SPESociety of Petroleum Engineers

SPE 14219

Rock Breakage Mechanisms With a PDC Cutterby D.H. Zeuch and J.T. Finger, Sandia Natl. Laboratories

This paper was prepared for presentation at the 60th Annual Technical Conference and Exhibition of the Society of Petroleum Engineers held in LasVegas, NV September 22-25, 1985.

This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by theauthor(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by theauthor(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Paperspresented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Permission to copy Isrestricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment of whereand by whom the paper is presented. Write Publications Manager, SPE, P.O. Box 833836, Richardson, TX 75083-3836. Telex, 730989 SPEDAL.

ABSTRACT

Some aspects of chip generation by apolycrystalline diamond compact (PDC)cutter moving through a rock can beunderstood by examining the shapes of thechips and the fracture patterns in theremaining rock. Data from laboratory ex­periments have led to general conclusionsabout the uniformity of chip generationmechanisms in different kinds of rock andabout crack nucleation position relativeto the cutter tip.

INTRODUCTION

Sandia National Laboratories becameinvolved in drilling research as part ofour management of U.S. DOE's GeothermalDrilling and Completion Program. Much ofour effort has been directed to poly­crystalline diamond compact (PDC) bitsbecause their one-piece construction.which eliminates the need for high tem­perature bearings. seals. and lubricants.is desirable for geothermal drilling.Sandia activities have been both experi­mental and analytical. ranging from fullscale field tests of experimental bits toderivation of mathematical models for themechanics and heat transfer of singlecutters moving through rock. Several ofthese projects have been described in pre­vious SPE papers.

The work described in this paper hasfocused on trying to better understand thestress field in the rock around the

This work was supported by the U. S.Department of Energy at Sandia NationalLaboratories under Contract DE-AC04­76DP00789.

cutter. the mechanism of chip generation.and the resulting damage patterns in therock. An understanding of the mechanismof chip formation by a single cutter caneventually lead to improvement in theability to design tools. There have beenmany attempts to develop numerical modelsof the chip formation process [e.g .•1-5]. However. there has been less effortdirected at obtaining experimental obser­vations of the cutting process (other thanexperimental determination of forces oncutters) that might be used to test themodels. Such observations might inclUdedirect examination of the cutting (e.g .•by high speed process photography). chipmorphology and characteristics. particlesize distributions. and damage induced inthe rock by cutting. There have been fewsuch studies to date [6-12]. In anearlier study [12]. good agreement wasdemonstrated between experimentally ob­served cutter forces and induced damage inthe rock. and those predicted by a finiteelement model [5]. This led to our con­clusion that the model showed definitepromise in representing the mechanics ofcut ting. In this report. we extend ourobservations to inclUde chip morphology.particle size distribution of thecuttings. and fracture surfaces on thechips. The results of these examinationshave shown that there is a remarkablesimilarity in chip morphology among verydifferent kinds of rocks and that theexperimental evidence is consistent with atheory of chip generation in which cracksare nucleated well in advance of thecutter tip.

Analysis of the forces on the cutterand the particle size distribution of thecuttings has also clarified some of thedifferences between cuts in virgin rock

2 ROCK BREAKAGE MECHANISMS WITH A PDC CUTTER SPE 14219

cutters were viewed perpendicular (seeFig. 4 for orientation) to the directionof cut, the damage patterns were againvery similar, although viewed parallel tothe cut they were somewhat different. Thegroove made by the worn cutter hadconsiderable damage extending laterallyoutside the width of the cutter, [12]whereas the groove of the sharp cutter didnot. This is probably a product of thegreater vertical force with a worn cutterand the stress concentrations of the sharpcorners at the sides of the wear flat.

Another set of tests, in which thecutter stopped before making a completetraverse of the sample, also showed goodsimilar i ty in the ledges lef tat the endsof the cuts (Fig. 4). The ledges were allconcave downward, and in some cases hadfractures forward of the ledge that wasleft at the last position of the cuttertip. (Fig. 6 and 7)

This evidence, combined with a chipmorphology in Which a typical Chip profile(Fig. 8) is concave downward and quitedissimilar to the conventional representa­tion of a chip being formed (Fig. 9),lends support to a theory of chip genera­tion that does not stipulate crack initia­tion at the cutter tip.

