SPE

md~mm=e

SPE 18370

Hydraulic Optimization of PDC Bits by Visualization Methods

by P. Bisson, J.C. Choux, and A, Provo, TOTAL CFP

-Mt 1SSS, $wIatY of Pefrofaum Englnaara

Thl$ ~ WOOpraparad fti pr@aantNlon ●t ma SPE Europaan Pefroloum Conference, London, UK Cofobar 1S-19, 19SS.
Thb papaf woo aahotad for praaanfatlo+Iby ●n SPE Program Cornmlftao fo!lowlng revlaw of Infofmstion oontalned In an abafraot aubmlffad by the
author(o), Oonfanfa of MO papar, u prnantad, hava not boon rwdawad by W SOMY of POtfOfOUMEn@naara and are subjact to corractlon by the
auffw#(9). Tha Matarlai, 00 praaorlfari( doaa not nacaaamify roffaof ●ny poaivon of m~Soabty of Pafrdaum EnaWmIfa offiom Or Mmwc. paWra
pmnntod at SPE maatfnga ara !wbjaci to pubfloatbn Wow by Edlfortal Oommlff@aaof m. Socloty of PafroWm En@raara. Parmlaakm to OQPyla
mtrbtad 10 ●n abatraot of nof rncia than 300 wcda. Ilfuafratbna may not ba oopbd. llta *aof ahou!d oontaln conapkutwa acknowtadgrnard of
whoraandbywiwrn tfra paper b prwantod. WfO PUMMOIM MOIIOW, WE, P.O. SOX-. R~, ~ 7MWSSSS, Tatox, 7309SS SPEDAL.

AMl!RM!s

The number of ways in which cutters are P.D.C. bits are being used more and more
distributed,and the number of shapes of frequently. The improvement in
PolycriatallineDiamond Compact (P.D.C.) infbtantaneous Rate of Penetration,
bits are increasing every day. These bits extended bit life plum the absence of any
can be fitted with either nozzles or mobile parts, limiting the risk of
crowfoot f hid outlets. The fluid will malfunction, explainsuch a trend.
spread aroun~ the bit in a more or less These bits, although expensivedue to the
complex manner depending on these outlets price of the technology and materials
and on the geometry of the annulus involved in their manufacture, make it
encountered.This fluid distributionwill possible to reduce coste per drillecl
have an influence on the efficiencyw:.>h ‘ meter, and hence the cost of wells.
which the cuttings are evacuated and on
the cooling action on the cutters, i.e. Consequently, bit tnanufacturers have
the capability to drill faster and for widened their range of products,modifying
longer. profiles, outter distribution and bit
hydraulics.The latter modifications,bit
The methods by which flow structurescan hydraulics, are undoubtedly the most
be determined, are high-performance difficult to grasp, as the physical
numerical methods and visualization phenomenonof fluid flow is so complex.
experiments. A proper understanding of the fluid
distributionaround the tool is necessary
This paper d:&cribes a testing sequence if the most is to be made of a given
for flow visualization,which has been profile and cutter distribution at the
developed in the laboratory, plue the cutting face.
necessary equipment.
When drilling soft formations,the R,O.P.
These tests revea1 the hydraulic can be very high (60 to 80 m/h during the
structures, and those lltostfrequently 12~11/4phase) : it is necessary to carry
encounteredwill be described. the cuttings to the annulus quickly so
that they do not get recrushed,
A concrete example showing the application just means extra and pointless wor!’h%
of our studieswill be given : the case of the cutters) and to avoid initiating
bit profile improvement used by one of balling-up phenomena (un-evacuated
our subsidiaries on an oil field under cuttings collect in the vicinity of the
development. bit).
In the c~se of harder and/or more abrasive
The limitations of our experiments, formations, the active part of the
allowing for the tests run and the cutters, consisting of synthetic poly-
eguipmentavailable,will be discussed,as Crystalline diamonds, is the seat of a
well as future lines of research. rise in temperature, whioh can cause
, partialor completedestruction.
Referencesand illustrationsat end of paper
.-
 . *

