Reservoir Formation Damage Episode 3 Part 6 pptx

Chapter 18Drilling Mud Filtrate and Solids Invasion and Mudcake Formation Summary Near wellbore mud filtrate and fines invasion during drilling operationsand the resulting formation dama

Part VI

Formation Damage Models for Field Applications

Fluids and Solids

Invasion, Sand

Production, and

Scale Formation

Chapter 18

Drilling Mud Filtrate and Solids Invasion and Mudcake

Formation

Summary

Near wellbore mud filtrate and fines invasion during drilling operationsand the resulting formation damage and filtercake formation are amongstthe most important problems involving the petroleum reservoir exploita-tion This chapter reviews the fundamental processes and their mathe-matical formulation necessary to develop models that can be used forassessment of the damaged zone, filtrate and fines concentrations, fluidsaturations, and the filtercake thickness and permeability alteration duringdrilling The effects of under and over balance drilling on near wellboreformation damage are discussed The models for simulation of the singleand two-phase flow situations in the formation with water or oil baseddrilling mud cases are described External particle invasion prior tofiltercake buildup and its effect on the formation damage by particleinvasion and retention and filtercake formation are described Thesemodels are demonstrated by various applications The models presentedhere can be used for accurate estimation of the near wellbore fluidsaturations and resistivity profiles, which are necessary for accurate well-log interpretation

Introduction

As illustrated in Figure 18-1 by Yao and Holditch (1993), drilling ofwells into subsurface reservoirs is usually accompanied with mud circula-tion in order to remove the frictional heat generated as the drill bitpenetrates the rock, to provide a lubrication for reduction of the frictional

608

UnlnvicUdZom Sandstone

Shale

Figure 18-1 Mud filtrate invasion in the near-wellbore formation (after Yao

and Holditch, ©1993 SPE; reprinted by permission of the Society of PetroleumEngineers)

effects, and to transport the cuttings of the rock produced during drilling.However, mud fines and filtrates can invade and damage the near wellboreformation as depicted in Figure 18-1 Typical drilling muds may be water-based, oil-based", or water-oil emulsion types Usually, certain types offine solid particles are added as weighting agents Drilling muds areusually non-Newtonian fluids (Briscoe et al., 1994) As shown in Figure18-2 using the data by Simpson (1974), the depth of filtrate invasionstrongly depends on the type of muds As can be seen, the depth ofinvasion is less with oil-based muds, more with water-based muds, and

in between with emulsion muds applied to a water-wet formation.Drilling of wells may be accomplished by overbalanced or under-balanced drilling techniques As explained by Bennion et al (1995), bothtechniques have certain advantages and disadvantages In the overbalanceddrilling, the downhole pressure of the circulating mud is maintained abovethe reservoir fluid pressure to prevent the reservoir fluids entering intothe wellbore Bennion et al (1995) state that overbalanced drilling is morecommon because the downhole pressures of the conventional muds areusually higher than typical reservoir fluid pressures Consequently, the

Time 1 ", day' Figure 18-2 Depth of filtrate invasion data of Simpson (1974) plotted against

the square root of time for different muds

overbalance pressure forces the mud filtrate and solids to invade anddamage the near wellbore formation and eventually form a protectivesealing filtercake over the formation face This problem can be alleviated

by underbalanced drilling Bennion et al (1995) state that underbalanceddrilling can be accomplished naturally using unweighted drilling muds

in geostatically overpressured reservoirs or using oil-based muds, whichare lighter than the water-based muds or foamed muds As a result, forcedinvasion of mud filtrate and fines into the near wellbore formation isprevented Bennion et al (1995) explains that underbalanced drilling isparticularly advantageous for high permeability, naturally fractured, andheterogeneous formations and for clayey formations that are sensitive tochemicals However, they explain that underbalanced drilling does notcompletely eliminate the formation damage, because underbalancedconditions cannot be maintained at all times during drilling, some drillingfluids can still enter the near wellbore formation by spontaneous imbibi-tion, and the formation face can be damaged due to insufficient lubricationand turbulence, and inefficient cooling For these reasons, the protectivesealing filtercake formed during overbalanced drilling is still beneficial

As shown in Figure 18-3 by Yao and Holditch (1993), the mud filtrateinvading the near wellbore formation mixes with and/or displaces thereservoir fluids (Civan, 1994, 1996; Phelps, 1995, Bilardo et al., 1996)

As a result, a damaged zone is created around the wellbore (Liu andCivan, 1993, 1994, 1996; Civan and Engler, 1994)

Transition/ CaketZone "

Figure 18-3 Detailed schematic of the various zones and the mud filtrate

invasion profiles at different times in the near-wellbore formation (after Yaoand Holditch, ©1993 SPE; reprinted by permission of the Society of PetroleumEngineers)