Many of the existing models ofdrag-bit cutting suggest that the cutteracts dominantly as a rigid indentorpassing through the rock, and that chipformation is due to tensile stresses gen­erated at the leading edge of the cutter[e. g., 9]. Butcher and Zeuch [14] pointedout that the superposition of a shearingforce on a rigid indentor tends to put thematerial in compression in advance of theleading edge of the indentor and intension at the trailing edge. In fact,very high forces are observed on thecutter, and hence on the rock, in direc­tions both normal to the surface of thecut, and parallel to the cut. The skewedstate of compression could actually tendto suppress fracture in advance of thecutter While enhancing it at the trailingedge as suggested by work on indentationof glasses [15-19]. Butcher and Zeuchsuggested that the initiation of the frac­ture leading to formation of a large chipmust occur in advance of the compressedzone. The fracture then propagates towardthe free surface, simultaneously relievingthe compressive stress at the leading edgeof the cutter and allowing the completechip to form. Although the precise mecha­nism remains to be determined, neither thesubsurface damage we observe [12], nor thechip and ledge morphology, appear to beconsistent with a simple indentation model.

Chip morphology and size distributionwere evaluated for Tennessee Marble,vermont Marble, and Westerly Granite.Shapes were virtually identical for all

rock types, although the more brittle andfiner grained granite was consistentlybiased toward finer particle dimensions.The similarity in shape indicated that thesame chip formation mechanism was opera­ting for the two very different rocktypes. This similarity also extends tochips formed by different cutters-­Friedman [10] used a coal-mining cutterqui te unlike the PDC that we used (Fig.10). and produced chip shapes essentiallyidentical to ours. Furthermore, using acutter similar to that employed byFriedman [10]. FiSh [7, Fig. 11] also ob­served chips and ledges virtually identi­cal to ours; that is, concave downward.In view of this striking similarity amongthree different studies, we are surprisedthat most models for chip formation de­scribe the process as a "tensile arc,"leading to a concave upward ledge and chip(Fig. 9). Clearly, further work is re­quired to resolve this discrepancy. Theremarkable consistency of the particleshape indicated that more might be learnedabout the stress distributions andprocesses leading to Chip formation byadditional study of the fragments andfracture surfaces, as in fractographicstudies of more conventional engineeringmaterials [20].

Examination of the fracture surfacesshowed that virtually all of the largerchips have an apparently crushed, powderyregion at the trailing edge of the frag­ment (Fig. 11) in addition to the charac­teristic, flat shape. This crushed zonecould have originated in several ways.!"irst, it could have occurred by scrapingof the chip as it overrode the lower frac­ture surface. This seems improbable, asthe upper surface shows no evidence ofcrushing, Which might be expected if thechip was held down with sufficient forceto grind the underside of the fragment.Second, fines produced during the cuttingprocess could have been forced under thechip prior to detachment. This,too,seems unlikely as some of the zones areclearly fractured and friable, rather thanmerely covered wi th powder. Nor is thereany petrographic evidence from the manytests which we have carried out to indi­cate that fine particles are forced intofractu~es. Such a mechanism has been pro­posed to enhance the lifting action re­quired to remove a large chip, but evi­dence for this is absent and it seems un­likely with a cutter of negative rakeangle, such as our PDC.

Finally, and most probably, it seemsthat this crushed zone might represent theregion of intense triaxial compression andperhaps crushing that was suggested byButcher and Zeuch. This would then be a"process zone" at the tip of which thefracture ultimately leading to the insta­bility nucleates. This process zone hasalso been observed in a different con­text. Toward the conclusion of a test,

SPE 14219 D. H. ZEUCH & J. T. FiNGER

EXPERIMENTAL RESULTS

After the experimental cuts were made,the rock samples were vacuum-impregnatedwith either a blue-dyed epoxy or a fluore­scent dye. The samples were then sliced,both perpendicular and parallel to theexperimental cuts (Fig. 4), to give thinsections and polished bulk sections of therock around the cutter path. The dyeshighlighted the cracks left in the rock bythe cutter's passage.

The chips and fines from the experi­mental cuts were used for particle sizedistribution analysis and for examinationof the fracture surfaces on the largechips. Size distribution was measured bypassing the cuttings through nested sievesof progressively finer mesh and weighingthe particles in each compartment. Thelarge chips from each cut were examinedmicroscopically to evaluate the natUre offractures forming the surfaces. Some ofthe large chips were encapsulated inplastic and sawn in half parallel to thedirection of the experimental cut. so thatan accurate chip profile could be seen.

