q

1, G. CARTER; MEMBER AIME. G. W, ZVANS

HALLIBURTON CO. DUNCAN, OKIA.

Abstract

Special cementing problems concerning gas storage wel!s and tubingless completion wells have indicated the need for better pipe-cement bonding. This presentation stresses some of the problems involved, and possible solutiim.r, The application of a resinsand, coating at specific points on the caring wilt give ugreater adhesion at the cement-pipe interface. llesistance of the cement-pipe bond to fluid intrusion and loading is increased. Laboratory and jield results as well as acoustic logs emphasize the improvement in bonding when a roughness factor {s introduced, Comparative bonding data both with and without mill varnish on the pipe are ako evaluated.

Introduction The two types of bonds to be considered ,ia this presentation are shear and hydraulic bond between cement and pipe. Shear bond is dellned as the bond that meclianic,ally supports pipe in, the hole, and is determined by, measuring the force required to initiate pipe movement in a cement sheath, .This force when divided by the cement-casing contact surface area, yields the shear bond in psi. Hydraulic bond blocks migration of fluids in a cemented area and is deter+ed by applying pressure at the pipe-cement interface until leakage occurs. The hydrau)ic pressure when Ieafcage appegrs at either end of the specimen is defined as the bond failure pressure irr psig. Gas bond tests were determI ined using compressed, air or nitrogen as the pressuring medhrm. Fig. 1 illua- , trates the sample configu~ation generally used for determining these bond

streng@s in the laboratory, Dimensionally, the specimens can be of any size without significant effect. Hydraulic bonding is of primary importance, particularly in isoIating zones irr a producing or injection well, as most cementing jobs provide adequate mecht+nical support (shear bond) to hold the pipe in place. Shear bond should lot be overlooked, however, in dettmul, ing WOC time prior to subsequent d$lling or ,eompletion work~ It is neeessary to consider the effect of various easing conditions and eompletiorr techniques if optimum bond is to. be obtained and maintained. Some of the primary considerations in planning a more successful cement job from the easing-bonding s~dpoint are: (1) pipe surface finish, (2) drilling fluid; (3) thermal and, stimulation stresseq and (4) casing equipment. Pipe Surface Fittfeh Hydraulid and gas bond were directIy affected by the pipe, surface finkh against which the cement was placed. Types of casing investigated wefe steel and plastic as used in field applications, Hydraulic bonds were determined o? steel pipe with the following surface finishes: new mill varnish, mill simle (chemical removal of varnish), sandblasted, used (rus~) and
FORCE

resin-sand coated. Two ~ typea of phtstic pipe used were filament wound and cen~ifugally east with both smooth and rough external surfaces. Titble 1 shows the hydraulic, gas, and shear bond strength measurements. of cement to steel and plastic pipe with varying external surface finishes. As may be seen, the rougher the external pipe firdsh, (13g. 2) the bigher the hydraulic bond using iiquids or g~. These bond strengths indicate the importance of preparing pi@ surfaces prior to placement of cement in the well. As an example, mill varnish exhibits the lowest bond strengths compared witlf other types of finishes, with, time dependency after setting being ~oted in new mill varnish pipe below 140F. Table 2 ilhrstrates the reduction in bondhg of cemerit to pipe that occurs at about two days with this type of surface finish, -Acoustic cement bond logs were run on a test well to verk!y results obtained in physical bond measurement, and correlation was obtained in that acoustic Io~ indicated a correspond~, ing reduction in bond from two to five days after cementing. It is evident tha~ this 19w physical bond strength period would be a critical time for running acoustic logs or completion practices on new mill varriish pipe, especially: where ,Ahe temperature is below 140F. ,

, ,

,,

El.

ill!?

F@ l--$hem

bond

test (left), and hydkkdk band test.topipe (right).

