CHAPTER I

INTRODUCTION

Thesis Objective and Scope

The objective of this research is to quantify the relationship between axial force,

spindle speed, travel speed, and other process parameters for friction stir welding (FSW)

at high spindle speeds (1500-5000 RPM), and correlate the results with a two

dimensional fluid flow model capable of predicting the forces and torque during FSW.

In order to build an efficient control system for robotic FSW, a valid force

prediction model needs to be established, and many basic physical mechanics of this

process need to be understood. The 2-D model presented here represents the initial steps

in developing a 3-D process model.

Process Overview

Friction stir welding (FSW) was invented and patented by W. M Thomas et al. [1]

of the Welding Institute in Cambridge, UK. In FSW, a cylindrical, shouldered tool with a

profiled probe is rotated and slowly plunged into the joint line between two pieces of

sheet or plate material, which are butted together. The pieces are rigidly clamped onto a

backing plate in a manner that prevents the abutting joint faces from being forced apart.

Frictional heat is generated between the tool shoulder and the material of the work pieces.

This heat causes the latter to reach a visco-plastic state that allows traversing of the tool

along the weld line. The plasticized material is transferred from the leading edge of the
tool to the trailing edge of the tool probe and is forged by the intimate contact of the tool

shoulder and the pin profile. It leaves a solid phase bond between the two pieces.

Applications

The current industries which utilize FSW are the aerospace, railway, land transportation,

shipbuilding/marine, and the construction industries. These industries have seen a push

towards using lightweight yet strong metals such as aluminum. Many products of these

industries require joining three-dimensional contours, which is not achievable using

friction stir welding heavy-duty machine tool type equipment with traversing systems

which are limited to only straight line or two-dimensional contours.

For these applications, industrial robots would be a preferred solution for

performing friction stir welding for a number of reasons, including: lower costs, energy

efficiency, greater manufacturing flexibility, and most significantly, the ability to follow

three-dimensional contours.

2
 Chapter II

PREVIOUS FRICTION STIR WELDING RESEARCH

In this chapter, an overview of the physical and mechanical properties of FSW

will be presented. Also, previous FSW research will be reviewed in order to show how

this research contributes to the science of FSW. An emphasis is placed on work that

relates to the topics of process parameter quantification and thermo-mechanical

modeling.

Terminology

To understand the process of friction stir welding and the focus of this research, it

is worth while to define certain terminologies and their usage in this thesis.

In FSW, the tool typically consists of a cylindrical shoulder with a profiled probe,

also called the pin. The material or materials being welded can be called the work piece,

part, sample, or plate. The joint where the samples are abutted will be referred to as the

weld line. The part used to support and clamp the sample is called the backing plate,

backing bar, or anvil.

The tool rotates at an angular velocity given in revolutions per minute (RPM),

which will be referred to as rotational speed (RS). The translational velocity at which the

tool travels along the weld line is called the feed rate or travel speed (TS), and will be

given in millimeter per second (mm/s) or inches per minute (ipm). The side of the weld

where the angular velocity and forward velocity of the pin tool are additive is called the

3
advancing or leading side. The other side where the angular velocity and translational

velocity are in opposite directions is called the trailing or retreating side.

As shown in Figure 2, forces act in three dimensional spaces. The force along the

X-axis, Y-axis, and Z-axis will be referred to as the translational (Fx), transverse (Fy), and

axial force (Fz) respectively, and will be given in Newtons (N). The moment (Mz) about

the axis of rotation will be referred to as the torque and given in Newton-meters (N-m).

Fz (axial force)

Direction of travel
Fx (translational force)

Weld line
Mz
Fy (transverse (torque)
force)
Trailing edge of
rotating tool

Leading edge of
rotating tool
Advancing side of
weld
Shoulder

Pin

Retreating side of weld

Figure 2.1 - Schematic of FSW.

4
Power however will be given in Watts (N-m/s). Figure 2 shows a schematic of the

process and the given terminologies.

Welding Materials

A wide range of materials can be successfully joined. These materials include

thermoplastics, lead, zinc, aluminum alloys, copper, silver and gold. Materials with

higher melting points (in excess of 1100°C) such as ferrous metals and alloys can also be

joined. However they require probes of high grade temperature resisting materials such

as tungsten [1].

Aluminum has been welded in single passes ranging from 0.050” to 2” in

thickness. Using a double pass method, welds up to 4” thick have been made [2].

Copper up to 2” thick has been welded. Welds up to 0.5” thick have been successfully

made in steel using the double pass method, and 0.37” thick magnesium alloy AZ61A has

been welded in a single pass [3].

Friction stir welding has successfully been performed in a variety of joint

geometries. Butt welds, corner welds, T-sections, overlap welds, and fillet welds have all

been done [2]. Circumferential welds have also been performed in the aerospace industry

for the manufacture of large cryogenic tanks [33].

Welding Tools

A FSW tool may be made out of a number of different materials. Choice of a

material for a tool is dependent on the type of metal material to be welded, particularly

the melting temperature of the material. An additional consideration is the desired travel

5
speed. Table 2.1 lists different tool material and the maximum operating temperatures

[4].

Table 2-1. Various tool materials. [4]

Material Approximate Max Work Temp (F)

H-13 1000

Ferro-TiC SK 1100

MP-159 1100

Stellite 6B 1600

Ferro-TiC HT-6A 1800

MAR-M-246 1900

Mo-TZM 2400

Rhenium 3600

Tungsten 3600

The tool has two basic parts; the shoulder and pin. The tool shoulder has two

general functions, create frictional heat at the tool/work piece interface and to cap the

plasticized material as it “stirred”.

The pin is a cylindrical pin projecting from the distal shoulder surface and has a

longitudinal axis co-extensive with the shoulder longitudinal axis. The pin must be large

enough to stay above the plastic stress level at operating temperatures. Current FSW

practice uses a pin having a surface profile consistent with the thread of a bolt, much like

6
the end of a machine bolt [4]. The purpose of profiling the pin is to reduce traverse loads

and improve material flow [5].

Tool pin shapes have taken the form of frusto-conical, inverted frusto-conical,

spherical, and pear shape, to simple conical, truncated cones, to slightly tapered cylinders

[5, 6]. Cocks et al. [7] introduced a pin which has a combined right handed and left

handed thread pattern. This “enantionmorphic” pin is said to produce welds of improved

mechanical properties [7].

For this research, a tool made of H-13 tool steel heat treated to Rc 48-50 with a

0.5” diameter cylindrical shoulder and a threaded cylindrical pin will be used.

Weld Microstructures

The heat and deformation generated during FSW produce four micro-structurally

distinct regions across the weld. They are the heat affected zone (HAZ), thermo-

mechanically affected zone (TMAZ), dynamically recrystallized zone (DXZ) or weld

nugget, and the unaffected material. [8]

The HAZ is the outermost portion of the weld which is modified by the thermal

field of the welding process but does not experience any deformation. It is similar to the

heat affected zones observed in welds prepared by more conventional fusion welding

processes. Inward from the HAZ is located the TMAZ, where the material experienced

plastic deformation due to the stirring process in addition to the heat-induced micro-

structural changes. At the center of the weld, where the heat and deformation are the

greatest, aluminum alloys undergo significant grain refinement within an onion-shaped

region called the weld nugget or DXZ, which is approximately the size of the rotating pin

7
of the tool. The unaffected, or parent-material, is material that is heated but not modified

by the thermal field of the weld.

Mechanical Properties

In whole-weld tensile tests, most precipitation-strengthened aluminum alloys

exhibit similar yielding and fracture behavior [9–12]. During these tests, the tensile

strain becomes localized in the HAZ on both sides of the weld nugget [13]. Fracture will

typically occur at this location and will usually be located on the retreating side of the

weld [12]. The localization of yield and fracture at the HAZ demonstrates the importance

of this region in controlling the mechanical behavior of friction stir welds. Despite this,

there have been few systematic examinations of the HAZ to determine the underlying

cause of this behavior.

Some studies [9–11, 14–17] have demonstrated that precipitates are significantly

coarsened in the HAZ relative to those observed in the unaffected base plate or weld

nugget. Sato et al. [18] examined different locations in the HAZ and weld nugget of a

6063 Al FSW and observed that the precipitates experienced increasing dissolution

toward the weld center. Su et al. [19] recently reported on precipitate evolutions

occurring in a 7050 Al FSW. They observed a coarsening of precipitates from the base

plate into the TMAZ, with increasing dissolution and re-precipitation occurring from the

TMAZ into the weld nugget.

Kwon et al. [20] investigated the influence of the tool rotation speed on the

hardness and tensile strength of the friction stir welding aluminum 1050 and concluded

that the hardness within the weld was higher on the advancing side than on the retreating

8
side. Also, that in the transition zone between the weld and the parent material the

variation in hardness was more drastic on the advancing side than on the retreating side

and that the hardness and tensile strength of the weld increased significantly with

decreased tool rotation speed.

Lee et al. [21] examined the microstructure and mechanical properties of FSW

6005 Al alloy with increasing welding speed and concluded that the tensile strength

increased as welding speed increased.

Experimental and Theoretical Modeling

In this section, previous works pertaining to thermal-mechanical modeling will be

reviewed. Since little is known about the physics involved during the FSW process, these

works will help to provide insight into the mechanics of FSW.

Thermo-mechanical Modeling

Ulysse et al. [22] attempted to model the friction stir-welding process using three-

dimensional visco-plastic modeling. The simulation was limited to one tool geometry

where the tool pin was 6.4 mm in diameter and its depth into the plate was 6.4 mm,

which is about 1/3 of the plate thickness. The pin was tilted by 3° from the vertical,

leaning away from the direction of welding. The tool shoulder was 19 mm in diameter.

The shoulder face was a 7° concave cone design.

In the model, a cylindrical shoulder recess was assumed in order to approximate

the shallow concave area as shown in Figure 2.2. The tool above the work surface was

approximated as a 20 mm high cylindrical shaft. The 3D finite-element (FE) friction stir-

9
welding simulations were conducted using the commercial software FIDAP [24]. The

mesh used for the FSW simulations are shown in Figures 2.2 and 2.3; about 33,000 eight-

noded (brick) elements and 29,400 nodes were used in this study.

Figure 2.2 – Enlarged view of Ulysse FSW tool FE mesh. [22]

In addition, only butt joints 19.1 mm AA 7050-T7451 (2.3% Cu, 2.25% Mg, 6.2%

Zn) thick plates were considered in this work. The model of the work-piece region was

60 mm wide by 100 mm in length as shown in Figure 2.3. The support table, located

underneath the work-piece, is not included in the analysis in order to reduce the size of

the numerical model. Therefore, heat transfer to the support table is ignored in this work.