2 HYDRAULICOPTIMIZATIONOF P.D.C. BITS BY VISUALIZATIONMBTHODS SPE 18370

These two important functions : evacuation A series of tests is adoptedwith a view tc

of the cuttingsand cooling of the cutters, the technical (confinedenvironment,etc.)
are providedby the drillingfluid. and economic (testing costs, material
requirements)constraints : Four tests are
It is therefore important to have a applied to each bit, the tracers used for
thorough knowledge of the distributionof visualizationare paint and particles.
this fluid around the bit for the follmwing
reasons : Preparationprior to testing :
to avoid the zones of low velccity
oppositethe active parts of the cuttnrs In order to create the same geometrical
to improve the path taken by the conditionsas those prevailingdown hole, a
cuttingstowardsthe annulus. prcfile which surrounds the bit iS
The purpoee of our tests is to provide as manufactured.A plexiglassblock (providing
accurate an image as poesible of the flow a transparent environment for
distrlhutionand its structures. visualization)is machined, using a lathe,
to slightly smallermeasurementsthat those
of the bit to be testec. The bit itself is
then used as cutting tool to finish off the
Figure 1 shows the complete installation, machining.After polishing the plexiglass,
which consists of a hydraulicbench set on a print, such as that left at the bottom of
a structure (shown in extreme horizontal the hole, is obtained.After this %ounter-
and vertical positions) and a closed formn and the bit have been eet in the
hydraulicloop. A central unit containsall hydraulic bench, we have a similar set-up
the variouscontrols. to that of a bit drillingdown hole.
The bench has been designed for There are three teste based on the ‘I?;ntn
visualization:there is no rock destruction marker :
during our tests, Each of our modules is - fresh paint test
provided with large portholes.Fig.2 shows - pointingtest
the principleof the bench : it consisteof - abrasionteet
two tubes, and a jack system to place the
bit in the desired position. A hydraulic 1) The ‘fresh paintlltest consists in
motor provides the driving force for bit coveringthe bit with a first coat of paint
rotation.The bench is water proof. which is left to dry. A second coat of
paint, whose colour contrasts as strongly
The hydraulic loop (fig.3) provides the as possible with the first coat is then
fluid-feed to the b~ohed (water). The applied.The bit is inetalledin the bench,
circuit is in a loop. Two and water is circulated at the required
centrifugal pumps (1# and 250 kW flowrate. The test only lasts for a few
respectively) are in parallel. minutes. Finally, the bit is photographed
Depending on the required circulationflow from every angle.
ratee, one or the other will be used (the
maximum flow rates are in the order of 3500 In this way, a pattern of lines of fluid
l/rein)
. A third of the total flow passes will mark the surface of the bit. These
through a heat exchanger.Apart from one of marks are the result of overlaying three
our tests, which demands up to one hour~s condition : two traneient (start-up and
circulation, very litte increase stopping of the pumps), and one steady-
temperatureis to be noted i state (circulationat a given flowrate).
dampening tank, with a 2111J;?pa%’&”~! Observationof the test shows that, during
integratedin the loop. A set of pressure, initial circulation,no paint is removed.
temperature and flow-rate gauges is During the transientstate due to stoppage
installed, and these both gather the of the circulation,the flow becomes less
parameters of the experimentsand control turbulent, due to a reduction of the
the personnel and the flowrate.We can therefore assume, as the
instal%?a?;~nd%in~teeting. paint is very viscous and lteteel-wettablelt,
that any possibleerror is slight.
Certain extremely phenomena
(monitoring the displa?e~?nt of solid Photographs taken during these tests are
particles materializing f~o:d particles) enclosed.
have to be gathered subsequent
analysis. A high-speed oamera (1000 to 2) The !Ipointingw test will only be run in
12000 shots/eec. full format) fixes these the case where there is a doubt regarding
phenomena on film. They are analysed using the flew direction after the fresh paint
a digitalizingtable connected to a raicro- test.
computer. This stores all the data, which
are processed by an appropriate software
package.