, .>

TABLE 2-BUNIS Sl\Ea:GTH . joint of used pipe at 560 ft, showing time Shear Hydraulic psi Days (Wnter.pds) [Nitroseo.pslg] higher amplitude, had sections of mill TYD. OF Nnbh Bend StreniNh 200 74 15 varnish still intact. Shear Hydraullc Gas i 170 63 yg Psls Psia PSI 210 72 3 i: As vvou~d be expected, plastic pipe 74 250 New (Mill ~arnlsh) -T 40 -zwith smooth exterior surfaces exhibits New [VatnlshChemlcslly ?6 .; 270 55 Removed) 104 300.400 70 CaslnS Cendltlen-new mill varnish similar hydraulic bond values to milf 123 New \Sa#asted) 300.700 150 I Cement-APl Class A cement Used 141 500-700 150 varnish casing., Thk type of casing WaIer 15.2 gal/eack New fSani:M#d-RnIn-Sand Curfng Temper@ure-100 F. with its lower collapse pressure flexes 2400 I 100-1200 400+ Pia%tlc more readily than steel pipe; thus lowparing resin-sand coated pipe to new Fllement Weund (Smooth] 79 210 er bond failure pressure was experi9? 270 mill varnish pipe. Centrlf\g.ally Cast %%[h) S1 220 enced with rough surface plastic pipe 101 310 than with the corresponding steel pipe, CementAPl Class A .ioment Drilling FhtIds WOtOr 5,2 sal/5ack Application of a resin-sand coating casing S110--2 h!. Inside 4 in. Curlnp Temperatur~O P. to the pipe exterior resulted in conVariations in hydraulic bond Curtng Time-l day siderable imfirovement in bond. Thk strengths were observed whenever pipe * type of surf;ce finish as compared to surfaces were wet with fluids other sandblasted pipe, illustrated in Fig. 2,. than water, Drilling fluids used in increases the shear bond strength This phenomenon of time dependency these tests included water-base, oilhas not, been evidenced on other types about 20 times and the hydraulic bond base and inverted oil-emulsion muds, of pipe surfaces or at higher temperaapproxirnately IWO, tinies (Table 1). Pipe surfaces coated with mud cause tures, Athigher temperatures, the rate Even more pronounced improvement / reduction of the cement-pipe bond, of cement expa,psion overcomes the was ~oted for resistance, to gas flow with oil-wet surfaces ex~ibiting the t~me dependency effect of mill varnishat the bonded interface, The resin- lowest bond on equivalent pipe tlnishes. At lower temperatures a longer sand coat provides a rougher surface es. Under these condititms rougher time is required for the cement toexfinish to which the cement can adsurfaces again provide higher bcmds here, therefoLe, faiiure occurs within pand sufficiently to overcome those than smooth surfaces. The bond of ceeffects, A, comparison between vtious ,the cement sheath (Fig. 4) after jet ment to an oil-wet surface is approxi-. pipe finishes according to an acoustic perforating rather than at the bond mately one-half that of a water-wet or bond log is shown in Fig. 3, This Iog interface. .The hydraulic bond failure dry surface (Table 3). This should be was run three and seven days after pressure using water on this type of an important consideration in any cementing. Subsequent logs showed pipe finish is about 5:1 compared to well completion program. Therefore, the n4w and used pipe sections to be new mill varnish pipe, Gas failure the practice of running a chemical completely acoustically bonded. The pressure is approximately 40:1, comwash ahead of primary cementing jobs to remove the circulatable mud rind return the casinjg to a water-wet surface condition is desirable, The ,,,, . ..... --., e use of friction-reducing additives in a cement slurry results in better mud re400. I ioo. 1200 moval, thus, providing a mor~ satis2400 RESIN-SAND factory bonding condition.: COATED TABLE I-BONDING PROPERTIES OF VARIOUS PIPE FIN ISHES
w.! niG .!: P3(F

Thermal and Stimulation Stresses

,. .141 500.7C0 1<0-250

. .. . .
Rti$lY

SANO SLAST6D

123

SO0700

150.250

,
MILL VARNIStf
.

, t 79 2oa I 0.20

,.

. -.=fig. >Grrelation ofextemal ~pefinishand bond. strengtl1s.

The direction in which pressure is applied and the length of time pressure is heId on the bonded interface are other important factors in hydraulic bonding. TWO major applications of pressure shouldbe considered: completion and stimulation pressure where maximum presstrre is inside .tfie casing; rmd production pressures where formation. pressures are maximum and tubing pressures are minimum, 13earden and Lane have pointed out that closed-in pressure after completion of a primary cement job is very detrimental to a cement-pipe bond. Additional work has been doqe on both hydraulic and shear bonding which verifie$ previous data. Fig. 5 shows diameter expadsion of umnfpported pipe with reference to interna[ #ressure. The heat of hvdration of a setting cement canflrotiuce a sirnilar~ effec~ ... .. ... to- internal uresiuiiriti of the- ctiing and cause ex~ansion ~f tlie pipe. Abi~ ity to dhsipate this heat depends upon

152s

, ,, :. . ,,
,?

,.
1

,! . .

