[email protected]

The Centrifuge and Mud Technology~ RAY WILLIAMS*AND JOHNMESAROS~ ABSTRACT Maiiagenieiit approves because well costs a r e reduced. Drilling engineers find t h a t penetration rates a r e increased aiid hole troubles reduced. Drilling crews a r e relieved from the work and hazard of adding large amounts of barite and massive chemical treatments t o the mud.

The decanter-type centrifuge has now operated continuously on weighted Gulf Coast muds since 1953. I t h a s become a valuable and versatile tool i11 the hands of the competent field-mud engineer. With it he can salvage barite and reject harmful light native solids. This makes it possible to reduce additions of water and chemicals, to eliminate contaminants, aiid t o produce a n effective packer mild. INTRODUCTION The centrifugal treatment of yeighted drilllng muds i s l ~ e c o ~ n i nincreasingly g prominent i n today's drilling operations. I t is f a s t becoming recognized a s the key tool which allows the mud engineer to prepare and maintain the low-solids muds needed t o drill n ~ o d e r n ,deep, high-l~ressurewells wit11 some semblance of economy. I t s efficiency a s a sollcls remover h a s been known f o r some time. No\v, its versatility a s a mud-engineering tool is I~eingexploited. The greatest value of centrifuging a s a mud-engineering tool is based on the problems of maintaining weighted muds. Even when no difficulties a r e being exper~enced,the normal maintenance of high-density mud 1s very costly. Watering back such a mud h a s been a n accepted procedure because the fine formation particles which a r e not removed by the shale-shaker screen or by settling in the tanks become a p a r t of the mud. The presence of these fine, lightweight formation particles affects the mud adversely in t h a t they occupy space. The volhne of the liquid phase is reduced when these additional tsolids a r e present. The character of these solids can be altered by chemical treatment so t h a t more of then1 can be tolerated before a mud loses its desired properties, but eventually their concentration i n t h e mud must be reduced. The practice of watering back, therefore, h a s been necessary to inaintaln the mud in i t s normal range of properties. I n high-density muds the problem i s intensified by the presence of large amounts of barite, which is the most expensive component of the mucl. The amount of water which is added t o the mucl a t any given time depends upon the condition of the mud. The addition of 10 percent t o the volume 1s common, and sornetinles i t is,necessary to add 20-percent water to get a mud into proper condition. This is i n addition to the normal amount of water which i s always going Into the mud while drilling is in progress.

I

The cost of preparing a barrel of 17-1b mud i s $15.85. Suppose t h a t a 1,000-11111 mud ,system needs t o be \i~aterecl back a n amount t h a t will require the jetting aivag of 100 bbl of mud. The mud, a f t e r being diluted, must be conditioned with weighting material and sufficient chein~calsto maintain the desired properties. I n actual practice, the conditioni n g takes place a t the same time a s the aclditiorl of the water, and such additives a s bentonite may not be used. The cost of getting the total system back to ~ t original s 17.0 Ib per gal weight and the use of the necessary chemical aclclitives will amount t o about the same a s would the preparation of a new mud. Conchtioning the mud system by the watering-back method ~ v i l l ,therefore, be equal In cost to the value of

I I

.

.

Table 1 Cost of Preparing 1 Bbl of 17-lb Mud - $ 1

I Rlatenal

Amount Required, Pounds per Barrel

Bentonite .................... 6 Driscose ..................... 1% Graphite ...................... 2 Quebracho .................. 1 Caustic ........................ 1 Weight material ..........475

*Phillips Petroleum Co.. Alvin, Texas : slnce removed to International

Drilllng Co Rome, Italy. SPhllllps Petroleum Co Bartlsjville. Okla. tPresented by Ray Wlll~arns'at the spring meetrng of the Southern Distr~ct.D~visronof Product~on,Shreveport. L a . March 1957.

COST O F DILUTING A MUD Remembering t h a t when a certain,volume of water i s added t o the system, room must be made f o r t h a t water, we ,should first look a t the econon~icfactors involved in watering back a mud. The preparation of a new mud having a density, f o r example, of 17.0 Ib per gal, \\rill involve t h e use of the typical amounts of materials a s given i n Table 1.