Examinations of the cut rock were partof the effort to deduce the stress fieldthat produced the chips. Thin section andpolished bulk sections of the cutter pathin the rock sample showed fracturesdipping into the rock in the direction ofthe cutter travel. These fractures werealso generally spaced at intervals con­sistent with the chip size and forcepeaks. We compared the fracture patternsin three kinds of rock---Tennessee Marble,Vermont Marble, and Westerly Granite[12]--and found that they were verysimilar to each other and to the fracturespredicted in the the two-dimensionalfinite element model developed by Swenson[5]. When the cuts made by sharp and worn

period­cutting

Generally the force data have beenconsistent with a chip generation modelthat requires the force to be periodic ­first increasing as the stresses aredeveloped that will form cracks and detacha chip, and then falling suddenly to alower value as the cutter enters the spaceformerly occupied by the chip. By usingthe speed of the cutter, the period of theforce variations can be converted to adistance between force peaks in the rock.This distance--ranging from 0.150 to 0.250inches--is consistent with the dimensionsacross the large chips from each run.Other aspects of the cutter forces, suchas the increase in cutting force over thefirst one to two inches of the cut andthen a decrease over the next few inches(Fig. 5) are probably the result of avarying accumulation of fines packed underthe cutter wear flat.

change, and to observe the forceicity characteristic of thisprocess (Fig. 3).

Experimental Materials Three kinds ofrock - Westerly Granite, Tennessee Marble,and Vermont Marble - were used in thesetests. Petrographic descriptions andmechanical properties of the first tworock types are given elsewhere [13]. Theprecise origin of the third rock. de­scribed to us only as "Vermont Marble". isunknown. Petrographic observation revealsthat it is a very pure, highly recrystal­lized and annealed. white marble. Averagegrain size is approximately 500 ~m, andthe calcite grains show very littletwinning or other evidence of deforma­tion. In hand specimen, the color is uni­formly milky white, and it is massive inappearance, with no evidence of layeringor other inhomogeneities such as stylo­lites. The cutting speed was 45 inches/minute and the depth of different cutsvaried from 0.025 inch to 0.100 inch, withmost cuts at 0.050 inch depth. One seriesof tests in Tennessee Marble used a worncutter to make a pass through virgin rockand then, after widening the original cut,a second pass of the same depth over thetrack of the first cut (Fig. 2). Thesetests compared the forces required to cutthe virgin rock and the damaged rock.

Description of Data The forces on thePDC cutter were recorded in analog form onmagnetic tape. and then digi tized at onemillisecond intervals for statisticalanalysis. The analysis is used to compareaverage forces in different intervals of acut, to compare the variance of forces indifferent cuts. to identify trends inforce distribution as cutting conditions

After the rock samples were cut on themi lling machine, we sent them to a petro­graphic lab where they were first vacuumimpregnated with dye and then sawn intothin sections and polished bulk sections.

EXPERIMENTAL PROCEDURE

Test Apparatus The experimental rockcuts were made with a milling machine mod­ified for laboratory work. The rocksamples (approximately 2" x 4" x 6") wereheld in a vise mounted on the mill table,and this was used to traverse the rockpast the cutter. Cuts were made with bothsharp and worn cutters (Fig. 1), eachhaving a 0.522 inch diameter compact witha negative 200 rake angle. A three-axisdynamometer, which held the cutter stud,was used to measure the cutting forces,which were recorded on magnetic tape. Thechips from each cut were collected in acardboard shield around the rock sample;after cutting, the shield and the top ofthe sample were brushed carefully to sweepthe chips into a plastic bag.

and in surfaces damaged by previous cutterpasses. The balance of this paper willdescribe the experimental procedure andrelate the experimental results to a modelof chip generation.

ROCK BREAKAGE MECHANISMS WITH A POC CUTTER SPE 14219

REFERENCES

3. Fractures are nucleated in therock in advance of the cutter tip.

surfaces. and sUb-surface damage. is verysimilar for different kinds of rocks andfor different cutter geometries.

[4] Ford. L. M.• C. E. Longfellow and M.Friedman.. 1983. The Interaction of aCoal Mining Tool wi thRock: Theory.Experiment. and Analysis Report No.SAND83--0128. Sandia NationalLaboratories. Albuquerque. NM. 55 pp.

1973. Fractureto Rock. Ph.D

University of

Hardy. M. P .•Mechanics AppliedDissertation.Minnesota. 231 pp.

[3 ]

5. Since all of these tests were doneat atmospheric pressure. the effects ofborehole and confining pressures areclearly of interest. A series of singlecutter tests in a pressurized test chamberwill be done soon. and the chipmorphology. fracture surface. andsubsurface damage will be compared to theresults presented here.

[1] Sikarskie. D. L.• and N. J. Altiero.1973. The Formation of Chips in thePenetration of Elastic-BrittleMaterials (Rock). Journal of AppliedMechanics. 40. 791-797.