Prediction of the near wellbore conditions, such as mud filtrate andfines invasion and distribution, is important for accurate interpretation ofthe well-logs used for measurement and monitoring of the properties

of the near wellbore formations and accurate estimation of the carbon content of the reservoirs (Civan and Engler, 1994; Phelps, 1995;Ramakrishnan and Wilkinson, 1997) Civan (1994) states: "This process

hydro-is complicated by the formation of a mud filtercake and its effect oninvasion by reducing the filtrate volume and the migration of fine particlesinto the porous formation Simultaneously, the properties of the fluidphases in porous media, such as density and viscosity, vary as a result

of mixing and interactions of reservoir fluids with the mud filtrate andfine particles." Therefore, for modeling purposes, the coupling of theexternal filtercake buildup and the near wellbore fluid invasion andformation damage is essential (see Figure 18-4)

Donaldson and Chernoglazov (1987) developed a "leaky-piston" filtrateinvasion and convection-dispersion filtrate transport model applicable tocases involving drilling muds that can mix with the formation fluid Thismodel considers the dispersion of the mud filtrate within the formationfluid in a single-phase fluid system to estimate the salinity variation inthe near wellbore region, but neglects the affect of mud fines invasion.This model was formulated for linear flow and the filtercake affect is

Mud A

Well

Reservoir formation

Figure 18-4 Mud-cake buildup over the wellbore sandface and filtrate

invasion in the near-wellbore formation (after Civan, ©1999 SPE; reprinted

by permission of the Society of Petroleum Engineers)

simulated by means of an empirically determined, decaying filter rateequation Civan and Engler (1994) extended and improved this model forthe radial filtrate invasion case applicable to actual openhole wells.Yao and Holditch (1993) and Bilardo et al (1996) have developed radialfiltration models for reservoirs containing some formation water Theyassumed that the mud filtrate mixes with the formation water as a singlephase Because their interest is in the development of models to estimatethe water phase saturation, they do not consider a convection-dispersiontransport equation for estimation of the brine salinity variation due to themixing of the mud filtrate with the formation brine However, the salinitywould be required for the resistivity measurements Phelps (1995) presents

a model to determine the fluid saturations in layered formations duringmud filtrate invasion Civan (1994, 1998, 1999) presented an improvedformulation of the multi-species and two-phase fluid transport in deforming

porous media; derivation of compressible and incompressible cake modelswith and without particle invasion; and an application for radial flowfiltercake buildup and mud filtrate invasion Olarewaju (1990) developed

an analytical model for permeability alteration around wells due to drillingmud filtrate invasion and mudcake formation Ramakrishnan and Wilkinson(1997) developed a radial model for water-based mud filtrate invasion.This model enables the determination of the saturations of the oil andwater phases and the salt concentration in the water phase They combineall the dissolved ions in brine into a single pseudo-component, called

"salt." Chin (1995) presents numerical models for formation invasion forvarious applications including formation damage, measurement-while-drilling, and time lapse analysis

In the following, single- and two-phase mud filtrate invasion modelsare presented

Simplified Single Phase Mud Filtrate Invasion Model

Similar to Donaldson and Chernoglazov (1987), Civan and Engler(1994) assumed that the mud filtrate mixes with the reservoir fluid andthe salt concentration varies This model implicitly assumes a piston typeimmiscible displacement of oil similar to the formulations by Collins(1961) and Olarewaju (1990) Thus, the fluid zone can be viewed in twoparts: the water phase and oil phase zones behind and ahead of the

displacement front, located at a distance, r e (t) In this case, the front

moves with time The formulation is also applicable when the mud filtratecan mix with the reservoir fluid (i.e., of the same wetting type) Thefiltrate mixture is considered as a pseudo-component The filtrate massbalance is given by:

where a and b are empirical parameters.

The dispersion coefficient is expressed as power-law function of thevolume flux (Donaldson and Chernoglazov, 1987):

D = fu* (18-7) where / and g are empirical parameters.

Next, three dimensionless groups are defined for computational venience and scaling purposes The dimensionless concentration isdefined as:

con-The dimensionless radial distance is given by:

Because the process is mainly convection dominated, we use Eq 18-11.The porous media peclet number is expressed as:

Pe = %&- (18-12)

where «0 and D 0 are some characteristic values that are the maximum

values of u and D, determined as following.