,,
TAB(E 3-SONBINO
, PROPERTIES ON SURFACE WETTING @F PIPE

,!

i.

formations thafi: through impermeable formations such as shales, due to posShea:~d TyP@ cl: Mud sible cement dehydration. Faster bond . 141 None strensh development will also occur 97 Water Bose 66 across higher temperature formations Inverted 011 Emulslo; 63 O!l-sase than opposite ,cooler zones. The use of Cprlns lime-l day t Casing SIZO-2 III, Indde 4 In. proper cement aMitives add different Type CasfnBusad Cement-APl Class A WMeflf types of cement can minimize thk Water 5.2 Eal/sacIc variation,e: CurIns Temperature-fJO F. During stimulation through a perforation, high internal, casing pressures the thermal conductivity and heat camay cause horizontal and yertical fracpacity of the displacing fhLi&. The turing of the cement sheath, Vertical build-up of heat inside the casing will fracturing of the cement sheath normnormally begin when the cement takes ally occurs after hydraulic bond fail- its initial set, with maximum increase ure is initiated when the cement sheath in temperature occurring at a later is forced into tension by expansion of stage. After the cement has set, the the pipe and pressure of Jhe fluid betemperature will slowly return to that .Sween the pipe and cement. of the formation, causing the casing Hydraulic bond failure extension is to contract, This oxpan$io,n and contraction of the pipe places an add~a time function dependent upon cement properties, presmre applied, and tional stress on the casing and c~ment viscosity of the pressuring medium. which could result in: ~ decrease in bond strength. Additional damage to The rate at which ,bond failure progresses has been measured in both labthe borid might occur should the casoratory and well Laboratory ining be closed in at the time the tem# .. tests. . r vestlgation entalleci measuremems on erature is rising inside the casing. 10-ft lengths of hydraulic bond test Temperature variations will occur specimens. Test-..tell work consisted of throughout a cement column in a we~l cementing casing in a normid manner, due to the various types of formations with communication being established encountered. Bond strengths normW between two sets of perforations by develop more readily across porous ;. .

use of a itraddie packer. These investi~aticms showed ~he bond failure rate ~fi~ water varied between 1.125

.> J-.!/ !., .. .>-. . . --LLLU .,..., :.-I I I I Ill I I ~I I 1.. . ~~

, METAL CENTRALIZER 540

i= 1
410

~~

F H--J
II
RUBBER USED

420

RESIN ?SANC COATED

430

440

450

. .1 IM
..,

SANDBLASTED

460

4?0
480 490

NEW PIPE SPACER

500

b+tllTlllll/

510

SANDBLASTED

520

530

t--t-u
I

Nllllllllllll.

PIPE

550

3~

1-

1111

560

t-t+ 1 1

II
, /

570

5s n

..-

lt-=:sAN:s:_

L$$Q
, J

.:

,; __._-,..

.. .

.s... .. .,.
7 D4YS

DAY s

,,,

~4@@-%%

Fig.

3Acotistic bortd logs.

w.

&Failure

within S.
159. -. 1

FEBRUARY,

1~64

.. --. . . ..> . .....

J
. .,

. .. ..

-,. .<

,.,

,:

..-,,,

..

..

,,

/ and L25 ft/rnin. Laboratory tests indicate tkt even with stronger backups to simulate denser formations, the rate is approximately the adme. Normally, gas pressuring causes faster bond failure -progression (linear rate) than water, oil or mud. For instance,, gas pressuring teats indicate that bond failure progression is in excess of 20 ft/rrrin. Prewtsre requirements with gas, 6n similar type surfaces, are considerably lower than for water, bsing approximately 1/20 as great, ( Inward pipe deflection has been calculated and then measured at 0.000015 in., at the time hydraulic bond failure <occurred on used pipe. This will ~vary with the type of surface finish and pressuring medium. The above value was measured using water, while more viscous fluids should result in higher deflectiori, and gases would g~ye Iower readings. Vertical bond failure wiii normally occur 30 either side of the pressure application point when a uniform cement sheath is. in place around the pipe. Unequal distribution of cement may cause bond failure to 6ccur at the weakest plane which could account for communication in multiple string tubingless completions.

!.