,

I

(Handling and breakage-10

5.85

Cost per Pound

Total Cost of Material

0.02 0.90 0.16 0.18 0.09 0.086

0.12 1.35 0.32 0.18 0.09 12.35

$14.41 percent). .......... 1.44

Total cost ............................................................. $15.85

MESAROS RAY WILLIAMSAND JOHN

186

100 bbl of mud worth $1,585 which w a s jetted away plus the cost of making 100 bbl of new 17-lb mud-another $1,585. The cost of this procedure w ~ l lbe $3,170. COST O F CENTRIFUGING A MUD The centrifuge process eliminates the need for wateri n g back a mud because the drilled solids a r e removed mechanically. The barites a r e removed and returned t o the mud system i n the form of a heavy sludge. The sludge must be thoroughly mised back into the rest of the mud i n the tanks. The light solids and water and chemicals a r e discharged into the reserve pit a s effluent. I f 100 bbl of mud a r e processed through the centrifuge, 1 lost and a small amount of all of t h e chemicals ~ 1 1 be barite will be rejected along with the clays, depending upon the efficiency a t which the n ~ a c h i n eis being made t o operate. L e t u s assume t h a t the machine i s t h r o w ~ n g away 15 percent of the barite w111ch it processes. Because there i s $1,235 worth of barite in 100 hbl of the mud, the centrifuge will recover 85 percent o r $1,050 worth of b a r ~ t e(see Table 2). The chen~icalsi n 100 bhl of this 17-lb mud a r e also lost In the process. The difference in the cost of the mud ancl t h e value of the barite recorerecl represellts the cost of centrifuging the mud. This cost illust 1)e in addition t o t h e $1,585 \vhicll was originally spent to prepare tile mud. One thousand and fifty dollars can be saved by using t h e celltrifuge one tillle to keep mud properly conditioned. The watering-hack metllocl is still i n comnlon practice, and in sollle areas i t is necessary to condition tile mud i n this manner a s often a s three o r four times a week. Where tile centrifuge 1s made to perform properly, i t is effectively renloving solids in s u c l ~quantity and over such a range of sizes t h a t these larger water dilutio11s a r e not necessary; n o r a r e the larger and often complicated chemical treatments of the past needed. CENTRIFUGE CONSTRUCTION AND OPERATION The reason t h a t weighted ~ n u d scan be processed profitably through a centrifuge is the difference i n specific g r a r ~ t ybetween barite and formation clays or solicls. The specific gravity of barite is 4.2 \vhile t h a t of foinlaTable 2 Cost of Centrifuging 100 Bbl of 17-lb M u d

Original cost to prepare 100 bbl ............................ $1,585.00 Cost of barites a n d chemicals lost in process ........ 535.00 Total cost ........................................................ $2,120.00 The total cost of centrifuging 100 bhl of 17-Ill mud i s $2,120. This method conlpares with tlle method by which 100 bbl of nlud i s jetted away to be replaced by water and new nlud a s follows: Cost of jetting and watering back ........................ $3,170.00 Cost of centrifuging ................................................ 2,120.00 Difference ...................................................... $1,050.00

{ i '.-J---

P,

t

-

EFFLUENT PORT INLET EFFLUENT PORT,

SLUDG~ Fig. 1-Flow

of Mud through Continuous-type Centrifuge

tion solids averages about 2.5. Several ~ n a n u f a c t u r e r sa r e now p r o v ~ d ~ ncentrifuges g \vhlch a r e adaptable f o r use on drilling muds. The decanter-type machine performs its separation by subjecting the fluid to centrifugal force while passing through a rotating cone. (See Fig. 1.) The ~ n u denters the rotating conical bowl a s a diluted slurry containing solids from the foilnation and heavy weighting m a t e r ~ a l . When the feed ~ n i s t u r e is first thro\vn against the walls of the r o t a t ~ n g bowl, the heaviest particles come out of the llquid and cling t o tlie rotating \vall. The size of the heavy particles determines a t what point in t h e cone the centrifugal force is g r e a t enough t o remove them. The force is greatest a t the point where t h e d ~ a m e t e rof the bowl 1s largest. The heavy barite solids \vhich become stacked against the howl a r e nloved toward the smaller end of the bowl by a slliral conveyor inside the 11o\vl. The conveyor moves in t h e same direction but a t a slightly slo\ver speed t h a n the bowl. The barite is then discharged a s a high density sludge through pelts In the bowl. The ligllt\veight solids will not he thrown out of the feed llquid a s easily a s will the heavies and, therefore, will flo\tr out of the bo\rl \vith the 11qu1d effluent or orerflow through port's a t the large end of the hoxrl. 111 Fig. 1, t.he slleecl a t wlllch the conical 11owl is rotat,ing \\,ill determine tlie amount of c e ~ ~ t r i f u gforce al which i s cle\~eloped. Because of the shape of the horizontally l~ositionedcone, there will always be fluid l y ~ n ga t t h e \vide end of the bowl. When the bowl is rotating, t h e liquid will be thro\x,n around the walls. This liquid foiqns t h e pool and its volume is determined by t h e position of t h e adjustable effluent discharge poi-ts. If t h e pool volume i s large, t h e particles a r e retained f o r a longer period of time than they \voultl be if the volume were ~ h ~the ~longer~ thef ~ ~ are ~retained , i,, the pool of liquid, the lollger they will be esl,osed to t h e forces of rotation. The two factors which most affect the centrifuge process are the celltrlfugal force ancl the length of tinle during \vllich the pal-ticles are sul,jected to force, By usillg lnultillles of gravity or gees to esI,ress the centrifugal force ancl seconds to measure retention time, tlle useful term "gee-seconds" can be produced by simple multiplication. If the pool volume i,s.reduced,the fine particles will not be peinlitted to remain under the influence of centrifugal