10
 Figure 2.3 – FE mesh of the welding model. [22]

Ulysse [22] modeled the large plastic deformation involved in stir-welding

processes by relating the deviatoric stress tensor to the strain-rate tensor. The TMAZ was

assumed to be a rigid-visco-plastic material where the flow stress depends on the strain-

rate and temperature and is represented by an inverse hyperbolic-sine relation as follows:

 1

1  n
Z  Q 
σ e = sinh −1    Z = ε& exp  Eq. 2.1
α  A    RT 
 

where α, Q, A, n are material constants, R the gas constant , T the absolute temperature

and Z the Zener-Hollomon parameter [42]. The material constants were determined

using standard compression tests. The mechanical model equations are complete after

appropriate boundary conditions are prescribed.

The temperature distribution is obtained by solving the energy equation,

expressed here as the conductive–convective, steady-state equation

11
 Eq. 2.2)

where ρ is density, cp the specific heat, u the velocity vector, k the conductivity, θ the

temperature and Q. is the internal heat generation rate. About 90% of the plastic

deformation is assumed to be converted into heat [23]. In this work, temperature-

dependent conductivity and specific heat coefficients for aluminum alloys were adopted.

The heat generation rate term can be expressed as the product of the effective stress and

effective strain-rate.

Comparisons of model predictions with experimental data are illustrated in

Figures 2.4-2.6. All temperatures are peak temperatures. The trend of the measured data

is also indicated for convenience in the figures. The following parameters were used in

the comparisons: (1.0 mm/s, 11.7 rev/s), (1.37 mm/s, 8.17 rev/s), (1.9 mm/s, 11.7 rev/s),

(2.593 mm/s, 11.7 rev/s), (3.54 mm/s, 8.17 rev/s), (1.9 mm/s, 11.7 rev/s), (2.593 mm/s,

11.7 rev/s), (3.54 mm/s, 8.17 rev/s), (3.5mm/s,25.5rev/s).

While various temperature measurements have been recorded, experimental

measurements to validate the present force predictions are not available. Analytical

predictions of axial (Fz) and shear forces on the pin are shown in Figure 2.5 as a function

of translational speed. It can be observed that increasing the welding speed, regardless of

rotational speed, has the effect of increasing the axial force thrust and shear force on the

pin. In addition, for a fixed welding speed, increasing the rotational speed has the effect

of decreasing the forces. Quantification of this relation over a wide parametric range is

the core topic of this thesis.

12
Figure 2.4 – FSW temperatures as a function of tool rotational speed. [22]

Figure 2.5 - Axial (Fz) and shear forces on the pin shown as a function of
translational speed. [22]

13
 Figure 2.6 – Axial and shear forces on pin as function of tool
rotational speed. [22]

Reynolds et al. [25] introduced a two dimensional model based on fluid

mechanics that modeled the solid state material transport during welding as a laminar

viscous flow of a non-Newtonian fluid past a cylinder. Only the tool pin was represented

in the simulation. The temperature and strain rate dependent viscosity of AA6061 was

based on the constitutive law of the flow stress of aluminum alloys using the Zener-

Hollomon parameter (Z) (Eq. 2.1). Also, temperature dependent thermal conductivity

and specific heat were used to calculate the heat transfer in the fluid.

Reynolds et al. [25] concluded that the force against the welding direction at the

pin increases with increasing TS at constant SS and decreases with increasing SS at

constant TS. The power increases with increasing SS at constant RS and remains

constant with varying TS and constant SS.

14
 Colegrove [30] used an advanced analytical estimation of the heat generation for

tools with a threaded probe to estimate the heat generation distribution. The fraction of

heat generated by the probe is estimated to be as high as 20%, which leads to the

conclusion that the analytical estimated probe heat generation contribution is not

negligible.

In parallel with the analytical model, Colegrove and Shercliff et al. [30, 31]

developed a material flow model, which addresses the influence of threads on the

material flow. An advanced viscous material model is introduced and the influence of

different contact conditions prescribed as the boundary condition is analyzed [31].

Schidmt et al. sought to establish an analytical model for heat generation during

friction stir welding based on different assumptions of the contact condition between the

rotating tool surface and the weld piece. The material flow and heat generation are

characterized by the contact conditions at the interface and are described as sliding,

sticking or partial sliding/sticking. Different mechanisms of heat generation were found

to be behind each contact condition. The analytical expression for the heat generation is

a modification of previous analytical models known from the literature [29, 30] and

accounts for both conical surfaces and different contact conditions.

Chen et al. [32] introduced a three-dimensional model based on finite element

analysis to study the thermal history and thermo-mechanical process in the butt-welding

of aluminum alloy 6061-T6. The model incorporates the mechanical reaction of the tool

and thermo-mechanical process of the welded material. The heat source incorporated in

the model involves the friction between the material and the pin and the shoulder. The

dynamics of the FSW thermo-mechanical process, the thermal history and the evolution

15
of longitudinal, lateral, and through-thickness stress in the friction stirred weld are

simulated numerically. The X-ray diffraction (XRD) technique is used to measure the

residual stress of the welded plate.

Chen et al. [32] suggested that the maximum temperature gradients in longitudinal

and lateral directions are located just beyond the shoulder edge, and that the longitudinal

residual stress is greater than the lateral residual stress at the top surface of the weld. The

prediction shows that the high stress is located in the region extending down from the

crown to the mid-thickness of the weld. A higher traverse speed induces a larger high

longitudinal stress zone and a narrower lateral stress zone in the weld.

Nunes et al. [26] modeled the tool torque to be totally due to the shear flow stress

of the metal acting perpendicular to the direction of tool rotation and integrated over the

surface. The welding power is equal to the torque multiplied by the RS and is given by

P = ω 0M z
Rs Rp
M z = ∫
Rp
2π R p σ d R p + 2π R p2 t σ + ∫ 2π R p2 σ d R p ,
0
Eq. 6.3

where Mz is the welding torque, ωo is the tool rotational speed, Rp is the pin radius, Rs is

the shoulder radius, σ is the shear flow stress, and t is the pin depth. Figure 2.7 shows a

comparison of Mitchell et al. [27] experimental torques to torques computed from the

Nunes et al. model [26].

16
 600

500

400
Torque (in-lbs

Measured
300
Computed
200

100

0
350 525 700
Speed (RPM)

Figure 2.7 - Comparison of Mitchell et al. [27] experimental torques to torques

computed from Nunes et al. model.

Flow Field Visualization

The visualization of material flow during friction stir welding is key to

understanding the internal workings of the process and is critical to process modeling.

Nunes et al. [26] state that the flow field around a standard threaded tool or tools

with threadlike features is comprised of two parts. The primary flow of material is

around the tool in the direction of rotation. There is a plug of metal that rotates with the

pin, and the flow gradient goes from the velocity of the plug at the plug surface to zero

some distance away. Nunes et al. [26] states that the plug radius occurs at a point that

creates a minimum torque for a given shear flow stress. The plug hypothesis is supported

17
by computational fluid dynamics work done by Dong et al. [33], which predicted that a

plug of metal rotates with the tool. The plug model can be seen below in Figure 2-8.

Figure 2.8 - Secondary Flow Caused by Threads.

Ulysse [22] was able to extract useful information along the path line of a material

particle traced in the welding region from the visco-plastic FSW model. The path line of

a material particle located just underneath the pin tip was traced back to its original

location in the shoulder recess. The temperature history of the particle along the path line

is shown in Fig. 2.9. As seen in the figure, the particle was traced back in time from t = 2

s (at the pin tip) to 0 s (near the shoulder region).

For about 0.5 s, the particle rotates around the top of the pin with a relatively

small downward or axial velocity Uz; and during this time, its temperature remains

18
relatively uniform. After the initial 0.5 s, the particle follows the pin rotation and

downward motion which sets up some periodicity in the temperature history.

Figure 2.9 - Temperature (in °C) history of a material particle near the pin. [22]

As the particle rotates, it crosses cold and hot regions located at the leading and

trailing edges of the pin, respectively, which explains the peaks in Fig. 2.9. The material

particle is also pushed downward, toward the tip of the pin, where temperatures are

relatively cold, which explains the decreasing trend in temperature.

Guerra et al. [28] investigated the flow of metal during FSW using a faying

surface tracer and a pin frozen in place during welding. It was shown that the material is

transported by a wiping of material from the advancing front side of the pin onto a zone

of material that rotates and advances with the pin. The material undergoes a helical

19
motion within the rotational zone that rotates, advances, and descends in the wash of the

threads on the nib and rises on the outer part of the rotational zone. After one or more

rotations, this material is sloughed off in the wake of the pin, primarily on the advancing

side. The second process is an entrainment of material from the front retreating side of

the nib that fills in between the sloughed off pieces from the advancing side.

Colligan et al. [29] followed material flow in 6061 and 7075 aluminum by

imbedding small steel balls as tracers into grooves cut into the work-piece parallel to the

weld direction. Grooves were cut parallel to the weld direction but at various distances

from the weld centerline and at various depths. After welding, the distribution of the

steel balls was revealed by radiography in both the plan and the cross-sectional views.

Results

are displayed nicely in the original paper but, in general, the work showed that the

material striking the pin on the advancing side of the weld would be displaced from the

rear of the retreating side of the pin.

20
 CHAPTER III

EXPERIMENTAL PROCEDURE

Both experimental and analytical results [22, 23] show that the axial force (and

other forces) can be reduced by increasing the spindle speed. The full range over which

this apparent relationship can be expected to hold true is not known. A full quantification

of the relationships between spindle speed and other process parameters for friction stir

welding are needed. These relationships are of fundamental importance to improved

weld productivity with the friction stir welding process and are key to the widespread use

of robots for FSW.

Vanderbilt University Welding Automation Laboratory FSW Experiments

To study this relationship, experiments were performed at the Vanderbilt

University Welding Automation Laboratory using a Milwaukee #2K Universal Milling

Machine fitted with a Kearney and Trecker Heavy Duty Vertical Head Attachment

modified to accommodate high spindle speeds. The weld sample, clamping fixture (or

backing plate), tool design, instrumentation, and machine modifications detailed below.

Sample Description

For this experiment, plates of AL 6061-T651 aluminum, nominally 0.250 inches

thick were friction stir welded. The samples were 3 inches wide by 18 inches long. The

21
tool depth was set to 0.145”. To ensure precise setting of the tool depth, the tool was

positioned along the weld line aft of the samples leading edge. The sample is clamped

via the scheme shown in Figure 3-1.

Figure 3-1: VU FSW backing plate.

The clamping system allows for 30 inches of travel and samples with 3” or 5”

widths. Using the horizontal spindle, the maximum travel distance was 12 inches and

limited to 3” width samples. See Appendix A for a detailed schematic of the backing

plate.