.. f

-..,,, ,

-1

Casing Rquipment

Tests have been conducted on various types of casing attachments, such as collar$, centralizers and wall cleaners, to evaluate their effect on ilydraulic and shear bonding.. Results indicate that casing intrusions into the cement column have little influence on hydrauiic or gas bond failure pressure,, but the rate of failure, progression is apparently decreased, There is the possibility that: in some instances casing attachments cani be responsible for a change jn direction of the hydraulic bond failure path. Attachments might direct bond failure from casing-cement interface to the formation-cement interface, which m~ fail at lower pressures as pointed out in previous work.] However, a large increase in shear bonding strength is obtained because of the! necessity to shear a portion of the cement itself rather than the pipecement interface, Ststnntary ~ Hydraulic and sh&r bond increases with surface roughness. Viscosity of the pressuring fluid will increase bond failure smessure as viscosity increasea. Oil-wet pipe surfaces reduce hya

draulic atid shear bond strength of cement to pipe, A change in casing internal pressure and temperature will cause a corresponding change in hydraulic and shear bond qtrengQs. ( Mill varnish has a detrimental effect on cement bond strengths. Hydraulic bond failure-is a function primarily of pipe expansion or contraction, Caiktg attachtptmts increase shear bond strength but have BO significant effect on hydraulic bond ,faihtre.. (hschrsion In plannin~ primary cementing jobs more consideration should be given tsx (1) casing surface finishes; (2) .cementitig placement techniques; and (3) ~ timing of operations in the cemented casing. Acknowledgment The authors wish to express their appr~ciation to the management, of Halhburton Co. for permission to prepare and publish thw paper, to those in this organization who offered sug-, gestions for the preparation of this paper; and especially to Glen Cordell, who helped obtain much of thedata. ,.

?.

, ;

, I

J
) ,.
J

~EXPANSION

OF PIPE

DIAMETER!

L . . . . L Evans, G. W. and Carter,L. G.: BOnd-References

ing Studies of Cementing Composition? to Pipe and Formations, Drill. and Prod. Prac., API (196!/) 72.

Vs
IiNTERNAL .
NOTE: CURVES WELO VALIO POINT TO 1

!i
,,.
, r

PRESSURE

Ei

2, Bearden, W, G. and Lane, R. D.: Engisteered Cementing Operations to. Eliminate WOC Time, Drill. and Prod. Prac., API (1%1). 3. Becker, H. and Peterson, G,: Bond. of $%%re~t Compositions for Cmnentmg , Proc., Sixth World Petrolerim Coneress. Frankfurt. Germanv (June 194. Jones, P. H. and Berdine, D.: Factors Influencing Bond Between ~ement and Formation, [Yrill. rmd Prod: Pr&., API (1940) 46. 5. Coppinger, J. E: and Goode, J. N.: Ire. pro~ed - Primary Cementing Practices Through Hydt-dic WeII-Bore Analysis, Spring Meeting of the Smithweatern D]strict, Div. of Produotiptr, API (March 1315, 1963) . ,; : i

6. SlagIe, K. A. and Smith, D, K,: Salt Cement for ShaIe and Bentoni tic Sands. Jour. Pet. Tech, (Feb., 1963) 187, . 7. Walker, T.: Report of Bond Lo@tg;, Spring Meeting of the Mid-Continent District, Div. of Production, API (April, +6,.

: :-. ~, r

I
. _._..
Fig. &Diamet&

;.

-. expansion of stusuppor-ted pipe with reference to internal pressure , . . ,.

160

Eb31To&NoT&.P1c2XUw MD BltJ- ~ ..! ; 6-RAPtitCAL SKETCHES OF L. .G. CARTER , I AND ~G. W, EVAN; APPEAR ON PACiE 168. . ., , JOURNAL OF PETROLEUM TECHNOLOGY ,. .,. 1 - ,, /:,

-i

,.

-,<.

-.,

i,

i / ,.
,
1

,..

.2

,/

.,

,,

Carter and Evans Author Bonding Study

,. : ) /

, , ,,

(right) and G. W. co-authors of ,the Technics] Article, A Study of Cement-Pip$ Bonding w~lch begins on , page 157. Carter received his BS degree in chemistry and mathematics from Southwestern State College in Durant, Okla,, in 1954. He held a teaching position before joining Halliburton Co. At present he is senio} chemist in the Cement Section of the Research & Development Lab working +,withoil- well cementing materials. Evans was awarded a BS degree in mechanical engineering by Okla%orna State U. in 1950. Employed, by Halliburton following graduation,. Evans is a development engineerf in the Mechanical. Re-, search & Development Dept., working on tools. services and equipment.

L. G.: Carter Evans (Ieft) are

. ,, , / ,,. ,

,{
k

.> f
,

,.. .. . /
,

.. . ..,:..,. -

.... . ..

.- .:. . .

-:--

. . . .. ..

,. .,