[2] Wang. J. K.. and T. F. Lehnoff.1976. Bit Penetration Into Rock--AFinite Element Study . Int.erna tiona 1Journal of Rock Mechanics and MiningSciences. 11. 11-16.

4. The ledge in the rock ahead of thecutter is consistent in shape. The lattertwo conclusions are important becauseaugmentation of PDC cutters by water jetsdirected in front of the cutters is aneffective technique. The geometry of theremaining ledge and the location of cracksahead of the cutter should be wellunderstood because they will affect thewater jet placement.

2. Even though the mechanism of chipgeneration appears to be similar betweendifferent cutters and between differentrock condition. the efficiency of thecutting process may change significantly.Cuts in pre-damaged rock remove more chipsper unit work than in virgin rock and cutswith sharp cutters obviously require lessforce than with dull cutters.

Sieve size analysis of the cuttingsshowed that. not surprisingly. the remainsof the second cuts were finer than thoseof the first cuts (Table 1). The averageforces were also slightly larger for thesecond cuts. but so were the masses ofrock removed (Table 2). It appears thatat least four phenomena are in play here.two which tend to make the forces higherand two which tend to make them lower.Removal of a greater mass of rock withsmaller average dimensions would create aconsiderably greater free surface for thefragments of the second runs. and pre­existing cracks would prevent new frac­tures from extending as far as they wouldin virgin rock. thus requiring more nucle­ations per length (or area) of fracture.Both of these factors would increase theforces required for the second cuts. butthe prior cracks would act to reduceforces by providing some of the free sur­face for the new chips and possibly byrelieving the compressive region at thecutter tip. thus easing new crack nuclea­tions. The latter mechanisms seem to bedominant. since the index of rock removalis higher for the second runs. and theincrease in efficiency from first tosecond runs is consistent (Table 2).

The last aspect of the experimentalwork was a comparison of cutting behaviorin virgin rock samples with cutting be­havior in a rock that was already dam­aged. The purpose of this was twofold:to see whether the chips and forcepatterns were similar in the two cases.and to compare the efficiencies of the twoprocesses. As shown in figure 2. thefirst and second cuts were the same depthrelative to the starting surface; beforethe second cut was made the first cut waswidened to relieve the confining effectsof the adjacent surface.

i.e .• at the end of a cut. a single. largefragment is removed. On the fracture sur­faces of several of these chips. a similarcrushed. powdery zone is observed (Fig.12). At the least. this must representthe zone in which the fracture has becomeunstable and begun to accelerate to itssteady state velocity. similar toinferences made from observations offracture surfaces in glasses. Whether ornot it can be demonstrated that thefracture is the result of triaxialcompression (i.e .• extension fractures) oris truly tensile in character. is notclear. but intense compression andshearing would explain the powdery. oftencrushed appearance of these zones.

CONCLUSIONS AND RECOMMENDATIONS

The exper imenta 1this work has ledconclusions:

[5] Swenson. D. V.. 1983. Modeling andAnalysis of Drag Bit Cutting. ReportNo. SAND83-0278. Sandia NationalLaboratories. Albuquerque. NM 87 pp.

as

evidence examined inus to the following

1. The mechanism of chip generation.shown by chip morphology. fracture

[6 ] Fish. B.Variables

G.. 1961. Thein Rotary Drilling.

BasicPart

SPE 14219 D. H. ZEueH & J. T. FINGER 5

[7] Fish, B. G., 1961, The BasicVariables in Rotary Drilling, Part 2.Mine & Quarry, February issue, 74-Bl.

[B] Gray, K. E., F. Armstrong and C.Gatlin, 1962. Two-Dimensional Studyof Rock Breakage in Drag Bit Drillingat Atmospheric Pressure. Journal ofPetroleum Technology, January issue,93-9B.

Wilshaw, 1973,Principles andMater. Sci.,

Partial Cone CrackBrittle Material

Sliding SphericalSoc. Lond. Ser. A,

Lawn, B. R., 1967,Formation in aLoaded with aIndenter, Proc. R.299:3077-316.

MCClintock. F. A., and A. S. Argon,1966, Mechanical Behavior ofMaterials, Addison-Wesley, Inc .•Reading, Massachusetts, 170 pp.

[16] Hamilton, G. M., and L. E. Goodman,1966, The Stress Field Created by acircular, Sliding Contact J. Appl.Mech., 33:371-376.

Proceedings (A. C. Stickland, ed.),pp. 253-25B, The Institute of Physicsand the Physical Society, London.