Tool Design

For this experiment, the tool was made from H-13 tool steel heat treated to

Rockwell c hardness 48-50. The shank diameter was 1”. The tool shoulder was flat with

a 0.50” diameter. The pin was cylindrical with a 10-24 threads per inch left hand pattern.

22
The pin length was 0.1425” and the diameter was 0.190”. Heat sinks were cut into the far

end of the tool shank near the shoulder to facilitate heat dissipation during welding. The

tool was rigidly mounted into the tool holder using a twist lock system. The tool lead

angle was set to 2º. Figure 3-2 shows a detailed schematic of the tool.

Figure 3-2: VU FSW tool with 0.5” shoulder and left hand 10-24 thread pattern.

Instrumentation

A Kistler rotating quartz 4-component dynamometer (RCD) was used for

measuring forces and torque on the rotating tool. The dynamometer (Figure 3-3) consists

of a four component sensor fitted under high preload between a base plate and a top plate.

The four components are measured without displacement. The four component sensor is

ground-insulated, therefore ground loop problems are largely eliminated. The

dynamometer is rustproof and protected against penetration of splash water and cooling

agents. For each component a 2-range miniature charge amplifier is integrated in the

23
dynamometer. The output voltages of the charge amplifiers are digitized and transmitted

by telemetry to the stator and then acquired by a PC. The stator is rigidly mounted

concentrically with the RCD with a 2 mm gap between them. A mount was fabricated

and bolted to the face of the vertical head. A detailed schematic of the stator mount can

be seen in Appendix A.

The Kistler data acquisition software DynoWare was used for data collection.

DynoWare records the three forces and torque during welding and allows the data points

to be exported to a tabularized text file. The data is then imported into MATLAB 7.0,

where it is run through a linear smoothing filter and is plotted. Please see Appendix C for

the linear smoothing filter program file.

Telemetry
Pickup

KISTLER

Figure 3-3: Kistler Rotating Cutting Force Dynamometer

24
Machine Modifications

Welding is performed on a Milwaukee #2K Universal Milling Machine fitted with

a Kearney and Trecker Heavy Duty Vertical Head Attachment modified to accommodate

high spindle speeds. The vertical head clamps the vertical sliding surface of the milling

machine. A Baldor VM2514, 20 HP, 3450 RPM, 3Phase 230 VAC motor is mounted to

the shoulder of the head and drives the vertical spindle via a Poly-V belt (Browning

380J16) and drive system. The motor is controlled by a Cutler Hammer SVX-9000 20HP

variable frequency drive. Please see Appendix A for a detailed schematic of the motor

mount.

To meet the operational speed requirements, a 1.33 pulley over drive ratio was

used. The large pulley’s (Browning 16J60P) diameter was 6” while the smaller pulley’s

(Browning 16J45P) diameter was 4.5”. The maximum speed using the above

configuration is 4800 rpm. The overdrive ratio was selected to prevent the possibility of

over-speeding the RCD, whose max operational speed is 5000 rpm. Over-speeding

would require the RCD to be recalibrated. Therefore the maximum rotational speed at

which data was collected for this experiment was 4500 rpm.

To reduce the inertial load of the vertical spindle, the gear train which coupled the

head to the milling machine drive was removed. The gearing system total weight was

approximately 50 lbs. This reduction of loading allows for more torque to be available

during welding.

The head was originally grease lubricated, and accommodated a maximum

operational speed of 1500 rpm. To suit the higher operational speeds for the experiment,

the lubricating grease was cleaned from the spindle’s tapered roller bearings. A Bijur

25
Fluid Flex Pressurized Lubricating System was used to lubricate the tapered roller

bearings.

The Fluid Flex system dispenses a mixture of compressed air (125 psi maximum)

and oil (DTE Lite ISO VG 32). The compressed air is filtered through an air

filter/regulator (160 Psi maximum) with a ¼ NPT inlet. The air enters the Fluid Flex

system and is reduced to a desired level and passed through a solenoid valve which

synchronizes the system with the spindle. Low-pressure air enters the fluid reservoir and

forces fluid from the reservoir. Separate lines carry an atomized mixture of air and oil

through the distribution lines in the system to the Jet Tip assembly for discharge onto the

tapered roller bearings (Timken #455 and #749) of the spindle. Figure 3-4 shows the VU

FSW test bed.

26
 20 HP Motor

Air/Oil Lube System

V-Belt and
Pulley System

Vertical Head

Air/Oil
Delivery Lines

Kistler Dynamometer

FSW Tool
Axial Position
Monitor

Weld Sample

Backing Plate

Figure 3-4 VU FSW test bed.

The experimental procedure is as follows:

1. Program the spindle speed.

2. Set the travel speed.
3. Set the tool depth.
4. Start the lubrication system.
5. Start the machine spindle.
6. Plunge the tool into the material.
7. Start the data acquisition.
8. Engage the table travel.
9. Travel 18”.
10. Stop the spindle.

27
 Using the procedure above, welds were made for the parameter set below.

Table 3-1: VU FSW Experimental Parameter matrix.

Feed Rate (ipm)

11.4 27 37.2 44.8 53.3 63.3

Spindle Speed (RPM)

1500 X X X X X X

2250 X* X X X X X

3000 X** X X X X X

3750 - X X X X X

4500 - - - X X X

*Actual Rotational Speed was 2000 RPM

** Actual Rotational Speed was 2500 RPM

The parameters for which welding was not conducted are those where the

rotational speed is too high for the travel speed and creates a weld with a deformed

surface as shown in Figure 3-5. The overheat phenomena occurred at the preceding

parameter set. For example, for the 3750 rpm and 37.2 ipm parameter set, the weld

experienced the overheat phenomena. Therefore a weld for 4500 rpm and 37.2 ipm was

not run because the surface deformation is assumed to only increase. Detailed

photographs of the deformed welds can be seen in Appendix C. A discussion of the

28
suggested optimum weld pitch, the ratio of rotational speed to travel speed for a specified

parameter pair, will be discussed in Chapter 5.

Figure 3-5: Typical weld deformation for overheat phenomena experienced during
experimentation. (Parameter Set: 3750 rpm and 27.7 ipm.)

29
 CHAPTER IV

THEORETICAL MODELING PROCEDURE

In FSW, a cylindrical, shouldered tool with a profiled probe is rotated and slowly

plunged into the joint line between two pieces of sheet or plate material, which are butted

together. Following tool penetration, the friction stir welding operation depends on

continuous refurbishment of the visco-plastic layer surrounding the rotating tool. The

term 'third body' has been used to describe the region containing the visco-plastic

material produced during frictional welding and friction surfacing [34]. This terminology

will be applied for the remainder of this thesis.

It is apparent that the development of a satisfactory 3-dimensional process model

for FSW will depend on how well the ‘third body’ region is handled, in particular how

the material properties in this region are determined. A logical first step in developing a

full 3-dimensional representation of the friction stir welding process is to develop a

working 2-dimensional model. For this thesis, only the pin bottom and the sample are

considered. This allows a basic approach where the model is modularly developed. The

model is validated in stages, which allows the contribution of various components to be

examined and understood. The key factors during FSW which were considered in this

model are detailed below.

Currently FSW process modeling typically incorporates either a solid or fluid

mechanics approach. Experimental results have been shown to correlate with models

using either approach. Due to the moderately high temperatures associated with FSW (up

to 480 °C) (Sato et al. [18]), and the relatively low melting point of Al 6061-T6 (652°C);

30
it is clear the weld material in the third body region enters what is called a mushy zone

[35].

A mushy zone is a temperature region where the material is not a true solid or

liquid, though it has aspects of the behavior of both. Understanding and accurately

modeling the third body region will lead to an optimal 3-D model.

FSW Modeling: A Fluid Mechanics Approach

Using a fluid mechanics approach, the third body region is approximated as a

viscous flow domain under high shear stress and strain-rates at moderately high

temperatures. The thermo-mechanical property of primary importance for a model such

as this is the material viscosity.

Mechanical Model: Part 1

For the fluid mechanical approach to FSW modeling, the determination of the

material viscosity is the logical first step to begin development of a 3-dimensional model

for friction stir welding. North et al. [36] experimentally correlated the material viscosity

during FSW with the viscosity of a fluid intermediate between two concentric cylinders

as first suggested by Couette in 1890 [45]. Figure 4-1 shows a schematic of the Couette

Viscous Flow model. The material viscosity is found to be [45]

µ = (r12- ro2) M / [4π r12 ro2 (ω1 – ω0)] Eq. 4.1

The inner cylinder has radius ro , and angular velocity ωo while the outer cylinder has r1

and ω1, respectively, and M is the torque per unit depth of the tool pin. Applying

31
Equation 4.1 to FSW; take ro to be the radius of the tool pin, and set ωo equal the tool

rotational speed.

Ta

q0

r0
ω0

r1

ω1
Figure 4-1: Geometry and boundary conditions for the simple Couette flow model.

The outer cylinder radius r1 is taken to be the radius of the tool pin plus the width of the

third body region to a point in space where the material is solid and does not rotate,

giving ω1 = 0. M is the experimentally measured steady state welding torque for the

parameter sets in Table 3-1.

Couette flow will be unstable if it satisfies the Rayleigh criterion for Couette flow

instability [37]. The criterion states that Couette flow will be unstable if

d
ωo ro2 < 0 Eq. 4.2
dro

32
 ω0r02 > ω1r12 Eq. 4.3

Equation 4.2 simplifies to ωo ro , which is always greater than zero for friction stir

welding. Since ω1 = 0 when applying Couette flow to FSW, Equation 4.3 simplifies to

ωor12 > 0, which is always true for FSW. There fore the flow is always stable for FSW.

The width of the third body region as shown in Figure 4-2, and is modeled as

[
Wr = α βφ 2 + R p φ + 1] −1 Eq. 4.4

where ω is the rotational speed of the tool pin, vf is the travel speed, and Rp is the radius

of the tool pin [ 38]. For notation purposes, from this page forward in this thesis, Rp will

be used as the notation for the tool pin radius.

Figure 4-2: Schematic for the approximation of the third body region. [38]

33
 The variables α, β, and φ are given by

α = 1/2 [(Rp) - λ δ2]-1 Eq. 4.5

β = (Rs2 – Rp2) / (hλ) Eq. 4.6

φ = ωvf-1 Eq. 4.7

where Rs is the radius of the shoulder, δ2 is the projected thread area, and λ (24 threads

per inch) is the number of threads per inch. Since the model presented here is two

dimensional, secondary flows created by the threads are neglected.

Figure 4-3 shows the material viscosity for FSW using the Couette flow model.

