (annular flow) and one for a high-shear-rate flow regimen (pipe flow). More discussion is presented in the example below.
Other models have been used too, including the Meter model (also called the Carreau or Krieger-Dougherty model), which describes structured particle suspensions well, and the Casson model, which fits OBM data well. However, neither of these models has been widely adopted by the drilling-fluid community.

Drilling Fluid Rheology Models.
Measurement of Viscosity


Shear rates in a drill pipe generally encompass the range from 511 to 1022 sec^-1 (Fann Speeds of 300 to 600 rpm), whereas in the annulus, flows are usually one to two orders of magnitude lower, such as 5.1 to 170 sec^-1 (Fann Speeds of 3 to 100 rpm). To change the Fann Speed to units of sec^-1, ψ (rpm) is multiplied by 1.7; to change the Fann Reading to units of dyne/cm^2, υ (degrees) is multiplied by 5.11. YP is actually in units of degrees but is usually reported as lb/100 ft2, since the units are
nearly equivalent: 1 degree=1.067 lb/100 ft2. Not only is the Bingham Plastic model easy to apply, it is also quite useful for diagnosing drilling fluid problems. Because electrochemical effects manifest themselves at lower shear rates, YP is a good indicator of contamination by solutes that affect the electrochemical environment. By contrast, PV is a function of the base fluid viscosity and concentration of solids. Thus, PV is a good indicator of contamination by drilled solids.
Application of the Power Law and Herschel-Bulkley Models to Rotary Viscometer Data
The Power Law model may be applied to Fann Readings of viscosity using the expression
Fann Reading = K * [Fann Speed]^n:
To obtain representative values of K and n, it is best to fit all of the Fann Readings to the model. A simple statistical technique, such as least squares regression, is quite satisfactory. If a programmable computer is not available but the flow regimen of interest is clearly understood, two Fann Readings will usually suffice to estimate the values of K and n. Naturally, this will lead to weighting of K and n to the Fann Speed range covered by those two Fann Readings. To estimate K and n in the simple
Power Law model from two Fann Readings, take the logarithm of both sides and substitute the data for the respective Fann Readings and Speeds:
Log (Fann Reading) = LogK + n Log (Fann Speed):
For example, assume pipe flow and a shear rate of interest covered by the 300 and 600 rpm Fann Readings. If the Fann Readings are 50, 30, 20, and 8 at 600, 300, 100, and 6 rpm, respectively,
Log [50] = LogK + n Log [600]
Log[30] =LogK + n Log [300].
Subtracting the second equation from the first eliminates K and producesone equation for n:
n= {Log [50/30]}=Log [600/300]= 0:737.
The value of K may be determined by substituting 0.737 for n in one of the equations above:


and the viscosity at 6 rpm is
Viscosity = {(Fann Reading)[511]=(Fann Speed)[1:7]} = 521 cP.
Thus, none of the rheology models should ever be used to extrapolate outside of the range used for the data fit.
The Herschel-Bulkley model is more difficult to apply, since the equation has three unknowns and there is no simple analog solution. Calculation of K, n, and ζ0 is generally carried out with an iterative procedure, which is difficult to do without a programmable computer. An alternative, which works fairly well, is to simply assign to ζ0 the 3-rpm Fann Reading, subtract that reading from all of the Fann Readings, and fit them to the Power Law model as described above.


Selecting, arranging, and operating solids-removal equipment to optimize the drilling-fluid cleaning process require accurate information about the intrinsic nature of the cuttings (drilled solids) and solid additives.
1. Nature of Drilled Solids and Solid Additives
Particle size, density, shape, and concentration affect virtually every piece of equipment used to separate drilled solids and/or weighting material from the drilling fluid. In the theoretically perfect well, drilled solids reach the surface with the same shape and size that they had when they were created at the drill bit. In reality, cuttings are degraded by physicochemical interaction with the fluid and mechanical interaction with other cuttings, the drill string, and the well bore.
Cuttings hydrate, become soft, and disperse in aqueous fluids and even in invert-emulsion NAFs with excessively low salinity. On the other hand, cuttings may become more brittle than the formation in highwater- phase-salinity NAFs and can be mechanically degraded by the action of the rotating drill string inside the well bore, particularly in deviated, slim-hole, and extended-reach wells. Cuttings are also degraded by mechanical action. Abrasion of the cuttings by other cuttings, by the steel tubulars, and by the walls of the well bore can lead to rapid comminution of the particles. In summary, cuttings recovered at the surface are generally smaller and frequently more rounded than at their moment of creation, depending on the nature of the cuttings themselves and the drilling fluid. Accordingly, the particle size distribution (PSD) seen at the flowline can range from near-original cutting size to submicron-sized particles.
The surface properties of the drilled solids and weighting material, such as stickiness and amount of adsorbed mud, also can play major roles in the efficiency of a rig solids-separation device. Large, dense particles are the easiest to separate using shale shakers, hydrocyclones, and centrifuges, and the differences in size and density among different types of particles must be well known to design the appropriate piece(s) of equipment for the separation process. Indeed, the optimum efficiency window for each device depends on all four of these parameters: concentration, size, shape, and density. Furthermore, since removal of some— but not all—particles is desirable, characterization of each and every type of particle with respect to those variables is critical. LCM serves as a good example of this. Usually economics dictates removal of large LCM along with cuttings using scalping shakers. Sometimes, however, large concentrations of LCM are required—as much as 50 to 100 ppb— in the circulating system. In such cases, a separate scalping shaker may be installed ahead of the regular battery of shakers to remove the LCM and recycle it back into the mud system [Ali et al ].
2 Physical Properties of Solids in Drilling Fluids
Particle sizes in drilling fluids are classified as shown in Table 2.2 [M-I llc]. PSD is measured using various techniques. For particles >45 mm diameter, wet sieve analysis is simple, accurate, and fast [API RP 13C]. Alternative methods include the American Petroleum Institute (API) sand test, which provides a measure of the total amount of particles >74 mm diameter [API RP 13B1]; microscopic image analysis, whose size limit at the low end depends on the type of microscope employed; sedimentation, for particles 0.5 to 44 mm diameter [Darley & Gray]; Coulter counter, for particles 0.4 to 1200 mm diameter [API RP 13C]; and laser granulometry (also called laser light scattering, diffraction analysis, and Fraunhoffer diffraction), for particles 1 to 700 mm diameter [API RP 13C].

