Scott F. Miller

Rui Li

Mechanical Engineering,University of Michigan,

Ann Arbor, MI 48109

Hsin WangOak Ridge National Laboratory,

Oak Ridge, TN 37831

Albert J. ShihMechanical Engineering,

University of Michigan,Ann Arbor, MI 48109

Experimental and NumericalAnalysis of the Friction DrillingProcessFriction drilling is a nontraditional hole-making process. A rotating conical tool is ap-plied to penetrate a hole and create a bushing in a single step without generating chips.Friction drilling relies on the heat generated from the frictional force between the tooland sheet metal workpiece to soften, penetrate, and deform the work-material into abushing shape. The mechanical and thermal aspects of friction drilling are studied in thisresearch. Under the constant tool feed rate, the experimentally measured thrust force andtorque were analyzed. An infrared camera is applied to measure the temperature of thetool and workpiece. Two models are developed for friction drilling. One is the thermalfinite element model to predict the distance of tool travel before the workpiece reaches the250°C threshold temperature that is detectable by an infrared camera. Another is a forcemodel to predict the thrust force and torque in friction drilling based on the measuredtemperature, material properties, and estimated area of contact. The results of this studyare used to identify research needs and build the foundation for future friction drillingprocess optimization. �DOI: 10.1115/1.2193554�

1 IntroductionFriction drilling, also known as thermal drilling, flow drilling,

form drilling, or friction stir drilling, is a nontraditional hole-making method. The heat generated from friction between a ro-tating conical tool and the workpiece is used to soften the work-material and penetrate a hole �1,2�. It forms a bushing directlyfrom the sheet metal workpiece and is a clean, chipless process.

Figure 1 shows schematic illustrations of the five steps in fric-tion drilling. The tip of the conical tool approaches and contactsthe workpiece, as shown in Fig. 1�a�. The tool tip, like the webcenter in twist drill, indents into the workpiece and supports thedrill in both the radial and axial directions. The friction force onthe contact surface produces heat and softens the work-material.The tool is then extruded into the workpiece, as shown in Fig.1�b�, pushes the softened work-material sideward, and piercesthrough the workpiece, as shown in Fig. 1�c�. Once the tool tippenetrates the workpiece, as shown in Fig. 1�d�, the tool movesfurther forward to push aside more work-material and form thebushing using the cylindrical part of the tool. The shoulder of thetool may contact with the workpiece to trim or collar the extrudedburr on the bushing. Finally, the tool retracts and leaves a holewith a bushing on the workpiece �Fig. 1�e��. The thickness of thebushing is usually two to three times as thick as the original work-piece. This leaves enough surface area for threading.

All work-material in the friction drilled hole contributes to formthe bushing. It eliminates chip generation and is a clean, chiplesshole-forming process. Unlike the traditional drilling operation us-ing cutting fluid to reduce friction and heat generation, frictiondrilling is a dry process. Occasionally, a small amount of cuttingfluid is used to avoid material transfer from the workpiece to thetool.

Publications on the subject of friction drilling are limited. Sixpatents have been awarded: four to van Geffen �1–4� in 1976–1980 and later one to Head et al. �5� and one to Hoogenboom �6�in 1984. France et al. �7–9� investigated the strength characteris-tics of friction drilled holes in metal tubing. Overy �10� and Bak

Contributed by the Manufacturing Engineering Division of ASME for publicationin the JOURNAL OF MANUFACTURING SCIENCE AND ENGINEERING. Manuscript receivedAugust 17, 2004; final manuscript received January 6, 2006. Review conducted by S.

S. Schmid.

802 / Vol. 128, AUGUST 2006 Copyright © 2

�11� discussed the design aspect of the friction drilled holes.Kerkhofs et al. �12� studied the performance of coated frictiondrilling tools. The research on the mechanics of the friction drill-ing process, particularly the measurement and modeling of thrustforce, torque, and temperature of the tool and workpiece, is lack-ing and has become the goal of this research.