Couette Viscosity for FSW

1400000

1200000

1000000
V isco sity (P a-s)

11 ipm
800000 27 ipm
37.2 ipm
44.8 ipm
600000 53.3 ipm
63.3 ipm

400000

200000

0
1500 2250 3000 3750 4500

Rotational Speed (rpm)

Figure 4-3: Couette Flow Viscosity for FSW.

34
Mechanical Model: Part 2

Applying the rigid-visco-plastic viscosity approximation used by Ulysse [22] and

Reynolds et al. [25], the material viscosity is found as

σe
µ= Eq. 4.8
3ε&t

where σe is the flow stress and is found by Equation 2.1. The material constants for the

constitutive law (Equation 2.1) for Al 6061 are: α = 0.045 (Mpa)-1, Q = 145 kJmol-1, A =

8.8632E6 s-1, n = 3.55, and R = 8.314 mol-1K-1. The material constants were first

published by Sheppard and Jackson [42]. The time average mean strain rate (έt) is found

as a function of the geometry, extrusion zone width, and FSW processing parameters as

suggested by Abregast et al. [38].

Eq. 4.9

Figure 4-4 shows the mean strain-rate for the parameter sets in Table 3-1.

35
 Strain-Rate vs. Rotational Speed
350

300

250
Strain-Rate (s^-1)

11.4 ipm
200 27 ipm
37.2 ipm
44.8 ipm
150 53.3 ipm
63.3 ipm

100

50

0
1500 2250 3000 3750 4500

Rotational Speed (rpm)

Figure 4-4: Mean Average Strain-Rate vs. Rotational Speed for Constant Travel
Speeds.

The flow stress depends on the strain-rate and temperature and is represented by

an inverse sine-hyperbolic as shown in Equation 2.1. As shown in Figure 4-5, using the

time average mean strain-rate found above, the flow stress for temperatures ranging from

20ºC to 720ºC were found for each parameter set in Table 3-1.

36
 Visco-Plastic Approximation of Flow Stress vs Strain Rate

350

300
T=293
T=393
T=493
T=593
250
T=693
T=793
F lo w S tr e s s (M p a )

T=893

200

150

100

50

0
0 50 100 150 200 250 300 350 400

Strain Rate (s^-1)

Figure 4-5: Flow Stress vs. Strain Rate using the Visco-Plastic model.

Then, applying Equation 4.8, the strain-rate/temperature dependent material viscosity was

found for each parameter set in Table 3-1.

Thermal Modeling

A very important factor during friction stir welding is the steady state welding

temperature. For this experiment the rotational speed was varied for the range of travel

speeds listed in Table 3-1. It was observed that the rotational speed is the primary factor

37
associated with heat generation. During the experiment, the rotational speed was

constant, which should result in a constant rate of heat generation.

Various models exist which predict the heat generation during FSW. In the model

by Chao and Qi [39], the heat generation comes from the sliding friction, where

Coulomb’s law is used to estimate the shear or friction force at the interface. Russell and

Shercliff [40] based the heat generation on a constant friction stress at the interface, equal

to the shear yield stress at elevated temperature, which is set to 5% of the yield stress at

room temperature. The heat input is applied as a point source or line source as in the

normal version of Rosenthal’s equations, but the solution is modified to account for the

limited extent of the plate width. Schmidt et al. [41] estimated the heat generation based

on assumptions for different contact conditions at the tool/material interface in FSW

joints. In this thesis, a sticking condition at the tool/material interface is assumed. The

heat generation model during FSW for a sticking interface [42] at the pin sides is

Qp = 2π(σy / √3) Rp2h ωo Eq. 4.10

where σy is the yield stress of AL 6061-T6, Rp is the radius of the pin, h is the height of

the pin and ωo is the rotational speed. The incoming weld material temperature was set to

300 K. Applying Equation 4.9 for AL 6061 – T6 at 300 K, gives σy = 241 MPa.

Substituting this result into Eq. 4.10 gives the heat generation shown in Figure 4-6.

38
 Rotational Speed Versus Heat Generation
for the Tool Pin Sides

10000

9000

8000

7000
H e a t G e n e r a t io n (W )

6000

5000 Series1

4000

3000

2000

1000

0
1500 2250 3000 3750 4500
Rotational Speed (rpm)

Figure 4-6: Heat Generation as a function of Rotational Speed for a sticking contact
surface.

Numerical Model

The computational fluid dynamics package FLUENT [43] was used to simulate

flow past a 2-dimensional pin for the weld parameters given in Table 3-1. The FLUENT

package includes the following software:

1) FLUENT, the solver.

2) prePDF, the preprocessor for modeling non-premixed combustion in FLUENT.

3) GAMBIT, the preprocessor for geometry modeling and mesh generation.

4) TGrid, an additional preprocessor that can generate volume meshes from existing

boundary meshes.

39
5) Filters(translators) for import of surface and volume meshes from CAD/CAE packages

such as ANSYS, CGNS, I-DEAS, NASTRAN, PATRAN, and others.

The flow field around the pin is modeled in 2D with the Z-axis of the pin

perpendicular to the direction of flow. The pin is modeled as a circle and a square flow

domain is created around it. Velocity at the inlet of the flow domain is specified by the

user. The procedure for implementing the model is:

1. Create the geometry (pin and the flow domain)

2. Mesh the domain

3. Set the material properties and boundary conditions

The geometry and mesh (Figure 4.7) are created using Gambit and are exported to

FLUENT.

Figure 4-7: 2-D mesh of pin and sample with the origin shown at the tool
pin center.

40
 The pin has a diameter of 0.190” and the sample dimensions are 6”x3”. Because the

area of most interest is near the pin, the mesh is densest near the pin and in the primary

flow path of the material which contacts the pin. A 10-row boundary layer is placed at

the pin wall. The flow domain is meshed using a quad map scheme using user specified

discretization values.

Process Model

The three dimensional friction stir welding process is represented here by only the

pin bottom and the sample. This 2-D model is the logical first step in developing a 3-D

model. The process is represented here as a steady state two-dimensional laminar,

incompressible, non-Newtonian flow past a rotating non-threaded tool pin. Plunge

conditions are ignored. The heat generation, as detailed earlier in this chapter, is assumed

to be mainly due to the contact condition from rotation of the tool pin, and the

contribution from forward travel of the tool pin is negligible.

Material Properties

The material properties used as input variables for FLUENT are detailed below.

Table 4-1 lists the constant properties for H-13 tool steel while Tables 4-2 and 4-3 list the

temperature dependent properties for the weld material (AL 6061).

41
 Table 4-1: FSW Tool Properties

Material Property Density (ρ) Thermal Conductivity (k) Specific Heat (Cp)

H-13 7805 (kg/m3) 202 W/(m-K) 871 (J/kg-K)

Table 4-2 Temperature Dependent Yield Strength of AL 6061-T6

T (K) σy (MPa)

311 241

339 238

366 232

394 223

422 189

450 138

477 92

533 34

589 19

644 12

42
 Table 4-3 Temperature Dependent Thermal Conductivity and
Specific Heat for AL 6061-T6

T (K) K [W / (m-K)] Cp [J / (kg-K)]

293 195 870

373 195 950

473 173 980

573 211 1020

673 212 1060

773 225 1150

873 200 1160

915 90 1170

973 91 1170

1073 92 1170

Boundary Conditions

The pin is assigned a constant rotational speed. The heat flux and heat generation

rate are determined by dividing Equation 4.10 by the pin surface area and volume

respectively. The wall thickness is set equal to the pin radius. The pin material is H-13

tool steel. The properties are listed in Table 4-1.

The lateral edges of the mesh are specified as a translational flow domain. This

feature specifies that the fluid translates between these boundaries with a user specified

43
velocity magnitude and direction. The magnitude is equal to travel speed during welding

and the flow direction is in the positive X direction.

The inlet velocity is set to the travel speed with the flow direction in the positive X

direction. The inlet temperature of the weld material is set to 300 K. The outlet pressure

is set to 101,325 Pa.

The fluid is AL 6061 with temperature dependent properties from Table 4-2 and 4-3.

The viscosity is for each found using Equation 4.1 (Couette ) or 4.8 (Visco-Plastic). The

weld material is assumed to be fed towards the tool at the user specified travel speed and

temperature.

Solver

The solver controls for the simulations were set to 2-D, segregated, laminar, implicit,

and steady flow. Using a segregated solver, the governing equations are solved

sequentially (i.e., segregated from one another). Because the governing equations are

non-linear (and coupled), several iterations of the solution loop must be performed before

a convergent solution is obtained. The iteration process consists of the steps illustrated in

Figure 4-8 and are outlined below:

The steps are as follows:

1. Fluid properties are updated, based on the current solution. (If the calculation has
just begun, the fluid properties will be updated based on the initialized solution.)

2. The u, v, and w momentum equations are each solved in turn using current values
for pressure and face mass fluxes, in order to update the velocity field.

3. Since the velocities obtained in Step 2 may not satisfy the continuity equation
locally, a ``Poisson-type'' equation for the pressure correction is derived from the
continuity equation and the linearized momentum equations. This pressure
correction equation is then solved to obtain the necessary corrections to the

44
 pressure and velocity fields and the face mass fluxes such that continuity is
satisfied.

4. Where appropriate, equations for scalars such as turbulence, energy, species, and
radiation are solved using the previously updated values of the other variables.

5. When inter-phase coupling is to be included, the source terms in the appropriate

continuous phase equations may be updated with a discrete phase trajectory
calculation.

6. A check for convergence of the equation set is made.

Update Properties

Solve Momentum Equations.

Solve Pressure-Correction (continuity) equation,

Update pressure, face mass flow rate.

Solve Energy, species turbulence and

other scalar equations

Converged? Stop

Figure 4-8: Overview of the Segregated Solution Method

These steps are continued until the convergence criteria are met.

45
 Using the implicit solution method, for a given variable, the unknown value in

each cell is computed using a relation that includes both existing and unknown values

from neighboring cells. Therefore each unknown will appear in more than one equation

in the system, and these equations must be solved simultaneously to give the unknown

quantities.

Laminar flow of the weld material was assumed because of the large viscosity

values (on the order of 106 Pa-s) that were found using Equation 4.1 and 4.8, leading to

very small Reynolds numbers (approx. 10-5) close to the pin. Due to the low Reynolds

number, only viscous effects are important and inertial effects may be neglected. The

material is assumed to translate past the rotating tool pin as it does in the actual

experiments.