THE CENTRIFUGE AND MUDlECHNOLOGP force long enough to be removed with the coarse particles.' Some of the slnaller barite particles \\.ill be rejected along with the clay i n this case. PARTICLE SIZES AND WEIGHTS I n understanding the mechanical separation process of a drilling mud, ~tis essential t h a t the distribution of t h e size of particles of various densities be considere(1.l If a mutl contains weighting material having a specific gravity of 4.2 and foi-nlation sollds having a specific gravlty of '7.5, the process woultl present 110 probleal if all of the particles were the sarne size and shape. Unfortunately, many of the ~ ~ a r i ~ u n - s u l f aparticles te will be snlaller t h a n some of the formation paiticles. The separation process is affected 11y the size of the particles a s well a s their specific gravity. Fig. 2 illustrates t h e particle-distnbutioll situation t h a t might exist in a drilllng mud. Q

w

W LL

f V)

G

-J 0

V)

PARTICLE SIZE-INCREASING Fig. 2-Possible

Types of Separation with the Centrifuge

The cuives show t h a t i t is impossible to separate and discard all of the clay solids ant1 retain all of the barite. Frequently, a s ~ n u c l la s 15 percent of the harite present i n a mud is less that, 10 lllicrolls in size."he ]lest best al,proach, then, i s to separate all of tile clay possible and retain a s muc1.l of the barite a s possible. The cut necess a r y t~ nlake this type of separation ~1 auld be a t the point where the curves intersect a t line 1. Some of the clays \\.auld retained in the sludge and solne barite would he rejected with the clay in the effluent. CENTRIFUGE APPLICATIONS Several considerations now become evident, and the relationship between mud control and centrifuge process hecomes significant. If the c u w e s sho\vn in Flg. 2 represent a \\.eightecI 12-11) per gal rnucl i n a fast-clrllling hole m a k m g > alarge amount of sohds conle into the mud, t h e mud engineer would clecicle t o make a cnt a s shown by line 9. Here, he may actually throw some e s t r a barite a w a y in his effort to relnove all of the drilled solids. H e will conslcler t h e economy and t h e need f o r a certain s e t of nlucl propelties under the existing drilling conditions such a s depth, drilllng rate, and the type of lnud he will need when the well gets deeper. If the cuives i n Fig. 2 represent a 17-lb mud being used while the drilling r a t e is medium, i t will be necess a r y to take a cut a t the point where line 2 intersects the cui-ve. This decision w o u ~ drequire the engineer to r u n IReferences a r e a t the end of the paper.

t -

o W w

187

BEFORE CENTRIFUGING CENTRIFUGING

----- AFTER

LL

V)

1 0

*

PARTICLE SIZE

-

INCREASING

Fig. 3-Particle-size Distribution in Centrifuged Mud is Continually Changing the centrifuge f o r longer periods of time until the drilling r a t e slo\\rs do\\ n. The condition of the mud is all import a n t in such decisions. I t is necessary t o recover all the l ~ a r i t epossible \\,hen centrifuging a heavily weighted mud, but ~t la more important to keep the best possible mud properties. The nlud man may, therefore, have t o slip the separation line back to the right of line 2 even t l ~ o u g hhe will knowingly discard valuable barite. The fallacy of using Fig. 2 to esplain \\,hy different separations a r e necessary should be emphasizetl by p o ~ n t l n gout t h a t the d i s t ~ i b u t l o ncurves f o r particle sizes i n a given mud a r e always changing (Fig. 3 ) . These distributions a r e dependent on the drilling rate, mutl weight, and the amount of time \vhicll the centrifuge h a s already been run.

GUIDES FOR FIELD U S E The practical application of a centrifuge is dependent 011 the engineer who nlust use it. He nlust be the same man \\.ho 1s responsible f o r the care of the mud. He ]>lust know the history of the mud and when he uses the centrifuge, he must know why he is using it. The centrifuge i s usually thought of a s a means of saving espensive b a n t e , which is true. Honever, i t nlust be remembered t h a t this is done by removing and d i s c a r d ~ n gclay (Fig. 4).

0 W W

z

O

PARTICLE SIZE-INCREASING Fig. 4-A Centrifuge Saves Barite by Discarding Clay

MESAKOS RAY WILLIAMSAND JOHN

188

SOLIDS CONTENT O F MUD Several means, a r e a t the disposal of the engineer f o r determining the condition of his mud a s related t o solids content. They are: 1. Stoimer viscometer. 2. F a n n V-G meter. 3. Hand centrifuge. 4. Oil retort.. The use of the oil retort is one of the better methods of deteimining whether the mud is in need of centrifuging or chemical treatment. To use i t or any of the other testers properly, the mucl engineer must know w h a t t h e l~ropertieswill be when t h e mud is i n the best possible shape. Suppose t h a t a funnel viscosity of 50 sec A P I indicates t h a t the mucl is iust a s he wants it. H e will also know t h a t some higher viscosity, perhaps 60 see, indicates t h a t the solids content of the mud is approaching a critical concentration above which the viscosity will rise very f a s t ancl the mud will be in poor condition. By using the oil retort, he can determine the icleal solids content f o r t h a t mud. Case I. If a f t e r several days of drilling, the funnel viscosity i s climbing towards 60 sec, there is a need f o r a solidscontent determination. T,he solids content of the mud will usually he found to have increased. Table 3 shows sonie sample retort determinations.