Classification of Particles in Drilling Fluids

With the Coulter counter, the solids are suspended in a weak electrolyte solution and drawn through a small aperture separating two electrodes, between which a voltage is applied and current flows. As each particle passes through the aperture, it displaces an equal volume of conducting fluid and the impedance between the electrodes increases in a manner that can be correlated with the particle size.
Laser granulometry is rapidly gaining popularity as the method of choice for PSD measurements. In laser granulometry, the solids are dispersed in a transparent liquid and suspended by circulation, if necessary, the slurry may be viscosified with a material like xanthan gum polymer. A beam of light is shone on a sample of the suspended solids, and the intensity versus the angle of the scattered light is analyzed to determine the PSD. Freshwater is used to disperse inert materials like barite. The drilling-fluid base fluid (saltwater, etc.) is used for all other solids (e.g., drilled solids). The sample is diluted to make it sufficiently transparent to obtain accurate readings. The instrument fits the particles to a spherical model to generate a histogram of number of particles versus particle size. For particles that do not fit a spherical model very
well, such as plates or rods, calibration with a known PSD of those particles is preferable. Laser granulometry results also depend on the step size chosen—for instance, for step sizes of 5 mm versus 10 mm, using 5 mm will generate two peaks that are each about half the size of a peak generated using 10 mm. If the step size chosen is too large, the reported PSD may miss some of the fine structure of the spectrum; on the other hand, a step size that is too small will generate excessive oscillations and the spectrum will appear to be very ‘‘noisy.’’