Governing Equations

FLUENT uses the solver configuration above to solve the conservation of mass,

momentum (Navier-Stokes equations), and energy equations. For two-dimensional

steady state incompressible fluid flow the continuity equation is

∂u ∂v Eq. 4.11
+ = 0
∂x ∂y

where x and y denote coordinates and u and v are velocity components

Neglecting gravitational and body forces, the conservation of momentum yields

the following form of the Navier-Stokes equations

46
 ∂u ∂u ∂p  ∂ 2u ∂ 2u  Eq. 4.12
ρu + ρv = − + µ  2 + 2 
∂x ∂y ∂x  ∂x ∂y 

∂u ∂u ∂p  ∂ 2v ∂ 2v  Eq.4.13
ρu + ρv = − + µ  2 + 2 
∂x ∂y ∂y  ∂x ∂y 

where ρ is the density, µ is the viscosity, and p is the static pressure. The stress tensor is

given by τ

 ∂u ∂v 
τ = µ  +  Eq.4.14
 ∂y ∂x 

Neglecting changes in potential energy and assuming that heat transfer obeys Fourier’s

law of heat conduction the steady state energy equation is written as

∂ 2T    ∂u   ∂v   ∂u ∂v  
2 2 2
 ∂T ∂T ∂T   ∂ 2T
ρCv + u + v  = k 2 + v 2  + µ 2  + 2  +  +   Eq. 4.15
 ∂t ∂x ∂y   ∂x ∂y    ∂x   ∂y   ∂y ∂x  

where Cv is the constant volume specific heat and k is the thermal conductivity. The

terms on the right hand side of Equation 4.13 represent energy transfer by conduction and

viscous dissipation.

47
 CHAPTER V

EXPERIMENTAL RESULTS AND DISCUSSION

To establish guidelines for understanding the experimental results and their

anticipated consequences, some terminology will be introduced here.

The ratio of spindle speed to travel speed will be referred to as the weld pitch

(wp), and has units of revolutions per inch. The weld pitch can be increased in one of two

ways; 1) increasing the tool rotational speed, or 2) decreasing the travel speed. Likewise,

the weld pitch can be decreased by reducing tool rotational speed or increasing the travel

speed or feed rate.

The contact condition at the tool pin can be described as sliding, sticking, or

partial sliding/sticking. The experimental results will be presented with respect to the

effects due to weld pitch variation. Table 5-1 lists the weld pitch for the parameter

matrix in Table 3-1.

As stated in the introduction, a goal of this research is to establish guidelines for

implementing FSW capable robots. A significant limiting factor when implementing

FSW capable robots is the axial force requirement necessary when welding. In the

experimental results to be presented here, the translational force, transverse force, axial

force, and welding torque were measured for the parameter sets listed in Table 3-1. The

raw data plots can be seen in Appendix A.

48
 Table 5-1: VU FSW Experimental Parameter matrix Weld Pitch.

Feed Rate (ipm)

11.4 27 37.2 44.8 53.3 63.3

1500
132 56 40 33 28 24
Spindle Speed (RPM)

2250
175* 83 60 50 42 36

3000
219** 111 81 67 56 47

3750
- 139 101 84 70 59

4500
- - 100 84 71
*Actual Rotational Speed was 2000 RPM
** Actual Rotational Speed was 2500 RPM

Axial Force (Fz)

The axial force was measured for the weld parameter sets shown in Table 5-1. The

raw data plots can be seen in Appendix B. The steady state axial force is presented in the

following figures as the average axial force during the weld. Each weld parameter set

was run a minimum of two times in order to verify the precision of the force data. The

steady state axial force was found by averaging the mean axial force of each run of a

weld parameter. Figures 5-1 and 5-2 show the steady state axial force for variable

rotational speed and travel speed, respectively. From Figure 5-1, it can be seen that the

axial force decreases as the rotational speed increases.

49
 Axial Force vs Rotational Speed at Constant Travel Speed

10500

9500
A x ia l F o r c e (N ) (k N )

11 ipm
27 ipm
8500 37.2 ipm
44.8 ipm
53.3 ipm
63.3 ipm
7500

6500

5500
1500 2000 2500 3000 3500 4000 4500

Tool Rotational Speed (rpm)

Figure 5-1: Axial Force vs. Rotational Speed for constant travel speeds.

Increasing the rotational speed and holding the travel speed constant leads to a

decrease in axial force. Increasing the travel speed and holding the rotational speed

constant leads to an increase in axial force. The decrease in axial force for increasing

weld pitch through increased rotational speed is shown in Table 5-2.

50
 Table 5-2: Percentage Decrease in Axial Force for Increasing Weld Pitch
by increasing Rotational Speed

TS (ipm) Min wp (rpi) Max wp (rpi) % Fz Decrease

11.4 131.6 219.3 17.1
27.0 55.6 138.9 9.1
37.2 40.3 100.8 4.6
44.8 33.5 100.4 7.3
53.3 28.1 84.4 12.4
63.3 23.7 71.1 9.8

Figure 5-2 shows the axial force for various travel speeds at constant rotational

speeds. Table 5-3 shows the percentage decrease in axial force for increasing weld pitch

by decreasing the travel speed.

Axial Force vs. Travel Speed at Constant Rotational Speed

11000

10000

9000
A x ia l F o rc e (N )

8000

1500 RPM
7000
2250 RPM
3000 RPM
3750 RPM
6000 4500 RPM

5000
11.4 27 37.2 44.8 53.3 63.3

Travel Speed (ipm)

Figure 5-2: Axial Force vs. Travel Speed for constant rotational
speed.

51
 Table 5-3: Percentage Decrease in Axial Force for Increasing Weld Pitch
by reducing Travel Speed

RS (ipm) Min wp (rpi) Max wp (rpi) % Fz Decrease

1500 23.7 131.6 36.1
2250 35.5 197.4 41.0
3000 47.4 263.2 44.6
3750 59.2 138.9 33.7
4500 71.1 100.4 16.8

Welding Torque

The effect of weld pitch variation on the welding torque is key to understanding the

friction stir welding process and successfully implementing FSW capable robots. The

torque was measured for the weld parameter sets shown in Table 5-1. The raw data plots

can be seen in Appendix B.

The steady state torque is presented in Figures 5-3 and 5-4. The steady state welding

torque is found by averaging the mean torque each run of a weld parameter set. Figures

5-3 and 5-4 show the steady welding torque for variable rotational speed and travel speed

respectively. Tables 5-4 and 5-5 show the percentage decrease in torque from increasing

weld pitch.

52
 Moment vs. Rotational Speed for Constant Travel Speed
53

48
TS = 11.4
TS = 27
43 TS = 37.2
TS = 44.8
M o m e n t (N m )

TS = 53.3
38
TS = 63.3

33

28

23

18

13
1500 2250 3000 3750 4500

Rotational Speed (RPM)

Figure 5-3: Torque vs. Rotational Speed for constant travel speed.

Table 5-4: Percentage Decrease in Torque for Increasing Weld Pitch

by increasing Rotational Speed

TS (ipm) Min wp (rpi) Max wp (rpi) % Mz Decrease

11.4 131.6 219.3 45.9
27.0 55.6 138.9 60.0
37.2 40.3 100.8 59.3
44.8 33.5 100.4 72.2
53.3 28.1 84.4 66.1
63.3 23.7 71.1 67.9

53
 Moment vs. Travel Speed for Constant Travel Speed
55

50

45

40
M o m e n t (N m )

1500 RPM
35 2250 RPM
3000 RPM
30 3750 RPM
4500 RPM

25

20

15

10
11.4 27 37.2 44.8 53.3 63.3

Travel Speed (ipm)

Figure 5-4: Torque vs. Travel Speed for constant rotational Speed.

Increasing the rotational speed while holding the travel speed constant leads to a

decrease in torque; while increasing the travel speed and holding the rotational speed

constant leads to an increase in torque. The decrease in torque for increasing weld pitch

through reduced travel speed is shown in Table 5-5.

54
 Table 5-5: Percentage Decrease in Torque for Increasing Weld Pitch
by reducing Travel Speed

TS (ipm) Min wp (rpi) Max wp (rpi) % Mz Decrease

1500.0 23.7 131.6 36.0
2250.0 35.5 197.4 35.3
3000.0 47.4 263.2 32.6
3750.0 59.2 138.9 19.1
4500.0 71.1 100.4 18.5

The welding power is shown in Figure 5-5 and appears to remain nearly constant for a

constant travel speed.

Spindle Power vs. Rotational Speed for Constant Travel Speeds

10000

9000

8000
M o m e n t (N m )

7000

TS = 11.4
6000 TS = 27
TS = 37.2
TS = 44.8
TS = 53.3
5000 TS = 63.3 ipm

4000
1500 2250 3000 3750 4500
Rotational Speed (RPM)

Figure 5-5: Spindle Power vs. Rotational Speed for constant travel speed.

55
Translational and Transverse Force

The translational and transverse forces were measured for the weld parameter sets

shown in Table 5-1. The steady state translational and transverse forces are presented in

the following figures as the average force (Fx or Fy) during a weld. Each weld parameter

set was run a minimum of two times in order to verify the precision of the force data.

The steady state translational or transverse force is found by averaging the mean

translational or transverse force of each parameter set. Figures 5-7 and 5-8 show the

steady state translational and transverse force for variable rotational speed respectively.

From Figures 5-6 and 5-7, it is apparent that the translational and transverse forces

have the general trend of decreasing with increased weld pitch, but not with the linear

trend as the axial force and torque follow.

Translational Force vs Rotational Speed for Constant

Travel Speeds
140

120
Translational Force (N)

100

TS = 11.4 ipm
80 TS = 27 ipm
TS = 37.2 ipm
TS = 44.8 ipm
60 TS = 53.3 ipm
TS = 63.3 ipm

40

20

0
1500 2250 3000 3750 4500

Rotational Speed (ipm)

Figure 5-6: Translational Force vs. Rotational Speed for constant travel speed.

56
 Transverse Force vs Rotational Speed for Constant
Travel Speeds
90

80

70

60
Translational Force (N)

50
TS = 11.4 ipm

40 TS = 27 ipm
TS = 37.2 ipm
TS = 44.8 ipm
30
TS = 53.3 ipm
TS = 63.3 ipm
20

10

0
1500 2250 3000 3750 4500

-10

-20

Rotational Speed (ipm)

Figure 5-7: Transverse Force vs. Rotational Speed for constant travel speed.

The steady state plots of the translational and transverse forces follow the general

parametric relationship of increased rotational speed/ decreased force, however as the

rotational speed is increased, a new relationship becomes apparent.

Viewing the raw data plots gives insight into the force behavior at higher rotational

speeds. Figures 5-8 and 5-9 show the raw data plots of the translational and transverse

force for various rotational speeds and 44 ipm travel speed. Increasing the rotational

speed for a constant travel speed creates a varying contact condition at the tool

pin/material interface.