Table 3 Retort Method lndicates Mud Needs Centrifuging

Funnel Viscosity 50 Seconds, . Percent Oil Water Solids Total

5 64 31 (ideal)

100

Funnel Viscosity 60 Seconds, Percent Oil Water Solids

5 61 34 (undesirable) 100 by volume

This example \rould indicate t h a t the mud needs to be centrifuged. In actual practice, the mud engineer will leave a standing recommenclation f o r the crews t o maint a i n the viscosity I~etween50 ancl 60 sec wit11 t h e centrifuge. A s f a r a s the crew is concerned, only the viscosity i s being controlled hy t h e centrifuge. Elther t h e solids content, a s tlete~mineclwith the retort, o r the funnel viscosity'is a useful index to the mud inan i n keeping t h e mud right. Case 2 Now suppose t h a t the mud from Case 1 is found t o have increased in viscosity t o 70 see. The retort method shows t h a t the solids content remains unchanged. This would immediately indicate t h a t the lnud has been contaminated and requires ~ h e i n i c a ltreatment r a t h e r than centrifuging. See Table 4.

I

Table 4 Retort Method Indicates Mud I s Contaminated

Ideal Funnel Viscosity 50 Seconds,

Undesirable Funnel Viscosity 70 Seconds,

Percent Oil 5 W a t e r 64 Solids 31

Percent Oil 5 W a t e r 64 Solids 31 100

100

ADVANTAGES O F LOW-SOLIDS MUDS The advantages of low-solids, low-viscosity systems a r e well known t o drilling engineers." Pumps r u n better and p a r t s need not be replaced a s often. The pump discharges the fluid with less pressure buildup. The drill stem i n which most of the pressure drop occurs can handle the same velocities a t reduced pressures. More of the hydraulic horsepower which is expended inside the dl-ill plpe and collars can be delivered to the bit nozzles in t h e form of greater velocities. The cuttings a r e removed more cluickly from beneath the bit teeth and a r e not reground to finer particles. Velocities In the annulus, too, a r e greater. This helps keep the hole cleaner and brings the c u t t ~ n g sto the surface faster. These advantages exist I?ecause the light formation particles a r e not present in the mud. The pump is not required to circulate them through the system. The additional pressure t h a t must be used when t h e unclesirable solids a r e present is no longer needed and can be used to gain other acl\~antages. This removal of drilled solids with the centrifuge h a s affected all phases of mud-engineering problems. New approaches a r e now being made t o such situations a s lost c ~ r c u l a t ~ o settling n, of cuttings, penetration rates, fluid velocities, and hole conditions. LOST-CIRCULATION EXAMPLE Use of the centrifuge and a working knowledge of flow properties must go hand i n hand. This was sho\vn during a lost-circulation prol~lemon a wildcat well on t h e Gulf Coast. Table 5 1s a list of pertinent well data. -

Table 5

Well Data

Mud Data

Depth, f t............. 11,985 Hole, in ................. 8% Drill pipe, in ......... 5 Pump stroke, in. 20 6 Liner size, i n ....... 43 Strokes per min.. Gage pressure, psi .................... 1,250

Mud weight, Ib per ga1..17.0 Fumlel viscosity, sec....62 Plastic viscosity, centipoises ................41 Yield point, 11) per 100 sq f t......... .....23

THE CENTRIFUGE AND MUDTECHNOLOGY

P.u = [11-,925 ( 1 1 ) / 2 2 5 ( 3 . 5 ) ]

As soon a s mud loss was recognized the p u n l p was slowed down from 43 strokes per min t o 28 strokes per min. This effected a reduction in t h e gage pressure from 1,250 psi to 750 psi, o r 500 Ib. The annular velocity w a s reduced fronl ft to 117 f t per Closer scrutiny was made of the problem t o deterinine what the actual effect of slowing the pump down was in reducing t h e pressure of the circulating fluid a t t h e I~ottomof the hole. Use was made of t h e pressure-drop equation f o r laminar flowing fluid^.^

P =

(lty/22.5P)

p3 = 166.6

=

P2 =

+

[ ( 4 1 ) (11,925) (2.9)/ 1,500 (3.5)'l 4.25 psi ( ~ ~ o t t o m - h o lcirculating e pressure a t 43

strokes per mln velocity)

01.