In this study, a noncontact infrared �IR� camera is applied tomeasure the tool and workpiece temperatures during friction drill-ing. The InSb-based IR camera setup used in this study can onlydetect thermal radiation above 250°C. A thermal finite elementmodel is applied to predict the distance of tool travel to reach thethreshold temperature after the initial contact. This enables theprediction of workpiece temperature and material properties in theforce model. The force model is developed to predict the thrustforce and torque generated during the friction drilling processbased on the pressure and area of contact between the tool andworkpiece. The high temperature in friction drilling changes ma-terial properties. The force model applies the experimentally mea-sured temperature to predict the yield stress and contact pressurebetween the tool and workpiece and to calculate the thrust forceand torque.

The experimental setup and process parameters for frictiondrilling are first introduced in Sec. 2. In Sec. 3, the measuredthrust force and torque are analyzed. The infrared camera for tem-perature measurement and its calibration are discussed in Sec. 4.The thermal finite element modeling is presented in Sec. 5 topredict the tool travel distance for the work-material to reach the250°C threshold temperature for infrared temperature measure-ment. The force model is discussed in Sec. 6. Modeling resultsand comparison with experimental measurements are presented inSec. 7.

2 Experimental Setup

2.1 Machine and Workpiece. A three-axis computer numeri-cal controlled vertical machining center, Milacron model Sabrewith a 7.5 kW spindle, is used for the friction drilling experiment.The overview of the friction drill test setup is shown in Fig. 2�a�.As shown in the close-up view in Fig. 2�b�, the drill is fixed to aspecially designed tool holder provided by Unimex Formdrill.

The workpiece is a 1.19 mm thick AISI 1020 cold-rolled car-

bon steel sheet. As shown in Fig. 2, the workpiece is held on top

006 by ASME Transactions of the ASME

of a Kistler model 9272A piezoelectric drilling dynamometer. Thethrust force and torque during drilling are measured. The experi-ment is carried out at 4000 rpm spindle speed and 165 mm/minconstant tool feed speed.

2.2 Tool. The tool is made of WC in a Co matrix. Key geo-metrical features of the drill are shown in Fig. 3. The drill consistsof five regions:

�1� Center region: The cone-shape center has the angle � andheight hc. The angle is usually blunt. The effect of bluntingis to generate more force and, therefore, heat at the start ofthe drilling. The center region, like the web in a twist drill,provides the support in the radial direction for the frictiondrilling process and keeps the tool from walking at the startof the process.

Fig. 1 Illustration of stages in the friction drilling process

�2� Conical region: This region has a sharper angle than the

Journal of Manufacturing Science and Engineering

center region. The drill in this region rubs against the work-piece to generate the friction force and heat and pushes thework-material sideward to shape the bushing. The angleand length of the cone-shape conical region are marked as� and hn, respectively.

�3� Cylindrical region: This region helps to form the hole andshape of the bushing. The length and diameter of this re-gion are designated as hl and d, respectively.

�4� Shoulder region: The shoulder of this region may touch theworkpiece to round the entry edge of the hole and bushing.

�5� Shank region: This is the area of the tool gripped by thetool holder of the machine.

The first three regions �center, conical, and cylindrical� deter-mine the thrust force and torque as well as the tool and workpiecetemperatures during friction drilling. The tool geometry is impor-tant to the shape of the bushing and process performance. Dimen-sions of the friction drill used in this study has d=7.3 mm, �=90 deg, �=36 deg, hc=0.970 mm, hn=8.490 mm, and hl=8.896 mm.