57
 Translational Force for Variable RS and Constant TS = 44.8ipm
200
1500 RPM
2250 RPM
180
3000 RPM
3750 RPM
160 4500 RPM

140

120

100
Fx (N)

80

60

40

20

-20
5 10 15 20 25
Time (s)

Figure 5-8: Raw data plot of Translational Force for various RS and TS = 44 ipm.

Transverse Force for Variable RS and Constant Ts = 44.8 ipm

1500 RPM
2250 RPM
150 3000 RPM
3750 RPM
4500 RPM

100

50
Fy (N)

-50

-100
5 10 15 20 25
Time (s)

Figure 5-9: Raw data plot of Transverse Force for various RS and TS = 44 ipm.

58
 In Figures 5-9 and 5-10, the translational and transverse forces are constant lines

of force for wp < 50.2 rpi and oscillates for wp > 50.2 rpi. The constant lines of force

indicate a constant pressure at the tool pin/material interface. The oscillation indicates

that the material at the tool pin/material interface does not apply constant pressure but

rather it sticks to the tool and drags along behind the tool as it rotates.

With the sticking contact condition, if the friction shear stress exceeds the yield

shear stress, the weld material at the tool/material interface will stick to the moving tool

surface segment. In this case, the matrix segment will accelerate along the tool surface

(finally receiving the tool velocity), until an equilibrium state is established between the

contact shear stress and the internal matrix shear stress. At this point, the stationary full

sticking condition is fulfilled [41].

For the sliding condition, if the contact shear stress is smaller than the internal

matrix yield shear stress, the matrix segment volume shears slightly to a stationary elastic

deformation, where the shear stress equals the ‘dynamic’ contact shear stress. This state

is referred to as the sliding condition [41].

The partial sliding/sticking contact condition is a mixed state of the two contact

conditions. In this case, the matrix segment accelerates to a velocity less than the tool

surface velocity, where it stabilizes. The equilibrium occurs when the ‘dynamic’ contact

shear stress equals the internal yield shear stress due to a quasi-stationary plastic

deformation rate [41].

The variation of the contact condition can reasonably be assumed to be induced

by increasing the rotational speed. Increasing the rotational speed causes a corresponding

increase in welding temperature. The over-heat phenomena (discussed in Chapter 3) that

59
occurred at certain welding parameter sets was always preceded by a sliding/sticking

contact condition for the lower weld pitch parametric set (rotational speed and travel

speed).

The raw data plots in Appendix B show that the contact condition is sliding for

wp < 50.2 rpi, and that the force is due to a pressure. For the for wp > 50.2, the contact

condition would appear to be sliding/sticking and the force on the pin is due to visco-

plastic drag of the weld material.

Understanding these conditions is key developing a three dimensional model

capable of predicting the forces and torques during FSW for various weld pitches and

tool geometries. Chapter 6 details the results of the two dimensional fluid flow model

discussed in Chapter 4 and is compared to the experimental results presented here.

60
 CHAPTER VI

2-D MODELING RESULTS AND DISCUSSION

It will be useful now to identify the limiting factors for implementation of FSW

capable robots. The primary limiting factor is the large axial force required.

The axial force was found to decrease by either increasing rotational speed or

decreasing travel speed. An increase in rotational speed decreases the axial force by

increasing the heat input into the weld material, thus raising the temperature of the weld

material. Reducing the travel speed increases the number of revolutions per unit length

of the weld, which increases the heat input per unit length of the weld.

Table 3-3 shows that increasing the temperature of the weld material from 38ºC to

371ºC decreases the yield strength from 241 MPa to 12 MPa, a 95% decrease. Therefore

the optimum weld pitch for FSW will occur at high rotational speeds and low travel

speeds.

Another potential limiting factor is the welding torque. The welding torque is

largely governed by the weld pitch as well as the tool geometry. Increasing the size of

the tool pin and particularly the tool shoulder, causes a corresponding increase in welding

torque. Increasing the rotational speed increases the temperature of the weld material,

decreasing the yield strength, which decreases the torque required to displace the weld

material to facilitate forward travel of the tool. As stated earlier, reducing the travel

speed increases the heat input to the weld, and reduces the yield strength, which reduces

the required torque.

61
 A significant goal of this research is to establish practical design guidelines for

developing FSW capable robots. Developing a 3-D model capable of predicting the

forces and torques associated with FSW would greatly aide the ability of scientists and

engineers to design, fabricate, and implement FSW capable robots.

In this thesis, a two dimensional model was developed to predict the translational

force, welding torque, and temperature on the tool pin for parameter sets listed in Table

5-1. All simulations were run using the computational fluid dynamics package FLUENT.

Chapter 4 details the determination of the material properties, boundary conditions,

mechanical modeling, thermal modeling, solver configuration and governing equations.

Welding Torque

Friction stir welding is a three dimensional process. In this thesis, the initial

modeling efforts are represented in 2-D. Though a 2-D model cannot fully represent a 3-

D process, if implemented correctly, the 2-D model may suggest the general trends of the

3-D process.

As stated in Chapter 4, the tool is represented by a 2-D rotating pin. In order for

the torque experimental results to be compared with the numerical model, the

experimental results must be scaled to represent the contribution of the pin during

welding. The rotating plug model suggested by Nunes et al. [26] was used to determine

the pin contribution.

With the rotating plug model, the tool torque is taken to be totally due to the shear

flow stress of the metal acting perpendicular to the direction of tool rotation and

integrated over the surface. The tool torque is given by

62
 Rs Rp
M z = ∫ 2πR p σd R p + 2πR p2 tσ + ∫ 2πR p2σd R p , Eq. 6.1
Rp 0

where Rp is the pin radius, Rs is the shoulder radius, σ is the shear flow stress, and t is the

pin depth. The three terms in Equation 6.1 are the contributions to the torque from the

tool shoulder, pin sides, and pin bottom respectively. Applying Equation 6.1 to the

experimentally measured steady state welding torque, the shear stress was computed for

each parameter set in Table 5-1, and can be seen in Table 6-1.

Table 6-1: Computed Shear Stress using the rotating plug model.

Feed Rate (ipm)

σ (Pa) σ (Pa) σ (Pa) σ (Pa) σ (Pa) σ (Pa)
RPM @ 11.4 @ 27 @ 37.2 @ 44.8 @ 53.3 @ 63.3
1500 7.92E+07 4.92E+08 5.58E+08 6.56E+08 7.80E+08 9.21E+08
2250 5.14E+07 1.35E+08 2.84E+08 3.13E+08 3.67E+08 4.58E+08
3000 2.63E+07 5.80E+07 1.15E+08 1.68E+08 2.34E+08 2.90E+08
3750 4.12E+07 6.19E+07 1.05E+08 1.47E+08 1.56E+08
4500 5.07E+07 8.97E+07 9.38E+07

63
From Equation 6.1, we see that the torque contribution from the pin side is

M z = 2πR p2 tσ . Eq. 6.2

Now, substituting the shear stress from Table 6-1 into Equation 6.2, Mp is found for the

corresponding parameter sets and can be seen in Table 6-2.

Table 6-2: Torque Contribution by the Tool Pin.

Feed Rate (ipm)

Mp (N-m) Mp (N- Mp (N-m) Mp (N-m) Mp (N-m) Mp (N-m)

RPM @ 11.4 m)@ 27 @ 37.2 @ 44.8 @ 53.3 @ 63.3

1500 8.3 10.4 11.4 12.2 12.7 12.9

2250 6.0 7.1 7.6 8.4 9.0 9.3

3000 4.7 5.3 5.7 6.1 6.4 7.0

3750 - 4.2 4.6 5.0 5.1 5.2

4500 - - - 3.4 4.3 4.2

Dividing Mp /Mz for the corresponding parameter sets show that the pin torque

contribution is approximately 24.9%.

The simulations were run for the Couette Flow and Visco-Plastic Flow viscosity

models. The simulation determined the shear stress at the pin wall, which was then input

into Equation 6-2 to determine the moment (welding torque) of the pin. Figures 6-1

through 6-6 show a comparison of the predicted simulation results to the experimental

pin torques for various weld pitches.

64
 At low weld pitches, the Couette Flow model did not correlate very well with the

experimental results. As the weld pitch increases, the experimental results and the

Couette Flow model begin to converge. This implies that the Coutte Flow model is more

predictive for very high weld pitches. At high weld pitches, the high heat input greatly

improves the weld material’s ability to flow. In general, the torque decreases as weld

pitch is increased.

Overall, the Visco-Plastic flow model was more accurate than the Couette Flow

model over the range of weld pitches. The Visco-Plastic Flow model also converged

with the experimental results as the weld pitch was increased.

Pin Moment vs. Rotational Speed for TS = 11.4 ipm

45

40

Experimental
35
Couette
Visco-Plastic
30
M o m e n t (N m )

25

20

15

10

0
1500 2250 3000

Rotational Speed (RPM)

Figure 6-1: Comparison of Predicted and Experimental Pin Moment vs. Rotational
Speed for 11.4 ipm

65
 In Figure 6-1, the Couette Flow model correlates very well for the 11.4 ipm travel

speed. The weld pitch at this travel speed was very high for all rotational speeds.

Lending further credibility to the theory that Couette Flow is more optimal for high weld

pitches.

Pin Moment vs. Rotational Speed for TS = 27 ipm

70.0

60.0 Experimental
Couette
Visco-Plastic
50.0
M o m e n t (N m )

40.0

30.0

20.0

10.0

0.0
1500 2250 3000 3750

Rotational Speed (RPM)

Figure 6-2: Comparison of Predicted and Experimental Pin Moment vs. Rotational
Speed for 27 ipm

At approximately 2300-2400 rpm, the Couette Torque and Experimental torque

are almost equal. The corresponding weld pitch is 202- 210 rpi. In Figures 6-2 through

6-6, the Couette torque is never less than the experimental torque, likewise, the weld

pitch is not higher than 200 rpi for the following plots

66
 Pin Moment vs. Rotational Speed for TS = 37.2 ipm
80.0

70.0

60.0

50.0
M o m e n t (N m )

Experimental
40.0 Couette
Visco-Plastic

30.0

20.0

10.0

0.0
1500 2250 3000 3750

Rotational Speed (RPM)

Figure 6-3: Comparison of Predicted and Experimental Pin Moment

vs. Rotational Speed

Pin Moment vs. Rotational Speed for TS = 44.8 ipm

100.0

90.0

80.0 Experimental
Couette
70.0 Visco-Plastic
M om ent (N m )

60.0

50.0

40.0

30.0

20.0

10.0

0.0
1500 2250 3000 3750 4500
Rotational Speed (RPM)

Figure 6-4: Comparison of Predicted and Experimental Pin Moment vs.