178 f t Per min annulal;

+

[11,925 (2.3)/22.5 ( 3 . 5 ) ] [ ( 4 1 ) (11,925) ( 1 . 9 5 ) / 1,500 (2..5)?] p2 = 400 psi (bottolll-hole circulating pressure a t 28

strokes per mill or 117 f t per velocity)

111111

annular

This calculation revealed t h a t by slowing the mud pump down from 43 s p ~ nt o 28 spm the pressure a t t h e hottom of t h e annulus was reduced 1 ) ~only 25 psi. T h o u g l ~ circulation i s frequently regained 1 ) ~tllis method, the f a c t still :emains t h a t the pressure eserted on a weakened formation had not been greatly altered. The approach to the problem mas i n l ~ ~ r O vthe e flow propel-ties of the mud. The centrifuge was used to fui-ther reduce the solids in t h e mucl, and a small alnount of water and chemicals were added f o r the purpose of reducing the y ~ e l dpoint of the mud. The weight of the mud was not altered from its original 17.0 lb per gal. A f t e r conditioning the mud, the yield point had heel1 reduced from 23 l b per 100 sq f t to 11 113 100 sq f t . The funnel viscosity was 50 sec. Calculations were again made with the new mucl properties a s shown in Table 6.

Table 6 Mud weight, Ih per gal .................................................. 17.0 Funnel viscosity, sec.................................................... -50 Plastic viscosity, centipoises...................................... -41 Yield point, 11) per 100 sq f t............................................11 pulllp strokes p e r min.. .................................................. 43

+

77.2

strokes per min of mud was reduced to 11 Ib per 100 sq f t )

pressure drop in pounds per square inch, over the length being calculated. n = 'plast'ic viscosity of the mud, in centipoises. = total length, ill feet, of the diameter considered. V z mean fluid velocity of tlle mud, in feet per second. t y = ~ i e l i point l of t h e mud, in pounds Per 100 sq ft. D = bit size minus pipe outside diameter, in inches f o r annular flow.

PI'=

[ ( 4 1 ) (11,925) ( 2 . 9 ) /

Ps = 244 psi (bottom-hole circulating pressore a t 43

+ (?~1~/1,.500~~)

PI = [11,925 ( 2 3 ) / 2 2 5 (.?.5)]

+

1,500 (.?.5)']

1Blt.evcin:

P

189

Conditioning this mud so t h a t the yield value was lo\\,ered resulted ill a reduction of the bottom-hole circulating pressure by 181 psi. This represents a sign~ficant ~mprovementin the hydraulic system. Besides solving one particular lost-circulation problem, here is very definite proof of the effect of high yield points on circulating pressure. F o ~ ~ n a t i osolids n in the mud cause high yield points. Reducing the punlp pressure, a s is the first tendency when mud is lost to the formation, actually gives only limited assistance t o t h e problem a t hand. The 500-psi change in the standl~ipe gage reading was in the most p a r t caused by reduced pressure drop inside the drill stem where the mud is in turbulent flow. I n turbulent flow a large' decrease i n pressure a c c o l , p a ~ e s only a small reduction in \7elocity The yield point of the mud is by f a r the most significant factor affecting the pressure drop when the mud IS i n laminar flow. The practice of loading up a mud with lost-circulation materials In thief zones is f a s t losing favor in many areas Tile presence of high collcentrations of sucll materials in a imparts flowing characteristics to the mud w111c11 a r e more likely t o increase t h a n t o decrease lost circulation. The materials t h a t stop loss of inud can also be the cause of lost circulation. The best approach t o this problem i s based on having the mud in such cond~tiont h a t pressures resulting from circulating the mud ~ v i l lhe minimized. THE CENTRIFUGE IS VERSATILE; SAVES ON WATER AND CHEMICAL The centrifuge is the best means by which the yield L~~~~ I,oillt of a lnud can be kept to a lllil~lllul,, chemical addltlons or large dilutions Lt
I,~