...
3XI
 ✎
●

.SPE 18370 P. BISSON - J.C. CHOUX - A. PROVO 3

ED P~
Once the first coat of paint has dried on The results obtained from the study of two
the bit, drops of paint of a contrasting types of bit a: fish-tailand b: blade (see
colour are placed where there is some fig. 5) will be used to illustrate the
doubt regardingthe direction of flow. The followingparagraphs.
bit is then placed in the bench and
circulation established as previously. The studies have revealedvarious ditferent
Observation of how the drops of paint flow structureswhich characterizethe way
spread over the surface of the bit, will the fluid flows on the bit : these are
clear up any doubts. referredto as IWorticeslt,Ilimpactpointsll,
ilcounter-currentS” or “zones of
3) The “abrasionlltest consists in covering considerable slowing of flown. These
the bit with several layers of paint, ea~h structures are generally revealed by the
one contrastingwith adjacent coats. Each wall flow (painttest).
coat must be left to dry for 24 hours
before the next coat is applied.The bit is A correlation between the path and the
then set in the bench and silica powder is particle flow rate is necessary to
introduced, the concentration being determine the extent and the influence of
adjustedaccordingto the rate at which the these structuresin the overall flow.
layers of paint erode.
The narrower the cross-section through
Circulation of the abrasive fluid iS which the Ciuid passes, the greater the
stopped at regular intervalsfor photos to , complexity of these structures. on fish-
be taken, and this series of photos will tail type bits (fig.6) macro-structures
shown the evolutionof the phenomenon. appear (the term “macro” is used when the
structure size is in the same or simila~
This test provides qualitativeinformation, order of magnitude, either as an inter-
as it determines the zones of maximum blade space, or as a channel),the fluid ia
turbulenceof the fluid. Moreover,when the not trapped,there is a constant influx ant
whole series of tests has been interpreted, outflow, without either becominq
the abrasion teet results can provide a disorganized. It would appear that these
confirmation or otherwise, of certain can be assimilated to virtual bodies ix
conclusions. constantcommunicationwith the rest of thf
flow.
Fig 4 shows a photograh of an abrasion
test, which, as it is normally in colour, It appears logical to assume that if thesf
is very meaningful. same structures were isolated (n(
communication) there would be al
accumulationof cuttings in them, and thul
There is also a test based on th& a possiblemodi:rication of the flow.
‘particle~t marker, which consists
introducing particles into the hydraulic With blade-type bits (fig 5b), where th{
system. cross-sectionsof passage are smaller, thl
flow generates complex micro-structure!
As these pass the bit, the high-speed whose dimensions and origifisa~e nOt S(
camera is triggered. Since the field of simple to define.
vision is limited, this test can be
repeatedif Pacessary. The fluid outlet type (nozzleor CrOWfOOt~
can be the cause of certain micro.
When the films have been developed structures (detectedby the “paint” test)
(approximately4000 shots/filmrepresentin g There are however,others, equally complex
0.5 to 1,5 fiecondsof filming) these are and occurring far enough away from th~
~&ilized by means of a dlgitalizin g fluid outlets to make it obvious that thes{
“ the film is projected onto this are not the source. It would thus appeal
table at a given scale. An electronicpen ~~at the presence of these micro-structure[
marks the position of a particle an d connected to that of neighbourim
follows it on the pictures until it structures (fig.7). Although the flo
disappears. The successive positions are patterns are extremely complex (as prove
recorded in the memory of a micro-computer. by the tracks left by the paint) we hav
Knowing the speed of filming, it is then noted a perfect repeatability of th
Ceasyto determine the instantaneousspeed patternmade by the lines of paint when th
of this particle. test is re-run in the same conditions.
This operation is performed for a large
number of particles.
The stored data can then undergo various
statisticalprocessing.
 . ●