Rotational Speed for 44.8 ipm

67
 Pin Moment vs. Rotational Speed for TS = 53.3 ipm
120.0

100.0

Experimental
Couette
80.0
Visco-Plastic
M om en t (N m )

60.0

40.0

20.0

0.0
1500 2250 3000 3750 4500
Rotational Speed (RPM)

Figure 6-5: Comparison of Predicted and Experimental Pin

Moment vs. Rotational Speed for 53.3 ipm

Pin Moment vs. Rotational Speed for TS = 63.3 ipm

140.0

120.0
Experimental
Couette
100.0 Visco-Plastic
M o m e n t (N m )

80.0

60.0

40.0

20.0

0.0
1500 2250 3000 3750 4500

Rotational Speed (RPM)

Figure 6-6: Comparison of Predicted and Experimental Pin

Moment vs. Rotational Speed for 63.3 ipm

68
Translational Force

Figures 6-7 through 6-12 show the comparison of the experimental translational

force to the predicted translational forces for the Couette Flow and the Visco-Plastic flow

model. In general the results follow the increased rotational speed/decreased force

relationship.

Translational Force vs. Rotational Speed for TS = 11.4 ipm

140

120

100
Experimental
Couette
80 Visco-Plastic
F x (N )

60

40

20

0
1500 2250 3000

Rotational Speed (RPM)

Figure 6-7: Comparison of Predicted and Experimental Translational Force vs.
Rotational Speed for 11.4 ipm.

69
 Translational Force vs. Rotational Speed for TS = 27 ipm
300

250

200
Experimental
Couette
Fx (N)

150 Visco-Plastic

100

50

0
1500 2250 3000 3750

Rotational Speed (RPM)

Figure 6-8: Comparison of Predicted and Experimental Translational Force vs.

Rotational Speed for 27 ipm.

Translational Force vs. Rotational Speed for TS = 37.2 ipm

400

350

300 Experimental
Couette
Visco-Plastic
250
F x (N )

200

150

100

50

0
1500 2250 3000 3750

Rotational Speed (RPM)

Figure 6-9: Comparison of Predicted and Experimental Translational Force

vs. Rotational Speed for 37.2 ipm.

70
 Translational Force vs. Rotational Speed for TS = 44.8 ipm
500

Experimental
450
Couette
Visco-Plastic
400

350

300
F x (N )

250

200

150

100

50

0
1500 2250 3000 3750 4500

Rotational Speed (RPM)

Figure 6-10: Comparison of Predicted and Experimental Translational

Force vs. Rotational Speed for 44.8 ipm.

Translational Force vs. Rotational Speed for TS = 53.3 ipm

700

600
Experimental
Couette
Visco-Plastic
500

400
Fx (N )

300

200

100

0
1500 2250 3000 3750 4500
Rotational Speed (RPM)

Figure 6-11: Comparison of Predicted and Experimental Translational

\ Force vs. Rotational Speed for 53.3 ipm.

71
 Translational Force vs. Rotational Speed for TS = 63.3 ipm
800

700

600

Experimental
500 Couette
Visco-Platic
F x (N )

400

300

200

100

0
1500 2250 3000 3750 4500

Rotational Speed (RPM)

Figure 6-12: Comparison of Predicted and Experimental Translational Force

vs. Rotational Speed for 63.3 ipm

The trend for the translational force seems to match the weld torque where the

Coutte Flow model is more accurate for very high weld pitches and the Visco-Plastic

flow model was more precise than the Couette Flow model over the range of weld

pitches.

Temperature Predictions

Figures 6-13 show the predicted temperature of the weld material at the tool

pin/material interface. The difference in temperature for the Couette Flow Model and the

Visco-Plastic flow model were less than 1% for the various weld pitches. Figure 6-13

72
shows that the temperature increases as the tool rotational speed increases. This further

supports the observation that the axial force decreases with increased rotational speed

because the weld material temperature is higher and decreases the yield strength. The

maximum predicted temperature is shown to occur at the highest rotational speed (4500

rpm), while the lowest temperature occurs at the lowest rotational speed of 1500 rpm.

Figure 6-14 shows the effect of travel speed on the weld temperature. The

predicted temperature decreases as the travel speed is increased.

Predicted Temperature vs. Rotational Speed

700

650

600

550
T e m p e ra tu re ( o C )

500

450

400 TS =11.4
TS = 27
TS = 37.2
350
TS = 44.8
TS = 53.3
300
TS = 63.3

250

200
1500 2250 3000 3750 4500

Rotational Speed (RPM)

Figure 6-13: Predicted Temperature vs. Rotational Speed for Constant Travel Speed.

73
 Predicted Temperature vs. Travel Speed

700

650 1500 RPM

2250 RPM
3000 RPM
600
3750 RPM
4500 RPM
550
T e m p e ra tu re (o C )

500

450

400

350

300

250

200
11.4 27 37.2 44.8 53.3 63.3
Travel Speed (ipm)

Figure 6-14: Predicted Temperature vs. Travel Speed for Constant Rotational Speed.

The predicted temperatures at 4500 rpm seem high. One possible reason is

because the contact condition at the tool pin material interface was assumed to be

sticking, while the experiments that the contact condition is partial sliding/sticking. More

experimentation at increased temperatures is necessary to improve the analysis. The heat

generation model used in this simulation has provisions for different contact conditions at

the tool pin/material interface. Future research will involve modifying the contact

condition used for determining the heat input during welding.

The steady state welding temperature was not measured during the experiment.

Therefore there is no way at this time to validate the predicted temperatures. However,

the results correlate well with the weld temperatures (up to 547 ºC) predicted by Song et

74
al. [44]. Similar temperatures were predicted by Sato et al. for Al 6063[18], which is

similar to Al 6061 in chemical composition and physical properties [44].

Figures 6-15, 6-16, and 6-17 show plots of the weld material temperature

gradients for 1500 rpm / 11.4 ipm, 2000 rpm / 11.4ipm, and 2500 rpm/11.4 ipm. As can

be seen in Figures 6-15 and 6-16, the contour plots of temperature profile are similar for

each parameter set. Therefore only the pin temperatures are presented here and can be

seen in Figure 6-13.

Figure 6-15: Temperature Contour Plot for 1500 RPM and 11.4 ipm.

75
Figure 6-16: Temperature Contour Plot for 2000 RPM and 11.4 ipm.

Figure 6-17: Temperature Contour Plot for 2500 RPM and 11.4 ipm.

76
The contour temperature plots show that as rotational speed is in creased, the preheating

temperature of the incoming weld material increases.

Velocity Profile

The velocity profile can be used to understand material flow during FSW. As

rotational speed is increased, the velocity of the weld material is expected to increase.

On the advancing side, the velocities are additive, while on the retreating side, the

velocities are subtractive. Figures 6-18, 6-19, and 6-20 show the velocity magnitude

profiles for the tool pin for 1500 rpm/11.4 ipm, 2000 rpm/11.4ipm, and 2500 rpm/11.4

ipm.

Figure 6-18: Pin Velocity Profile for 1500 rpm/ 11.4 ipm.

77
Figure 6-19: Pin Velocity Profile for 2000 rpm/ 11.4 ipm.

Figure 6-20: Pin Velocity Profile for 2500 rpm/ 11.4 ipm.

78
 In general, the velocity profile for 1500 rpm and a given travel speed are very

similar to Figure 6-18 for the Couette and Visco-Plastic flow mode. As rotational speed

is increased, the velocity profiles wee very similar to Figures 6-19 and 6-20.

The velocity profiles presented here resemble the shape of the third body region

proposed by Abregast et al. [38]. However, as rotational speed is increased, the profile

begins to balance and is not skewed toward to advancing side as shown in Figures 6-18.

It should be noted that the velocity profile for 1500 rpm/11.4 ipm does not resemble the

Couette Flow diagram in Figure 4-1. However, as the velocity profiles in Figures 6-19

and 6-20 become more concentric with the tool pin, as assumed by the Coutte Flow

model (Figure 4-1), the Couette Flow predictions improve.

79
 CHAPTER VII

CONCLUSION AND RECCOMMEDATIONS

Experimental

The experimental data shows that the increased rotational speed/decreased force

relationship exists for rotational speeds ranging from 1500 rpm to 4500 rpm, and for

travel speeds from 11.4 ipm to 63.3 ipm. The percentage decreases in axial force for

increasing weld pitch can be seen in Tables 5-2 and 5-3. Reducing the travel speed was

observed to be a more effective method for reducing axial force. Increasing rotational

speed was observed to be a viable option for reducing axial force as well. However, the

present limitation to fully exploiting these relationships is the overheating of the weld

material observed at high weld pitches. Appendix C shows photos of welds which

experienced overheating and subsequent surface deformation. An investigation into

possible methods for extracting heat away from the weld material will be necessary for

high weld pitch FSW to be achieved. Another possible solution would be the

implementation of force feedback control to high weld pitch FSW setups.

The experimental data shows that the increased rotational speed/decreased torque

relationship exists for rotational speeds ranging from 1500 rpm to 4500 rpm, and for

travel speeds from 11.4 ipm to 63.3 ipm. The percentage decreases in torque for

increasing weld pitch can be seen in Tables 5-4 and 5-5. Increasing rotational speed was

observed to be a more effective method for reducing the welding torque. Reducing travel

speed was observed to be a viable option for reducing the welding torque as well.

80
 The steady state plots of the translational and transverse forces follow the general

parametric relationship of increased rotational speed/ decreased force, however as the

rotational speed is increased, a new relationship becomes apparent.

Increasing the rotational speed for a constant travel speed creates a varying

contact condition at the tool pin/material interface. As the rotational speed is increased,

the contact condition at the tool pin transitions from sticking to partial sliding/sticking.

The transition of the contact condition was observed to signal that the process parameters

were approaching a critical weld pitch. The critical weld pitch is the weld pitch at which

an increase in weld pitch led to the over-heat phenomena.

The welding power was observed to be nearly constant for a constant travel speed,

which correlates well with the rotating plug model suggested by Nunes et al. [25].

2-D Model

At low weld pitches, the Couette Flow model did not correlate very well with the

experimental results. As the weld pitch increases, the experimental results and the

Couette Flow model begin to converge. This implies that the Coutte Flow model is more

predictive for very high weld pitches.

Overall, the Visco-Plastic flow model was more accurate than the Couette Flow

model over the range of weld pitches. The Visco-Plastic Flow model also converged

with the experimental results as the weld pitch was increased.

The steady state welding temperature was not measured during the

experiment. However, the results correlate well with the weld temperatures (up to 547

ºC) predicted by Song et al. [44]. Similar temperatures were predicted by Sato et al. for

81
Al 6063[45], which is similar to Al 6061 in chemical composition and physical properties

[44].