,

190

MESAROS RAP WILLIAMSAND JOHN

The history of a well near Bay Clty, Texas, i s also the story of t h e versatility of a centrifuge. The cost of s troublesome land-based well was o17er drilling t h ~ very a million dollars. A high pH lime-lmse mud was used to drill through 2,000 f t of heaving Anahuac shale. During this tlme cuttings did not appear on the shale-shaker screen. Instead, soft mud balls were present in the nlud inclicati n g the type of problem the fowlation was -presenting. Fifteen hundred to 2,000 111 of lime were required each tour t o keep t h e solids i n the system converted t o calcium. On drilling out of the shale a t 7,500 f t , a protection string w a s set and there was no longer a need f o r a lime-base mud. A t this time the lime content of t h e mud w a s 8 Ib per bbl. T11e centrifuge continuously, processed the mucl and in 10 days the 8 Ih per bbl lime content was reduced t o just a trace The pH of the nlud was maintained a t 11.5 t o 12.0 to preserve the starch which mas still present in the mud. Later the pH was allowed to come down t o a value of 9.5. T h ~ scentrifuged mud mas eventually used a s a fine completion fluid. Not a drop of mud h a d - t o be jetted away while drilhng from a depth of 7,000 f t t o total depth, which was 14,500 ft. Such low-solids centrifuged muds make very successful packer muds. High-temperature gellation is not a problem with such a fluid. , SALVAGE RESERVE MUD The Phillips Petroleum Company's centrifuges a r e made t o l u n about 75 percent of t h e time t h a t drilling is in progress. When the centrifuge is not used on the circulating system i n t h e well, i t is frequently used t o process lnud t h a t h a s been stored in a "duck's nest" or small separatecl section of the resellre pit. This nlud lnay be processed into a good high-density packer mud which is stored in tanks f o r future use. I t may process resenre mud ancl deliver the barite into the drilling well f o r weighting-up purposes. A t one time i t w a s the practice to locate discarded weighted muds in nearby areas ancl deliver them t o a central location i n the Chocolate Bayou Field. The centrifuge would then retrieve the barite from t h e muds f o r use i n weighting up or preparing other muds. The total cost involved i n this method \\-as usually t h e $1.00 or $1 50 per bbl f o r hauling the mud. Muds n;eighing 1 lb per gal and up \\?ere processed in this manner and u l t i n ~ a t e l y saved t h e company consiilerable sums of money. OPERATES DURING T R I P S Sometimes drilling rates a r e so f a s t t h a t t h e centrifuge h a s difficulty keeplng the solids content down t o the

amount consistent with good mud properties. The centrifuge should then be allowed to run on the surface s y s t e ~ nduring trips and other shut-down periods. This type of operation will pemnlit the circulating mud volume to be reduced sufficiently t o permit the use of more water and barite without having t o jet away p a r t of the weighted mud. F o r example, during a t r i p 100 bbl of a 17-lb mud may be centrifuged. The light clay solids and liquid a r e discharged a s effluent and the barite is retulmecl t o t h e pits a s a 23 111 per gal semi-fluid sludge \vhich is then redispersed into the system hy agitation. The volume of t h e 23 111 per gal mud i s 55 l>bl. W a t e r i s added t o keep the \veight a t 17 Ib per gal in the mucl tanks. Room h a s been made available f o r 45 bbl of water in t h e system without t h e need f o r jetting. The water can be added i n such amounts a s t o condition the mud so t h a t the viscosity is lower, the flow properties a r e improved, and p a r t of t h e light foimation solids have been removed. The conditioned surface mud when circulated through the hole again mixes with the rest of t h e mud and the system h a s been improved while the pipe was out of the hole. CONCLUSIONS To the mud engineer \vho works with \i.eighted muds, the centrifuge h a s become a tool \\.it11 which he can make and maintain better muds t h a n ever before. H e can control the funnel viscosity a s well a s other properties with the machine. To the operating companies the centrifuge h a s proved t o be a means by which significant reductions in well costs a r e effected by barite savings. Drilling engineers see the centrifuge a s a means to better penetration rates and fewer liole problems. Drilli n g crems like the centrifuge because i t is automatic ancl foolproof. A s the centrifuge runs, the crews a r e aware t h a t the machine is saving them the work t h a t would be required to add new materials to the mud. ACKNOWLEDGMENT The authors wish to express their appreciation t o Phillips Petroleum Company f o r permission t o publish this paper. REFERENCES 'Bobo, Roy A. ancl Hoch, R. S : The Mechanical Treatment of Weighted Drilling Mlids, AIME P a p e r No. 290-G. "obo, Roy A: Personal commun~cation,J u n e 1956. "heless, N. H. and Ho\re, J a c k L: Low-Solids Muds Improve Rate of Drilling, Cut Hole Time, Drillirty, 70, Aug. (1953). "Clark, E. H. JY:Bottom-Hole Pressure Surges While R u n n ~ n gP ~ p e ASME , P a p e r No. 54-Pet. 22.

APPENDIX CENTRIFUGE CALCULATIONS The size of the pool volunle is a n essential factor in obtaining the best separation. The volume of the pool is determinedby t h e position of the effluent poi-ts a s shown i n Fig. A l . This volume may be determined by a simple

calculation. The difference between the volume of a core and the volume of a cone will be t h e volume of the pool. Follolvlng a r e the calculations used to determine t h e pool volume of a n 18-in. decanter-type centrifuge. ,

THE CENTRIFUGE A N D MUD TECHNOLOGY

80: 1 GEAR RATIO

X

1000=GEE -SECONDS

Fig. Al-Pool Volume is Dependent upon Position of Effluent Ports

Fig. AS-Efficiency

Volu~neof f r u s t r u m of cone = ?h TI!. [( R R')' % ( R - R')?] R' = 5 in. R = 9 in. H = 22.7 in. 5)" ( 9 - 5)'/31 F u l l volume = 2 2 . 7 ~ / 4 [ ( 9

N = revolutions per minute. R = bowl radius, 111 inches. 2.84 = constant.