Note that the filtration rate varies in a range of

where

<7max = 4=0 = a

according to Eq 18-6 Thus, Eq 18-5 yields:

Thus, the volume flux varies in the range of

in which the convection time scale is given by:

The dispersion coefficient varies in the range of:

(18-19)

(18-20)

In Eq 18-27, the parameters a and (3 are given by:

The exterior radius, r e (t), of the invaded region can be determined

from the following volumetric balance:

(18-33)

Civan and Engler (1994) considered the mixing of the mud filtrate withthe resident fluid within a fixed, but sufficiently long, radial distance

(r e = constant] and obtained a numerical solution of the model using the

Crank-Nicholson finite difference scheme as described in Chapter 16.Figure 16-8 shows the solution obtained using the parameter values given

in Chapter 16 Based on Figure 16-8, the depth of filtrate invasion as afunction of time is plotted in Figure 18-5 Note that Figure 18-5 resembles

to the depth of invasion curves given by Simpson (1974)

Two-Phase Wellbore Mud Invasion

and Filter Cake Formation Model

The following assumptions are made: (1) oil/water system, (2) aqueousphase density varies by the salt content, (3) radial and horizontal flow,and (4) homogeneous formation

Civan (1994) and Ramakrishnan and Wilkinson (1997) modeled theradial invasion of water-based drilling muds into near-wellbore formation

in oil reservoirs Ramakrishnan and Wilkinson (1997) neglected the effects

of the fluid compressibility, capillary pressure, gravity, salt dispersion,the porosity variation, and fine particle invasion Therefore, the simplifiedmodel obtained under these conditions can be solved conveniently by themethod of characteristics Civan's (1994) model includes all of these effects.Therefore, it is much more complicated and requires a much more compli-cated numerical solution scheme In the following, the formulation of the

Engler (1994) plotted against the square root of time for different muds.

filtrate invasion problem is presented by combining the features of theformulations given by Civan (1994) and Ramakrishnan and Wilkinson (1997).Ramakrishnan and Wilkinson (1997) consider a two-phase fluidsystem in the near wellbore formation The aqueous phase forms the

wetting phase, denoted by W, and the oleic phase forms the nonwetting phase, denoted by N Neglecting the change of volume by mixing, they

express the density of the aqueous phase as a volumetrically weightedaverage of the densities of the aqueous phase at two extreme cases,namely, the density of the saturated solution, p^,, and the density of purewater, p^ Hence,

The initial condition is given by:

P

(18-40)

where a and P are some empirical constants, and S Wc and S Nr denotethe irreducible saturations of the aqueous and oleic phases, respectively.Thus, assuming that p^ is constant at reservoir conditions and sub-stituting Eq 18-35 into Eq 18-36 yields:

1+-* (18-53)

Substituting Eq 18-49 for the total volumetric flux, and Darcy'sequations for the aqueous and oleic phases into Eq 18-48, results in thefollowing equation for the pressure of the aqueous phase:

In the field, usually the mud pressure, P mud , is maintained constant Thus,

the filtrate invasion rate varies Note that Donaldson and Chernoglazov(1987) and Civan and Engler (1994) used empirical correlations for thefiltrate invasion rate Whereas, the filtrate invasion rate can be estimated

by means of Darcy's law assuming incompressible filtercake and constantviscosity filtrate, as following:

(18-59)

Application Ramakrishnan and Wilkinson (1997) neglected the porosity

variation and the capillary and gravity effects, and defined dimensionlessdistance and time, and normalized saturated solution volume fraction andsaturation, respectively, as:

_

(18-63)

Therefore, neglecting the capillary and gravity terms in Eq 18-52 andapplying Eqs 18-51 through 57 into Eqs 18-43 and 41, respectively,yields the following aqueous phase saturation and saturated solutionconcentration equations:

35 dF =

ac FW ac

The initial conditions are:

sion rate, q(t), is an unknown function, but it can be uniquely determined

by an inverse problem approach if the resistivity of the near wellboreformation is measured as a function of the radial distance Figure 18-6shows their solutions for low, medium, and high connate water saturations.However, the filtrate invasion rate can be directly predicted by applyingCivan's (1994) formulation as following

Applying the same simplifying assumptions used in deriving Eqs

18-64 and 65 to Eq 18-54 and Eq 18-60 and P W= PN = P, yields the

following simplified pressure equation:

ax = osubject to the following initial and boundary conditions:

Figure 18-6 Saturation and concentration profiles for (a) low, (b) medium,

and (c) high connate water saturations (after Ramakrishnan and Wilkinson,

©1997; reprinted with permission from Ramakrishnan, T S., and Wilkinson,

D J., "Formation Producibility and Fractional Flow Curves from Radial

Resistivity Variation Caused by Drilling Fluid Invasion," Phys Fluids, Vol 9,

No 4, April 1997, pp 833-844, ©1997, the American Institute of Physics)