4 NYDRAULTCOPTIMIZATIONOF P.D.C. BITS BY VISUALIZATIONMETNODS SPE 18370

This phenomenon is encounteredon all the

bits. The size and evolutionof each vortex Fig 12 shows the resultsgiven by the paint
will depend on the type of bit. test, the direction and modulus of the
velocity vectors given by the particle
On fish-tail type bits, the fluid being test, and the side of the bit where the
circulated is the seat of several non- counter-currents (hatched areas) are
stationary structures, whose range of encountered,respectively.
action is clearly definedby the paint test
(fig.6and figs.). The direction of the nozzle In relationto
each of the channelsis of importance
On blade-type bits ( fig.5), the fluid
being circulatedis no longer the seat of a
macro-structureas the cross-sections of
the passage of the fluid are small. These ION OF ~
vortices are not establishedas on Mfish-
tailn bits, but are local phenomena which once the tests have been completed, a
do not govern the overallflow (fig.9). general interpretation of results is
necessary.The structureof the flow to be
~Daat Ddtlii establishedmust be in agreement with the
An impact point is a zone where the fluid results and observations relating to the
hits the surface of the bit and splashes three tests (paint, abrasion, particles),
out in all directions.The ‘paintM test is each consideredseparately.
an excellent means of showing up this
phenomenon, but is not easy to detect with The procedure usually followed consists in
particle mon~e~dndg, eith& because the considering first of all the paint test.
enlargement to the event With a certain amount of experience, a
correctly is too great, or because visible preliminary estimation of the three-
access is poor. dimensional flow can be established,The
abrasion test makes it possible to detect
The impact points or zones are usually to the most turbulentzones, where usually the
be found in large-sizedchannels,or in the paint test cannot be interpreted. The
vicinity of straight-set nozzles or particle test, used for the quantitative
crowfootswith straightoutlet profiles. informationit supplies, serves to %racka
a particle within a structure which we
It im very difficult to define the origin consider to exist (interpretation)but
of the impact point due to the extremely which it has not been possible to evidence
complex hydraulicphenomena Involved,which (tests)”.
are the result of many parametres,such as
the type of fluid outlet, bit profile, or The particle test allows one to define the
shape of the channelcross-sections. average direction and intensity of the
fluid velocity vector at a given point
In the case of the bit shown on fig 8, the (fig*12 : channel showing a counter-
print of the bit left in the formation is current).
parabolic,and the central body containing
the nozzles is a cone (fig 10), the impact Once the interpretationhas been completed,
point therefore appears to be due to the the overall distributionof the flow can be
space geometryencounteredby the fluid. assessed (fig 13). Cross-sectionsgive a
three-dimensionalview of the flow : fig.lo
In the case of the b-type bit (fig.5),the shows a turbulentflow structuregenerating
impact points occur beyond the straight-set an impact point such as that seen, in
nozzle. At the outlet of this nozzle, the fig.8.
fluid breaks on a blade, each part flowing
along the edges of the channel, and then In order to check the validity of our
hitting the surface of the bit (see fig interpretations, we asked ARTEP
11) In the other channels, fitted with
● (Association de Recherche sur les
nozzles set at an angle of 8 or 11, no Techniques d*Exploitationdu P&role - a
Impactpoint was detected. petroleum research association grouping
ELF, IFP and TOTAL CFP) who are
investigating the possibilities of
The flow at the wall (“paint” test) numerical analysis for defining flow, to
generallyreveals counter-currents,i.e., a run a numerical calculation on a tested
local back-flow of the fluid towards the bit. There iS good agreement between
nozzle. Impact points, vortices and figs.14a (interpretation)and 14b (graphic
obstaclesgeneratethis type of current. printout- Calculationon Cray XMP).
. .