+

+

+

=

2 . 7 ~ / 4 [(14)" (4)'/3] = 5 . 6 7 5 ~ [ I 9 6 $ 5.51 = 5.61'5~ (201.5) = 5.588.9 CIC.in.

Less core = S.5SS.9 -r

=

1 ,80~i ell

(5)"(7!2.;')

= 8.588.9 - 1,782.9

ill.

Volume of pool = 1,806 crc ill. = 1,80!;/2S1 = 7.8 gal Once t h e effluent ports a r e positioned and the pool volume is known, t h a t value is a constant. The retention time m a y be detei~ninedif the feecl r a t e is laiown. pool volume in gallons Retention time = feed r a t e in gallons per ~ilinute The answer is in minutes \vhich a r e usually converted t o seconds t o avoid the use of fractions. This is done by multiplying by 60. The r a t e a t which particles pass through the pool depends upon the r a t e of feed. The degree or fineness of the separation will i n t u r n be dependent upon the time of retention. The centrifugal force t o which the particles a r e e s posed while they a r e retained in the 11okvl may be e s pressed a s n~ultiplesof gravity o r gees. The fornlula f o r detelqnining gees exerted hy a centrifuge is: Gees = (2.84) (10.') (A7*) or Gees = L2.84 ( N ' R ) ] /100,000

DISCUSSION G. R. Gray (Drilling Fluids, Inc., Houston, Texas) (written): Tlie savings i n barite made possible by t h e use of the centrifuge on heavily weighted nluds have been demonstrated in several previous papers. These savings a r e significant and adequately justify the use of the machine. I believe, however, t h a t other benefits less easily proved m a p a c t ~ ~ a l leffect y greater savings i n drilling costs t h a n the obvious reduction in cost of barite.

of Barite Recovery from an 18-lb Mud

The two factors \vhich most affect the centrifuge process, therefore, a r e the centrifugal force (gees) and tlie length of time during which tlie particles ape subjected to t h a t force. Multiply~nggees and retention tiine in seconds provides the ternl, gee-seconds. Gees

x retention time in

seconds =

gees x pool volume x 60 gallons per minute flowing (feed r a t e )

or

gees x 468 Gee-seconds = feed rate, gallons per minute Fig. A2 shows the ,relationship between gee-seconds and the centrifuge efficiency f o r maxinlunl barite recovei-y from a n 18 lb per gal mud.

CENTRIFUGE ADJUSTMENTS Tlie adjustments which can be nlade in order t o obtain t h e best efficiencies on the centrifuge are: 1, speed, 2 , volunie of pool, S, locatlon of feecl plpe, 4, viscosity of feecl, 5 , feed rate. Generally speaking, the follo\i
192

RAY WILLIA~VS AND JOHNMESAROS

is the use of the centrifuge t o reduce the density of a saturated saltlwatdr mud by removing salt cuttings while rock salt i s being drilled. Another instance i s t h e removal of salt contamination a f t e r a salt-water flow. Without the centrifuge, either continued large chemical treatments or estensive dilution i s generally necessary t o control mud properties a f t e r a salt-water flow. The centrifuge can be used t o change the type of drilling mud a s conditions i n t h e hole change with depth. F o r example, a lime-treated ~ n u dcan be changed with t h e centrifuge t o a lo\v-alkahnity mud when high temperature begins to adversely affect the properties of the lime-treated mud. Periodic shutdowns t o condition mud a r e elinlinated by t h e regular use of t h e centrifuge. Do the authors have any d a t a on the importance of this f a c t o r ? In thew calculation of the saving of barite the authors imply t h a t the chemical costs a r e increased i n direct propol-tion to the anlount of nlud t h a t is centrifuged. What a r e the relative costs f o r materials required t o control viscosity ancl filtration of, let u s say, a 17-lb per gal mud with and without t h e centrifuge? Was t h e cost of the centrifuge includecl i n the estimations made in the p a p e r ? What can be done t o recover the oil from a n oilemulsion mud ? How does the mud engineer lay out a program f o r t h e operation of the centrifuge? What e x t r a attention i s required ? Mr. Mesaros (written) : The authors a r e in agreement with Dr. Gray's statement t h a t barite recovery i s a n important economic factor t h a t i n itself justifies the centrifuge. I t is also agreed t h a t t h e other benefits of a centrifuged mud \vhich a r e less easily proved may effect greater savings in d r ~ l l i n gcosts. I t must he remembered, however, t h a t the real purpose of the centrifuge i s t o remove foimation solids from a nlucl so t h a t the 1?1ud will remain i n good cond~tionf o r greater periods of time. All other benefits and advantages from using the machine a r e secondary ancl rdsult from the removal of the solids. The solids-content vs. penetration-rate examples which Dr. Gray requests a r e not available. They were not needed to prove t h a t the centrifuge was a n effectlve machine. T h ~ stype d a t a a r e difficult to produce in the field because of the ~ m p o r t a n tp a r t t h a t bit weight, punlping volumes, ancl general drilling practices play i n penetration rates. I t is generally accepted t h a t a low-solids mud will permit f a s t e r drilling. The lower solids, of course, a r e reflected a s good flow characteilstics in the mud which in turn can be shown by calculation to produce lower pressures when being pumped through the system. I t is not necessarily true, however, t h a t a centrifuged mud is always a low-solids mud. The centrifuge removes foinlation fines which a r e i n a size range of perhaps up t o 10 niicrons. These particular fines, we I~elieve,a r e the ones \vhich first impart poor characteristics t o a mud. The sands and all other coarse particles not i n this range and which do not readily break down with agita-