3 Thrust Force and Torque in Friction DrillingThe measured thrust force and torque in friction drilling AISI

1020 steel are shown in Figs. 4�a� and 4�b�, respectively. Underthe constant tool feed rate, a peak thrust force of 700 N occurs at2.5 mm of tool travel from the initial contact with the workpiece.This position of maximum thrust force is marked as A. A separatefriction drilling test was conducted to stop and move back the toolat position A. The workpiece was sectioned and polished to revealthe deformation shape, as shown in Fig. 5�a�. The sheet metal

Fig. 2 Experimental setup in friction drilling: „a… overview and„b… close-up view of the spinning tool contacting the workpiece

workpiece was bent and dented but not perforated. This reveals

AUGUST 2006, Vol. 128 / 803

that the mechanical indentation of the workpiece by the tool tip isthe key deformation mechanism at the start of friction drilling.Insufficient heat was generated to soften the work-material due tothe low tool peripheral speed at the tip. No discoloration at theindentation of the workpiece was observed to confirm increasedworkpiece temperature. Low peripheral speed also accounted forthe small torque at the beginning of the contact. The torque rap-idly rose to 1.5 N m after the tool reached position A.

The thrust force drops rapidly to 300 N, less than half of thepeak value, after the tool moved by another 2 mm to position B,4.5 mm tool travel from contact. Another separate friction drillingtest was performed to stop the tool at position B. The deformationof the workpiece is shown in Fig. 5�b�. The workpiece is on theverge of penetration by the tool tip. It indicates that, like themechanics in conventional drilling �13�, the center of the drillcontributes a majority of the thrust force. Change in color wasobserved inside the hole due to the increased temperature. Thetorque increased to about 1.7 N m at position B.

In the next 5.3 mm, the tool penetrates to position C �9.8 mmtool travel from the start of contact� close to the maximum torquecondition. From position B to C, the thrust force remains at about

Fig. 3 Key dimensions of the friction drilling tool

Fig. 4 Thrust force and torque in friction drilling of AISI 1020carbon steel sheet

300 N and the torque gradually increases to the maximum at

804 / Vol. 128, AUGUST 2006

2 N m. The deformation of the workpiece at tool position C isshown in Fig. 5�c�. The conical region of the tool has perforatedthe workpiece and pushed the work-material aside to form a bush-ing. Rings of discoloration were observed inside and outside the

Fig. 5 Cross-sectional view of deformation of the workpiece infriction drilling at positions „a… A, „b… B, „c… C, and „d… D in Fig.4

hole, indicating the further increase of workpiece temperature.

Transactions of the ASME

Between tool position C and D, both the force and torque start todecrease. Position D at 12.7 mm, as shown in Fig. 4, marks theend of friction drilling. The complete hole in friction drilling isshown in Fig. 5�d�. After position D, the force is reduced to al-most zero �or even negative� but torque remains high, in the1.1–1.4 N-m range, while the tool retracts to leave the hole in theworkpiece.

In summary, Fig. 4 illustrates a typical pattern of thrust forceand torque for friction drilling at a constant tool feed rate. Par-tially filtered noise is present throughout the graphs in Fig. 4. Thehigh peak force is not desirable since it deforms the sheet work-piece and may shorten the tool life.

4 Workpiece Temperature MeasurementA Raytheon Radiance IR camera system was used to measure

the tool and workpiece temperature in friction drilling. The cam-era system was set up to capture temperature from a side view ofdrilling. The IR camera has an InSb focal plane array detectorwith a 256�256 pixels focal plane array. It is sensitive to thermalradiation wavelength from 3 to 5 �m. The IR image capturingwas set at 60 frames/s with 0.1 ms integration time. The ND2filter was used. The minimum detectable temperature by the IR

Fig. 6 Infrared camera pictures of friction drilling: „ a… 2 s, „b…tool and workpiece

camera system was 250°C.

Journal of Manufacturing Science and Engineering

Figures 6�a�–6�f� show three IR camera images of friction drill-ing at positions E, F, and D �the end of tool forward movement�and three IR frames when the tool is retracting from the work-piece. Figures 6�e� and 6�f� show the heated tool exiting from theworkpiece. The intensity of a pixel in the image corresponds tothe tool and workpiece temperature. The brightest spot with thehighest intensity in Fig. 6�c� indicates the maximum temperatureof the workpiece.