[17] Johnson, K. L., J. J. O'Connor and A.C. Woodward, 1973, The Effect of theIndenter Elasticity on the HertzianFrac ture of Br i t tIe Ma ter ia Is, Proc.R. Soc. Lond. Ser. A, 334:95-117.

[lBJ

[19] Lawn, B. , and R.Indentation Fracture:Applications., J.10:1049-10Bl.

[20]

January issue,1. Mine & Quarry,29-34.

Friedman, M., and L. M. Ford, 19B3,Analysis of Rock Deformation andFractures Induced by Rock CuttingTools Used in Coal Mining. In: RockMechanics: TheorY-Experiment­Practice. Proceedings of The 24thU.S. Symposium on Rock Mechanics,Texas A&M University, CollegeStation, Texas, June 20-23, 19B3, pp.713-723. The Association ofEngineering Geologists, Short Hills,New Jersey, B5B pp.

Hood, M., 1977, A Study of Methods toImprove the Performance of Drag Bi tsUsed to Cut Hard Rock, Chamber ofMines of South Africa Report No.35/77, Project No. GT2 N02, 135pp.

[10] Friedman, M., 19B3, Analysis of RockDeformation and Fracture Induced byRock Cutting Tools Used in CoalCutting, Report No. SAND83-7007,Sandia National Laboratories,Albuquerque, NM, 39 pp.

[9 ]

[ 11]

[12] Zeuch, D. H., D. V. Swenson and J. T.Finger, 19B3, Subsurface DamageDevelopment in Rock Dur ing Drag-bi tCutting Observations and ModelPredictions. In: Rock Mechanics:Theory-Experiment-Practice. Proceed­ings of The 24th U.S. Symposium onRock Mechanics, Texas A&M university,College Station, Texas, June 20-23,1983. pp. 733-742. The Associationof Engineering Geologists, ShortHills, New Jersey, 85B pp.

[13] Krech, W. W., F. A. Henderson and K.E. Hjelmstad, 1924, A Standard RockSui te for Rapid Excavation Research,U.S. Bureau of Mines Report No. 7B65,Washington, DC 29 pp.

[14] Butcher, B. M. , and D. H. Zeuch,19B3, Interaction of a Coal MiningTool with Rock, Sandia NationalLaboratories Technical Memorandum toJ. R. Kelsey (6241), dated Oct. IB,19B3. (Unpublished)

[15] Bi11inghurst, P. R., C. A. Brookesand D. Tabor, 1966, The SlidingProcess as a Fracture InducingMechanism. In: Physical Basis ofYield and Fracture, Conference

Table 1

Percent of total cuttings passingthrough a given mesh size

(average of four runs)

Mesh Size. lJ.m First Run Second run

4000 77.6 80.72000 54.7 57.71000 35.8 37.9

500 24.6 25.7250 15.0 15.5125 8.4 9.0

38 2.2 2.6

Table 2

Ratio ofPairs Average Weight of Specific Rock Second Run

of Runs Force. lb. cuttings. gm Removal. mci/J To First

1a 456 3.228 10.45Ib 530 3.432 9.56 .92

2a 480 3.665 11. 282b 459 4.155 13.36 1. 20

3a 510 3.819 11.043b 528 4.648 J.2.98 1.17

4a 473 3.239 10.094b 543 4.544 12.35 1. 23

SHARP

A. '--- -Y

SINTEREDDIAMOND

LAYER~

TUN GSTEN ""","¥'-L---I____

CARBIDESUBSTRATE

WORN

Fig. 1-Slde view of sharp and worn poe cullers.

TUNGSTENCARBIDESTUD

C. L-.. --V

Fig. 2-Sequence of overlaid cuts in Tennessee marble.

£!·600

w 400~ 200o 0u.

£!W 600(,) 400ll: 200o 0u.

o 2

TIME, seconds

Fig. 3~Vertlcal (bottom) and horizontal (top) forces on the cutter.

LEDGE

3

Flg.4-0rientation of thin and·pollshed bulk secllons.

800I I

7001- ...

'" ...'"600 - ...

•'" •D •£! 500 - D -

DW •(,)a:.0u. 400 - D -

300

'" FIRST RUNS - V200 I- D FIRST RUNS - H -

... SECOND RUNS - V

• SECOND RUNS - H

100 I0 2 4 6 8

TIME, secondsFig. 5-Average cutterforces·over successive 2·second Intervals.

Fig. a-Profile of typical chips (arrows show cut direction).

Fig. 12-Crushed zone on terminal chip.