figure 2.6
Figure 2.6 shows typical laser granulometry PSD curves for feed, liquid effluent (overflow), and solids discharge (underflow) for a field mud processed by a centrifuge. The efficiency of the device may be calculated from these data. PSD curves for each piece of equipment allow a more detailed understanding of what the device is doing and whether the equipment is optimally configured for the fluid being processed. There are calls within the drilling industry now to make laser granulometers standard equipment on critical wells, particularly high temperature/ high pressure and extended-reach wells, where the equivalent circulating density (ECD) is likely to exceed the fracture gradient.
Adsorbed mud, as well as swelling and/or dispersion of the cuttings resulting from interaction with the mud, can affect the PSD of cuttings. Comminution (degradation) of drilled solids has a strong impact on rheology and the total amount of mud adsorbed on the solids, inasmuch as the forces between the particles and the amount of mud adsorbed on them is proportional to their surface area. Drilled solids generally become comminuted while in the well bore and mud pits, as well as during passage through solids-control devices, through abrasion and chemical interaction with the base fluid. Surface-area increase due to comminution is proportional to the decrease in particle diameter. For example, breaking up a 100-mm-diameter particle into 5-mm particles will increase the total surface area by a factor of 20. Consequently, the
amount of mud adsorbed on the solids in this case will increase roughly by a factor of 20 as a direct result of comminution. Low-shear-rate viscosity will also increase significantly with this increase in total surface area, though the relationship is not strictly linear.
Average particle density, also termed ‘‘true’’ or ‘‘intrinsic’’ density, has units of weight/volume. Specific gravity (SG) is the ratio of the density of the material in question to the density of water and is, of course, unitless. Since the density of water is close to 1 g/cm3 over a wide range of temperature and pressure, the values reported for average particle density and SG are essentially the same. Average particle density should not be confused with bulk density (as often given in the Material Safety Data Sheet), which is a measure of the density of the packaged material. The LeChatelier flask method is the standard for determination of the average particle density of barite and hematite [API 13A]. In this method, one measures the incremental change in volume accompanying the addition of 100 g of the weighting material to a precisely measured volume of kerosene. A more convenient, but less accurate, method for determining density of weighting materials is the air pycnometer
[API RP 13I]. Another convenient method, which is rapidly gaining in popularity, is the stereopycnometer [API RP 13I]. In contrast to the air pycnometer, the stereopycnometer is as accurate as the LeChatelier flask method, and it can be used to measure density of any kind of particulate, including drilled cuttings. The stereopycnometer employs Archimedes’ principle of fluid displacement (helium, in this case) and the technique of gas expansion [API RP 13I].
Particle shape, partly described by the so-called aspect ratio, is not fully quantifiable. Neither is it possible to incorporate the broad spectrum of particle shapes in drilling fluids into particle-separation mathematical models. At this time, an old simple classification scheme is still used: granule, flake, fiber [Wright].
Concentration of particles in a mud is generally measured using aretort (an automatic portable still). The volume percentage of lowgravity solids (% LGS)—clays, sand, and salt—and the volume percentage of high-gravity solids (% HGS)—weighting material—are calculated from the measured volumes of the distilled fluids and the density of the mud. The calculated % LGS serves as an indicator of the effectiveness of the solids-control equipment on the rig. Occasionally both the overflow
and underflow solids from each piece of equipment are reported. Unfortunately, inaccuracies inherent in the retort, combined with the common practice of using an average density for the LGS and an average density for the HGS, can generate considerable uncertainties in % LGS. This is particularly true for low-density fluids, where a slight error in reading the retort will generate misleading—usually high— values of % LGS. However, if the calculated % LGS is below the target
limit (typically 5%), and dilution is not considered excessive, the solidscontrol equipment is considered to be efficient. (Calculation of solidsremoval
efficiency is presented in Chapter 15 on Dilution.) It should be noted that % LGS includes any clays that are purposefully added to the drilling fluid (for viscosity and filtration control). If a fluid contains 20 lb/bbl bentonite, it already contains 2.2% LGS before it acquires any drilled cuttings; in such fluids, the target limit of % LGS may be somewhat higher than 5%.
Concentration of particles affects mud properties, particularly rheology, which in turn affect the amount of residual mud on drilled solids. For noninteracting particles, the Einstein equation describes the effect of particles on the effective viscosity, μe, fairly well:


where μ is the viscosity of the liquid medium and μ is the volume fraction of the inert solids. This effect is independent of particle size, as long as the particles are suspended in the medium. The Einstein equation represents the effect of ‘‘inert’’ particles like barite fairly well, at least until their concentration becomes so great that the particles begin interacting with each other. Most particles in drilling fluids, however, have strong surface charges and interact strongly with each other at any concentration. Since all particles are enveloped by drilling fluid, attractive forces among strongly interacting particles (e.g., clays, drilled solids) generally lead to higher internal friction, hence a higher viscosity. Repulsive forces, such as are generated in muds containing high levels of lignosulfonate or other anionic polymers, will tend to exhibit lower viscosity. Because of these attractive/repulsive forces, strongly interacting particles generate an internal ‘‘structure’’ in a fluid, which manifests
itself most clearly at low fluid velocities. Thus, in most drilling fluids, significant deviations from the Einstein equation are the norm, as is discussed in more detail in the next section.
The viscosity of a drilling fluid must be maintained within certain limits to optimize the efficiency of a drilling operation: low-shear-rate viscosity needs to be high enough to transport cuttings out of the hole efficiently and minimize barite sag, while high-shear-rate viscosity needs to be as low as possible to maintain pumpability, remove cuttings from beneath the bit, and minimize ECD of the mud. In an analogous manner, for efficient operation of solids-control devices, the concentration
of drilled solids needs to be maintained within a specified range [Amoco]. The upper end (e.g., 5%) is particularly important, but the lower end (typically higher than 0%) is also important for most devices.
Stickiness of cuttings and its effect on the performance of solidscontrol devices are only beginning to be investigated. Various properties of the mud, along with lithology of the formation being drilled, are known to affect stickiness of particles, especially cuttings [Bradford et al.]. Generally, separation efficiency of any solids-control device decreases with increasing stickiness of the cuttings. Rheology, shale inhibition potential, and lubricity of the mud all can affect the stickiness of particles, which in turn affects performance of solids-control equipment, especially shale shakers. To handle gumbo (very sticky cuttings consisting primarily of young water-sensitive shale), operators will install special gumbo removal devices ahead of the shakers. To aid in conveyance of gumbo, the shaker screens are kept wet with a fine mist and
angled horizontally or downward toward the discharge end. Gumbo cannot be transported effectively on a linear motion or balanced elliptical
motion screen that is sloped upward.