Because the thermal emissivity of the workpiece and tool ma-terial are not known, calibration tests are required to find relation-ship between the IR intensity and a known temperature. In cali-bration tests, the tool and workpiece, with a thermocoupleattached, were first heated in an oven. During the cooling, the IRcamera was used to record the thermal radiation intensity of anobject �either the workpiece or tool� corresponding to the thermo-couple temperature reading. Two calibration curves relating the IRintensity and workpiece and tool temperature were generated.

Using the workpiece intensity versus temperature calibrationcurve, the maximum workpiece temperature in each IR framecould be obtained. Results are shown in Fig. 7. At the start, thereis a time interval, �t, needed for the workpiece to reach the

, „c… 5 s, „d… 8.5 s, „e… 10 s, and „f… 12 s from the contact of the

3 s

250°C threshold temperature detectable by the IR camera. The

AUGUST 2006, Vol. 128 / 805

tool travel during �t is designated as �. In the next section, a finiteelement thermal model is applied to predict �t and �.

The workpiece temperature gradually increases as the toolmoves through the workpiece and displaces the work-material. Asthe bushing is formed, the workpiece temperature reaches a maxi-mum value of 760°C at 3.2 s from the initial contact. The work-piece temperature then gradually decreases. The peak workpiecetemperature is above the crystallization temperature, between500 to 700°C, of AISI 1020 steel. Metallurgical studies show thegrain size of the workpiece is refined near the friction drilled hole.

The tool temperature is difficult to measure since the workpiecematerial surrounds the tool during friction drilling. In Fig. 6�b�, asthe tool enters to the workpiece, the peak tool temperature is about580°C. In Fig. 6�f�, for the retracting tool, the tool is cooled andpeak temperature is estimated at 400°C. A more dedicated experi-mental setup to aim at the tool tip once it just penetrates theworkpiece is necessary to acquire a more accurate estimation oftool temperature.

In Sec. 6, the temperature versus distance of tool travel data isapplied to find the temperature-dependent yield stress of the work-material to predict the contact pressure for thrust force and torquemodeling in friction drilling.

5 Thermal Finite Element Modeling to Predict �t and�

The heat transfer at the start of the friction drilling process wasmodeled using the thermal finite element method to predict thetime �t and distance � of the tool travel for the workpiece to reach250°C. The ANSYS 7.0 finite element software package and itsmesh generator were applied. The finite element mesh, as shownin Fig. 8�a�, consists of the eight-node, eight degree-of-freedomaxisymmetric element to model a disk plate. The finite elementmesh includes 4073 nodes. The initial contact and heat generationdue to friction drilling occur at the top center of the cylindricaldisk. The model is semiempirical as the experimentally measuredtorque is used as an input for the heat flux. Deformation of theworkpiece during friction drilling is not considered. The thicknessof the plate is 1.2 mm, the same as the workpiece. The radius ofthe disk is 20 mm, large enough not to influence the temperatureresults near the center of the disk. Properties of the workpiece areassumed to be temperature-independent. The density is7870 kg/m3, the heat capacity is 486 J /Kg°C, and the thermalconductivity is 51.9 W/m K. The initial temperature of the work-piece and the surrounding temperature are 25°C. The convectioncoefficient is set at 40 W/m2 K.

Three time steps, designated as steps I, II, and III, are used tosimulate the heat transfer in friction drilling in the first 0.35 s. Asshown in Figs. 8�b�–8�d�, A1, A2, and A3 are the areas in steps I,II, and III, respectively. Tool wear is not considered, i.e., a per-

Fig. 7 Maximum workpiece temperature and the minimumtemperature detectable by the infrared camera

fectly sharp tool is used to find the tool-workpiece interface areas.