tion can also build up in concentration until the mud h a s to be conditioned by water dilution and jetting. This coarser fraction of formation particles is retui-ned t o t h e system \i-ith the barite sludge and does not affect the mud properties a s readily a s the fines. Phillips centrifuged muds (lo not always shuw a decided decrease i n solids content over muds t h a t were not centrifuged. This is pai-ticularly true of high-density muds where only a small amount of fines c a n affect t h e tnud adversely. Obset-vation of solids contents of niucls can serve only a s a guide in their treatment, but the proof of t h e effectiveness of the c e n t r ~ f u g eis found i n improved flow properties of the mud. Data were not collected to show the importance of less shut-down time f o r mud conditioning due t o t h e presence of the centrifuge. This fact 1s recognized ancl is impoi-tant and a s was pointed out in the paper, t h e tnud is always being conditioned while drilling takes place. Also conditioning may be had by running t h e centrifuge on the surface system cluring trips and other non-circulating periods. We a r e sure t h a t mud-conditioning time has diminished on rigs uslng centrifuges 11ecause the mud is supposecl to be 111 good condition all the time. Less time is now used f o r circulatiori of the lllud prior to logging or other non-drilling operations. This time used t o be a s much a s 8 o r 12 hours. Now, i t is frequently one circulation or less. The authors did not imply t h a t the cl~emicalcosts a r e increased in dlrect proportion to the amount of mud t h a t is centilfuged. The f a c t is t h a t fewer chemicals, particularly d i s p e r s ~ n g chemicals, sl~oulcl be required because the clays which they disperse a r e not present, inasmuch a s they \\.ere removed mechanically. Filtration-control agents will be r e q u ~ r e dt o lower the water loss a s usual. Sometinles the fines when removed from the mud will upset the pal-ticle-size d i s t r i b u t i o ~suffi~ ciently to requlre a sack or t\vo more of D r ~ s c o s et h a n would ordinarily be required had the fines been present. Dr. Gray's request f o r d a t a of chemical cost requirements of a 17-lb lnud with and \vithout the benefit of a centrifuge cannot be fulfillecl without further lengthy studies. When t h e centrifuge \\,as first put into use in 1953 i n the Chocolate Bayou Field, a 30-day ol3servation of cl~emicalrequirements of the mud let1 t o an estimation t h a t about 80 percent fewer chemicals were required f o r daily maintenance. This figure i s not now considered realistic because i t was based on factors ancl conditions which were not t l v e test conditions. I n t h e Chocolate Bayou Field mud costs per average well were $125,000 prior t o the advent of the centrifuge. The average cost today is about one half this amount. Here again i t would be unrealistic t o suppose t h a t the centrifuge effected all of this reduction. During the past five years drilling techniques and equipment have so improved t h a t it i s safe t o say t h a t t h e centrifuge only contributed t o this mud-cost reduction. We do not know how to recover the oil from emulsion muds. The oil is valuable and the problem is not insurmountable. I t just has not been done yet.

%,

The lnud engineer must r e n ~ e ~ n b einr laylng out a centrifuge program t h a t he must still rely on his rnucl engineering efforts to keep a good clrllllllg mud. He lllust st111 use water and chemicals. He must also renienlber that there a r e times t h a t the centrifuge need not be 1x11. If the flow properties of the mud a r e gootl tile clay solids a r e suffi~ientlylow, then furtheFuse of the centrifuge is wasteful. Every bake1 of lnud t h a t is centr~fuged unnecessarily may contain $3 o r $4 \vorth of chelllicals and 011. I n a day, $1,500 worth of chemicals inay be

1

(

I

wasted. It is necessary to have criteria f o r cletel-nlining whell to stapt and' stop the lnachine. Ordinarily, the viscosity, yielcl of the mud, and plastic viscosity a r e goocl indicators of the need f o r centrifuglng. Generally, when the lnud needs to be jetted to make room f o r more water, the centrifuge should be ~ u n l l i n g to maintain the required density while adding water. One or more of the afore~nentionedflow properties should be the guide f o r operation of the centrifuge.