SP% 18370 P. BISSON - J.C. CHOUX - A. PROVO 5

\
or ouR~

The results of our studies have been used The followingcan be given as some possible
to improve the performancesof a bit being origins :
used to drill development wells on the - type of fluid outlet (nozzle,crowfoot,
Villeperdue field. This field lies in the orientation,inclination)
ParisianBasin, and is operatedby TOTAL in - the evolution of the flowing channel
associationwith TRITON FRANCE, cross-section
- the shape of the cross-sections.
The wells are Hslim holesw, a 7!!casing is In fact, the fluid distribution that we
set at about 200m TVD, the 611phase hits give, is the result of an interpretation
the target at about 2oOOm TVD. The wells and not a Wscientificimage!’such as that
have a ‘JH or “catenary” profile. The which could be provided by a laser
formations crossed are mainly marls and velocimetry.
limestones,with some thin beds which may
containanhydritesand/or dolomites. Consequencesof the lUKlrL-&@
The purpose of hydraulics optimizationis
The first wells were drilled with a 6’1bit, to drill faster (better R.O.P.) and for
like the one shown in fig 15.a). Following longer (prolonged bit life). The reasons
tka hydraulic study, a certain number of for the long life of P.D.C. bits are fairly
modifications were . recommended. These well known : the chips of rock must be
concerned both the profile of channel evacuated,and the cutting element must be
cross-sectionsand the fluid outlets. The properly cooled. Consequently, the
final design of the bit is shown in objectivewill be to avoid counter-currents
fig.15b. or low-circulationzones in the vicinity of
Table 1 shows the performancesof the two the cutting edges. The problem to solve is
bits. As can be seen, there is a not directly related to the type of
considerableincreasein the average R.O.P. formationdrilled, which only plays a part
(+25%) and an outstanding lengthening of in the over-heating process.
the life of the bit (>100%).
In the case of the R.O.P. the problem is
more difficultto define : it is reasonable
1~~ to assume”thatthe structureof the flow of
a bit, designed to drill sticky olays at
These are of two types : 30m/H and that of the same bit designed to
- Definition of the overall structure of drill compact shales at 10m/H will not be
the flow, due to the extreme complexityof the same. In the latter case the channel
the phenomenon cross-sectionscan be reduced so as to have
- Consequencesof the hydraulicson ROP and high flow velocities, the cuttings being
bit life. smaller and friable,whereas there would be
a risk of balling up in sticky shales.What
ion of the overw stmre of the is the best geometry to satisfy bot}icases
~ at the same time? Is it best to have very
test carried out investigates a turbulent structures which “churn up~~the
specific domain of flow, and provides a cuttings, or on the contrary, laminar-type
certain number of data. These, taken flows. These are the questions we are
separately, do not provide a complete trying to answer.
understanding of the flow. Extensive
interpretation work is then neceseary, ruTURB DEVBLORMMIX
which should give a definition of the
overall structure satisfying all the Numerical calculation should allow for a
resultstogether. larger number of simulations, whilst
keeping the number of experimentsdown .
The macro-structuresof the flow are fairly Once a profile has been defined which
well known and located by IIpaintl’ and optimizes the hydraulics according to
IIparticletl tests ; however, this is much precise criteria (made possibleby powerful
more difficult for the micro-structures, computers)an experimentalphase will still
and they can well have an important be n?c:ssary for a final validationof the
influenceon the former. calculations.
The laser velocimetrywill make it possible
Very frequently, the origin of the to investigate a complete flow without
hydraulic structures is difficult to disturbingit ; the final image of the flow
determine, and if a bit profilethiastto be we obtain with this measuringprocedure,is
optimized, it is necessary this an unquestionable representationof the
structurebe known, so as to evaluate the flow : the characteristicflow parameters
probable changes in flow following our are measured.
suggestedmodifications. The future evolution of our research work
will be orientedin these two directions.

233
 . .

6 .. --.-.-—-
HVDSAULIC -----.--.--——.
OPTIMIZATION ..—
OF P.D.C.
. . BITS BY VISUALIZATIONMETNODS SPE 18370

The authors thank TOTAL CFP for having

permitted them to publish this paper, and
ARTEP for having authorizedthe publication
of a graphics printout,and Deborah Keramsi
for help in writingthe article.

Table1
DRILLING BIT PERFORMANCES
,
oiwrw. oEslaN NEWCESIGN
IMPROVEMENT
Ug.las H&m
—

NUMBER OF BITS 4 8
Average footage
per bit (d
621 1340 >100 %
. ... .. .... ... . .. .... . .. . .. . .... .. ... .,, ,,, ,,. .. ... . .. ... ... . .. . .. . ... ...

Max. footage 766 2615

Average ROP
per bit (m/H)
9.41 11.73 25 %
... .... ... .. .. .. . ... .. .... ... ... ... .... ... .... .. .. ... ... ... .... .... .... .. ... .. . ... ... . .. . ...

Max. ROP 9,66 15,15

.

SPE 18370

Fig. l-HydrmIllc teat bonoh (general VIOW).

L (
3-Hydfaullc
Fig. !OOp.

,-J h4d9rm

Fig,2-ToMbonoh 33s
(flow shoot).
 ●

WE 18370

m 0

8) DEll

Fig. S-Bit @llw (a) flohtall tyfm

and (b)blada tyPoi

*. 7--vl@w
a 8 m&m-ww9.
Me
 .
sFE 18370

fl@ O—&nxmlVtw-lmwl
pointXl@Voltw.

Fig. O-vhw M ● -Voftox.

lllde/h”,tbm 11..

.,— .—

Fig. 10-CroaS a~ctlon of ● flow.

FIo. 11-WOW ol M Impxot point.

 .

sPE 18Y7°

m-m

b
/
/
I
I
I
I
I
Fig, 12-lnton@ty, dlrootlon of velooity veotore-@ counter current,

Fig, 13-Overall dlstrlbutlon of flow.

. .

r---m

b) II
a)

Fig, 14-Comparlaon of raaulta: (a) hydraulio toot bormh and (b) rwmortcal ●nalyaia
(CRAY XMP).

Fig, 16-Taat benoh applloatlon: (a) original doolgn and (b) now doslgn attcr mGdlfloation,