806 / Vol. 128, AUGUST 2006

Step I ranges from 0 to 0.118 s with 0.32 mm tool travel fromthe start of the tool contact. The interface between a perfectlyshaped tool and the workpiece at the 0.32 mm tool travel ismarked as A1 in Fig. 8�b�. A1 represents a conical surface of thetool tip. A heat flux is applied on surface A1 from 0 to 0.118 s.The heat flux, q, generated by friction is calculated by

q =2�Tn

60Aia �1�

where T is the experimentally measured torque, n is the rotationalspeed of the drill, Ai is the area of tool-workpiece interface, and a

Fig. 8 The axisymmetric finite element mesh and the thermalmodel at three time steps

is the fraction of frictional energy converted into heat.

Transactions of the ASME

In the study, n=4000 rpm and a=0.9. The ramp-type heat flux,i.e., linearly increasing from zero heat flux at the start of contactto the q at 0.118 s �end of step I�, is applied on the surface A1.The experimentally measured torque at 0.118 s is used to calcu-late q for step I.

Step II ranges from 0.118 to 0.235 s when the tool tip travelsfrom 0.32 to 0.65 mm. As shown in Fig. 8�c�, the interface sur-face between the tool tip and the workpiece at the 0.65 mm tooltravel is marked as A2. The step-type heat flux, i.e., uniform from0.118 to 0.235 s, is applied. The torque T at 0.235 s is used tocalculate q for step II.

Step III ranges from 0.235 to 0.35 s. The tool tip travels from0.65 to 0.97 mm into the workpiece. The step-type heat flux withtorque T at 0.35 s is applied.

Results of the maximum surface temperature of the workpieceversus time obtained from the thermal finite element modeling areshown in Fig. 9. The temperature gradually arises from25 to 500°C in about 1 s. The temperature reaches 250°C at �t=0.075 s and �=0.21 mm. The experimentally measured tempera-ture starting from 250°C is plotted in Fig. 9. Beyond 250°C, thetrend of temperature rising obtained from experiment and simula-tion match. This validates the finite element estimation of �t and�.

6 Thrust Force and Torque ModelingA model, based on the pressure and contact area between the

tool and workpiece, is established to predict the thrust force andtorque in friction drilling. Two elemental shapes used to model thecontact area, as shown in Fig. 10, are the tapered cylinder andstraight cylinder.

As shown in Fig. 10�a�, the shape of tapered cylinder is definedby three parameters: two heights, h1 and h2, and an angle, �. Thisangle � can be either � or �, depending on the center or conicalcontact region of the tool �Fig. 3�. A uniform pressure p, whichcan be estimated by the yield stress of the rigid-plastic work-material at a given temperature, is acting on the surface. Twocoefficients of friction, � and �a, are defined to calculate forces inthe tangential and axial directions. In the tangential direction, theworkpiece is sliding on the fast rotating tool surface at surfacespeed of 1530 mm/s without lubricant. As investigated by frictionstir welding �14�, this results in a relatively high �. In the axialdirection, the tool is penetrating the workpiece at very slow speed

Fig. 9 Comparison between the experimental and thermalmodeling temperature

of 4 mm/s. The �a is expected to be lower than �.

Journal of Manufacturing Science and Engineering

Equations for thrust force and toque in the tapered cylinder area�Fig. 10�a�� with an inclusion angle � can be derived as

F =�h1

h2

p sin�

2dA +�

h1

h2

�ap cos�

2dA

= �p�h22 − h1

2�tan2 �

2+ �ap��h2

2 − h12�tan

�

2�2�

T =�h1

h2

�prdA =

2��p�h23 − h1

3�tan2 �

2

3 cos�

2

�3�

In the straight cylinder area, the thrust force and torque are

F = 2��apRh3 �4�

T = 2��pR2h3 �5�Four assumptions are made to simplify the force model:

�1� The tool is perfectly sharp at the tip and all corners. Fiveparameters that define the perfect tool in Fig. 3 are hc, hn,hl, �, and �.

�2� The deformation of the workpiece is negligible.�3� The coefficients of friction are independent of temperature

and speed and do not change during the friction drillingprocess.

�4� No displaced work-material contributes to the force model-ing, i.e., only the overlapping area between the tool and

Fig. 10 Two basic areas for contact between the tool andworkpiece in friction drilling force modeling

Fig. 11 A band of discoloration in the drilled hole

AUGUST 2006, Vol. 128 / 807

Fig. 12 Six stages in friction drilling force modeling

808 / Vol. 128, AUGUST 2006

undeformed steel sheet is used in the force modeling. Thisassumption can be justified by the experimental observa-tions that, as shown in Fig. 11, a blue color strip inside thefriction drilled holes is always along the region where thetool contacts the original unformed workpiece. The bluecolor indicates high temperature and more active contact.The work-material, after being displaced by the rotatingtool, is hot. It deforms easily and does not create significantresistance to the deformation.

The contact between the double angle friction drilling toolshown in Fig. 3 and the undeformed workpiece sheet are dividedinto six stages, as shown in Fig. 12. Figure 12 shows the geo-metrical illustrations of different overlapping area between thetool and workpiece during the six stages of friction drilling. Thedistance of tool travel in the axial direction is designated as h.

Stage 1: Only the center region is contacting the workpiece. Asshown in Table 1, h1=0 and h2=h for the center region in stage 1.

Stage 2: Both the center and conical regions are contacting theworkpiece. In the center region, h1=0 and h2=hc, i.e., the wholecenter region remains inside the workpiece. For the tool andworkpiece used in this study, the thickness of the worpiece, t, islarger than the height of center region, hc. In the conical region,h1=hc+h* and h2=h+h*, where h* is an offsetting height, as il-lustrated in Fig. 13, to calculate heights h1 and h2 for the conicalregion. The h* can be derived as

h* = hc

tan��

2�

tan��

2� − hc �6�

Stage 3: The center region has penetrated through the work-piece and h1=h− t and h2=hc. In the conical region, h1=hc+h*

and h2=h+h*.Stage 4: Only the conical region with h1=h− t+h* and h2=h

+h* is in contact with the workpiece.

tool-workpiece contact in six stages

h2� Conical �h1, h2� Cylindrical �h3�

¯ ¯

hc+h*, h+h*¯

hc+h*, h+h*¯

h− t+h*, h+h*¯

h− t+h*, hc+hn+h* h−hc−hn¯ t

Fig. 13 Geometrical relationship to calculate h*

Table 1 Parametric representation of the

Stage Range of tool travel Center �h1,

1 0hhc 0, h2 hch t 0, hc3 th t+hc h− t, hc4 t+hchhc+hn

¯

5 hc+hnhhc+hn+ t ¯

6 hc+hn+ thhc+hn+hl¯

Transactions of the ASME

Stage 5: Both the conical and cylindrical regions are in contactwith the tool. In the conical region, h1=h− t+h* and h2=hc+hn+h*. In the cylindrical region, h3=h−hc−hn.

Stage 6: Only the cylindrical region is in contact with the work-piece and h3= t.

In summary, the input of the model includes: �1� tool geometry�hc, hn, hl, �, and ��; �2� workpiece geometry �t�; �3� work-material temperature-dependent stress-strain properties; �4� coef-ficients of friction �� and �a�; and �5� tool travel from contact �h�.The output of the model is the thrust force and torque. Effects ofthe strain-rate, deformation of the workpiece, and tool wear arenot considered.

7 An Example of Modeling Results and ComparisonWith Experimental Measurements

The friction drilling experiment conducted in Sec. 3 is modelednumerically. The measured maximum workpiece temperature inFig. 7 is used in the temperature-dependent stress-strain relations,obtained from �15�, to determine the yield stress at each point inthe model. The yield stress is used to represent the pressure p. Inthis study, � is set at 2.0, which is adopted from the value in

Table 2 Numerical example of the tool-w

Stage Range of tool travel Center �h1

1 0h0.97 0, h2 0.97h1.19 0, 0.973 1.19h2.160 h−1.19, 04 2.160h9.460 ¯

5 9.460h10.650 ¯

6 10.650h18.356 ¯

Fig. 14 Comparison of the experiment versus model predicted

thrust force and torque in friction drilling

Journal of Manufacturing Science and Engineering

friction stir welding �16�. The high coefficient of friction betweenthe fast rotating tool and the stationary workpiece is reasonablesince no lubricant was used. The �a is set at 0.4.

The thickness of the workpiece, t, is 1.19 mm. Based on Eq.�6�, h*=2.015 mm. Values for range of tool travel and the h1, h2,and h3 from the tool contact with the workpiece in the six stagesare summarized in Table 2. Experimental and modeling results ofthe thrust force and torque are presented in Fig. 14. Square sym-bols represent the model-predicted thrust force and torque. Thesesymbols are connected by lines and compared to experimentalmeasurements, which are adopted from Fig. 4. Ranges of the tooltravel in six stages are also marked in Fig. 14.

The model predicts that the peak thrust force occurs at about2 mm tool travel. It is smaller than the experimental measurement�2.5 mm� due to the assumption of no workpiece deformation.Deformation of the workpiece, as shown in Fig. 5, will delay thetool travel in reaching the peak thrust force. The predicted peakforce occurs in stage 3. The magnitude of peak thrust force isabout same, 750 N for the model and 700 N for experiment. Thedrop of the thrust force to about 300 N at the start of stage 4 issuccessfully predicted in the model. The gradual reduction of thethrust force in stages 5 and 6 is also well predicted.

The model prediction of the torque shows an early rise to1.6 N m at 2.3 mm tool travel. The experimental measurementshows the torque reaches the same level at 3 mm. The effect ofworkpiece deformation delays the tool travel from reaching thepeak value until late in stage 3. The torque is well predicted instage 4. In stage 5, when the tool starts to contact the cylindricalregion, the discrepancy of torque becomes more apparent. Thevariation of coefficient of friction due to temperature and speedare the likely causes for such discrepancy. As the tool penetratesinto the workpiece, the peripheral speed between the tool andworkpiece gradually increases. In the meantime, the workpieceand tool temperatures rise significantly. Only one coefficient offriction in the radial direction, �, is used to cover such wide rangeof contact conditions. More detailed investigations of the tribo-logical phenomenon between the tool and workpiece are requiredfor more accurate prediction of torque in friction drilling. Whenthe tool is retracting �distance from contact from12.7 to 14.4 mm�, the torque is still measurable in the range be-tween 1.1 and 1.4 N m. This indicates the existence of residualpressure on the tool after the hole is formed. The torque in thisstage is not modeled.

8 ConclusionsExperimental analysis and modeling of the friction drilling pro-

cess were conducted. Experimentally measured thrust force andtorque under constant tool speed were measured and analyzed.The workpiece and tool temperatures were measured using an IRcamera. A mathematical model based on the contact area betweenthe undeformed workpiece and the rigid, unworn tool was pro-posed. The temperature-dependent stress-strain properties of thework-material were included in the force model. The model-

kpiece contact in six stages „Unit: mm…

� Conical �h1, h2� Cylindrical �h3�

¯ ¯

2.985, h+2.015 ¯

2.985, h+2.015 ¯

h−3.205, h+2.015 ¯

h−3.205, 11.475 h−9.460¯ 1.19

or

, h2

.97

predicted thrust force and torque had good correlation with ex-

AUGUST 2006, Vol. 128 / 809

perimental measurements. Some discrepancies still existed andshowed the limitation of the simplified friction model used in thisstudy.

This research studied the mechanics in friction drilling. Morecomprehensive models based on the finite element method is re-quired. The model will require better understanding of the tribo-logical phenomenon, heat partition between the tool and work-piece, and prediction of coefficients of friction in radial and axialdirections. Research work has been conducted across differentwork-materials and will be further extended to friction drillingwith variable tool feed rate for process parameter and tool geom-etry optimization.

AcknowledgmentA portion of this research was sponsored by the High Strength

Weight Reduction Materials Program and the High TemperatureMaterials Laboratory User Program, Oak Ridge National Labora-tory, managed by UT-Battelle, LLC for the U.S. Department ofEnergy. Program management by Dr. Phil Sklad and Dr. PeterBlau is greatly appreciated. The assistance of Y. Zhu is also ac-knowledged.

References�1� Geffen, J. A. van, 1976, “Piercing Tools,” U.S. Patent No. 3,939,683.�2� Geffen, J. A. van, 1979, “Method and Apparatuses for Forming by Frictional

Heat and Pressure Holes Surrounded Each by a Boss in a Metal Plate or theWall of a Metal Tube,” U.S. Patent No. 4,175,413.

�3� Geffen, J. A. van, 1979, “Rotatable Piercing Tools for Forming Holes Sur-

810 / Vol. 128, AUGUST 2006

rounded Each by a Boss in Metal Plates or the Wall of Metal Tubes,” U.S.Patent No. 4,177,659.

�4� Geffen, J. A. van, 1980, “Rotatable Piercing Tools for Forming Bossed Holes,”U.S. Patent No. 4,185,486.

�5� Head, G. D., Le Master, W. C., Bredesky, L. P., and Winter, D. C., 1984,“Flow Drilling Process and Tool Therefore,” U.S. Patent No. 4,428,214.

�6� Hoogenboom, A. J., 1984, “Flow Drill for the Provision of Holes in SheetMaterial,” U.S. Patent No. 4,454,741.

�7� France, J. E., Davidson, J. B., and Kirby, P. A., 1999, “Strength and RotationalStiffness of Simple Connections to Tubular Columns Using Flowdrill Connec-tors,” J. Constr. Steel Res., 50, pp. 15–34.

�8� France, J. E., Davidson, J. B., and Kirby, P. A., 1999, “Moment-Capacity andRotational Stiffness of Endplate Connections to Concrete-Filled Tubular Col-umns With Flowdrilled Connectors,” J. Constr. Steel Res., 50, pp. 35–48.

�9� France, J. E., Davidson, J. B., and Kirby, P. A., 1999, “Strength and RotationalResponse of Moment Connections to Tubular Columns Using Flowdrill Con-nectors,” J. Constr. Steel Res., 50, pp. 1–14.

�10� Overy, K., 1978, “Flowdrilling—Bush Formation in Thin Metal,” CME, Chart.Mech. Eng., 25�7�, pp. 70–71.

�11� Bak, D., 1987, “Friction, Heat from Integral Bushings,” Des. News, 43�11�, p.124.

�12� Kerkhofs, M., Stappen, M. V., D’Olieslaeger, M., Quaeyhaegens, C., and Stals,L. M., 1994, “The Performance of �Ti,Al�N-Coated Flowdrills,” Surf. Coat.Technol., 68/69, pp. 741–746.

�13� Strenkowski, J. S., Hsieh, C. C., and Shih, A. J., 2004, “An Analytical FiniteElement Technique for Predicting Thrust Force and Torque in Drilling,” Int. J.Mach. Tools Manuf., 44, pp. 1413–1421.

�14� Schmidt, H., Hattel, J., and Wert, J., 2004, “An Analytical Model for the HeatGeneration in Friction Stir Welding,” Modell. Simul. Mater. Sci. Eng., 12, pp.143–157.

�15� Boyer, H. E., 1987, Atlas of Stress-Strain Curves, ASM International, p. 157.�16� Feng, Z., Gould, J. E., and Lienert, T. J., 1998, “A Heat Flow Model for

Friction Stir Welding of Aluminum Alloys,” Hot Deformation of AluminumAlloys II, T. R. Bieler, L. A. Lalli, and S. R. MacEwen, eds., The Minerals,

Metals and Materials Society, pp. 149–158.

Transactions of the ASME