The velocity profiles presented here resemble the shape of the third body region

proposed by Abregast et al. [38]. However, as rotational speed is increased, the profile

begins to balance and is not skew towards to advancing side as shown in Figure 6-18. It

should be note that the velocity profile for 1500 rpm/11.4 ipm does not resemble the

Couette Flow diagram in Figure 4-1. However, as the velocity profiles in Figures 6-19

and 6-20 begin to resemble the Coutte Flow model (Figure 4-1), the Couette Flow

predictions improve.

82
 APPENDIX A

Travel Speed = 11.4 ipm

Axial Force for Variable RS and Constant TS = 11.4 ipm

12000
1500 RPM
2000 RPM
2500 RPM
10000

8000

6000
Fz (N)

4000

2000

-2000
10 20 30 40 50 60
Time (s)

Moment for Variable RS and Constant TS = 11.4 ipm

1500 RPM
60
2000 RPM
2500 RPM
50

40
M z (N -m )

30

20

10

10 20 30 40 50 60
Time (s)

83
 Translational Force for Variable RS and Constnat TS = 11.4 ipm, z =0.145"
300
1500 RPM
2000 RPM
250 2500 RPM

200

150

100
Fx (N)

50

-50

-100

-150
0 10 20 30 40 50 60 70
Time (s)

Transverse Force for Variable RS and Constant TS = 11.4 ipm

1500 RPM
2000 RPM
2500 RPM

200

150

100
F y (N )

50

-50

0 10 20 30 40 50 60
Time (s)

84
Travel Speed = 27 ipm

Axial Force for Variable RS and Constant TS = 27 ipm

1500 RPM
12000
2250 RPM
3000 RPM
3750 RPM
10000

8000
F z (N )

6000

4000

2000

5 10 15 20 25 30 35 40
Time (s)

Moment for Variable RS and Constant TS = 27 ipm

1500 RPM
60
2250 RPM
3000 RPM
3750 RPM
50

40
M z (N -m )

30

20

10

5 10 15 20 25 30 35 40
Time (s)

85
 Translational Force for Variable RS and Constant TS = 27 ipm

1500 RPM
2250 RPM
3000 RPM
250
3750 RPM

200

150
Fx (N)

100

50

-50

-100
5 10 15 20 25 30 35 40
Time (s)

Transverse Force for Variable RS and Constant TS = 27 ipm

1500 RPM
200 2250 RPM
3000 RPM
3750 RPM
150

100

50
F y (N )

-50

-100

5 10 15 20 25 30 35 40
Fx (N)

86
Travel Speed = 37 ipm

Axial Force for Variable RS and Constant TS = 37.2 ipm

10000
1500 RPM
2250 RPM
9000
3000 RPM
3750 RPM
8000

7000

6000

5000
Fz (N)

4000

3000

2000

1000

-1000
5 10 15 20 25 30
Time (s)

Momnet for Variable RS and Constant TS = 37.2 ipm

1500 RPM
60 2250 RPM
3000 RPM
3750 RPM

50

40
Mz (N-m)

30

20

10

-10
5 10 15 20 25 30
Time (s)

87
 Translational Force for Variable RS and Constant TS = 37.2 ipm
250
1500 RPM
2250 RPM
3000 RPM
200 3750 RPM

150

100
Fy (N)

50

-50

-100
0 5 10 15 20 25 30
Time (s)

Transverse Force for Variable RS and Constant TS = 37.2 ipm

250
1500 RPM
2250 RPM
3000 RPM
200 3750 RPM

150

100
Fy (N)

50

-50

-100
0 5 10 15 20 25 30
Time (s)

88
Travel Speed = 44.8 ipm

Axial Force for Variable RS and Constant T S = 44.8 ipm

1500 RPM
2250 RPM
3000 RPM
12000 3750 RPM
4500 RPM

10000

8000
Fz (N)

6000

4000

2000

0
5 10 15 20 25
Time (s)

Moment for Variable RS and Constant T S = 44.8 ipm

70
1500 RPM
2250 RPM
60 3000 RPM
3750 RPM
4500 RPM

50

40
Mz (N-m)

30

20

10

-10
5 10 15 20 25
Time (s)

89
 Translational Force for Variable RS and Constant TS = 44.8ipm
200
1500 RPM
2250 RPM
180
3000 RPM
3750 RPM
160 4500 RPM

140

120

100
Fx (N)

80

60

40

20

-20
5 10 15 20 25
Time (s)

Transverse Force for Variable RS and Constant Ts = 44.8 ipm

1500 RPM
2250 RPM
150 3000 RPM
3750 RPM
4500 RPM

100

50
Fy (N)

-50

-100
5 10 15 20 25
Time (s)

90
Travel Speed = 53.3 ipm

Axial Force for Variable RS and Constant TS = 53.3 ipm

14000

12000

10000

8000
F z (N )

6000

4000

1500 RPM
2000 2250 RPM
3000 RPM
3750 RPM
0 4500 RPM

2 4 6 8 10 12 14 16 18 20
Time (s)

Moment for Variable RS and Constant TS = 53.3 ipm

60

50

40

30
Mz (N-m)

20

10

1500 RPM
2250 RPM
0
3000 RPM
3750 RPM
4500 RPM

-10
2 4 6 8 10 12 14 16 18 20
Time (s)

91
 Translational Force for Variable RS and Constant TS = 53.3 ipm
100
1500 RPM
2250 RPM
80
3000 RPM
3750 RPM
60 4500 RPM

40
F x (N )

20

-20

-40

-60
2 4 6 8 10 12 14 16 18 20 22
Time (s)

Transverse Force for Variable RS and Constant TS = 53.3 ipm

80

60

40

20

0
Fy (N)

-20

-40

-60

-80 1500 RPM

2250 RPM
3000 RPM
-100 3750 RPM
4500 RPM

-120
0 5 10 15 20 25
Time (s)

92
Travel Speed = 63.3 ipm

Axial Force vs Rotational Speed for COnstant TS = 63.3 ipm

15000
1500 RPM
2250 RPM
3000 RPM
3750 RPM
4500 RPM

10000
F z (N )

5000

0
2 4 6 8 10 12 14 16 18
Time (s)

Moment vs. Rotational Speed for Constant TS = 63.3 ipm

70 1500 RPM
2250 RPM
3000 RPM
60 3750 RPM
4500 RPM

50

40
M z (N -m )

30

20

10

-10
0 2 4 6 8 10 12 14 16 18
Time (s)

93
 Translational Force vs. Rotational Speed for TS = 63.3 ipm
200
1500 RPM
2250 RPM
3000 RPM
150
3750 RPM
4500 RPM

100

50
Fx (N)

-50

-100

-150

-200
0 2 4 6 8 10 12 14 16 18
Time (s)

Transverse Force vs. Rotational Speed for Constant TS = 63.3 ipm

150

100

50
Fy (N)

-50

0 2 4 6 8 10 12 14 16 18
Time (s)

94
 APPENDIX B

Deformed Weld for 3000 rpm and 27 ipm.

Deformed Weld for 3750 rpm and 27 ipm

95
Deformed Weld for 3750 rpm and 37.2 ipm

Deformed Weld for 3750 rpm and 44.8 ipm

96
Deformed Weld for 4500 rpm and 44.8 ipm

97
 APPENDIX C

MATLAB CODE TO POST EXPERIMENTAL DATA

load F1.txt
tx = F1(:,1);
Fx1 = F1(:,2);
l = length(Fx1);
for i = 1:l
Fx1(i) = Fx1(i)/1;
end
n = input('Filter Scale: ');
for i = 1:n
sum = 0;
for j = 1:i
sum = sum + Fx1(j);
end
Fx1smth(i) = sum / i;
end
for i = (n+1):l
sum = 0;
for j = 0:(n-1)
sum = sum + Fx1(i-j);
end
Fx1smth(i) = sum / n;

end

98
 REFERENCES

[1] Thomas W.M., et al. 1991, Friction Stir Butt Welding. International Patent
Application # PCT/GB92/02203, GB Patent Application #9125978.8, and GB Patent #
2,306,366.

[2] http://www.twi.co.uk/j32k/unprotected/band_1/fswjoint.html

[3] Nicholas E. D., and Kallee, S.W., 2000, “Friction Stir Welding – A decade on.” IIW
Asian Pacific International Congress. Sydney 29 October to 2 November

[4] Heideman R. J. et al., 2000, Friction Stir Welding Tool. U.S. Patent #6,053,391

[5] Carter, Robert W., “Pin Tool Design” Proceedings of the NASA MSFC Friction Stir
Welding Workshop and Showcase, Huntsville, Al, September 2003

[6] Thomas, W.M., et al., 1995, Friction Welding. U.S. Patent # 5,460,317

[7] Cocks, E. E., et al., 2000, Enantiomorphic Friction-Stir Welding. U.S. Patent #
6,029,879

[8] Su, J. Q., 2003, Acta Materialia. 51 713-729

[9] Mahoney, M.W., Rhodes, C.G., Flintoff, J.G., Spurling, R.A., and Bingel, W.H.,
1998, Metall. Mater. Trans. A. 29A (1998) 1955-1964.

[10] Heinz, B., Skrotzki, B., and Eggler, G., 2000, Mater. Sci. Forum. 331–337 1757-
1762.

[11] Svensson, L.E., Karlsson, L., Larsson, H., Karlsson, B., Fazzini, M., and Karlsson,
J., 2000, Sci. Technol. Welding Joining. 5 285-296.

[12] Reynolds, A.P., Lockwood, W.D., and Seidel, T.U., 2000, Mater. Sci. Forum. 331–
337 1719-1724.

[13] Lumsden, J.B., Mahoney, M.W., Pollock, G. and Rhodes, C.G., 1999, Corrosion. 55
1127-1135.

[14] Murr, L.E., Liu, G. and McClure, J.C., 1998, Mater. Sci.. 33 1243-1251.

[15] Rhodes, C.G., Mahoney, M.W, Bingel, W.H., Spurling, R.A., and. Bampton, C.C.,
1997, Scripta Mater. 36 69-75.

99
[16]. Jata, K.V., Sankaran, K.K, and Ruschau J.J., 2000, Metall. Mater. Trans. A, 31A
2181-2192.

[17] Liu, G., Murr, L.E., Niou, C.S., McClure, J.C., and Vega, F.R., 1997, Scripta Mater.
37 355-361.

[18] Sato, Y.S., Kokawa, H., Enomoto M., and Jogan, S., 1999, Metall. Mater. Trans. A,
30A 2429-2437.

[19] Su, J.-Q., Nelson, T.W., Mishra, R., and Mahoney, M., 2003, Acta Mater. 51 713-
729.

[20] Kwon Y.J., Shigematsu I., Saito N., 2003, Scripta Materialia. 49 785-789

[21] Lee W. B., Yeon Y. M., Jung S. B., 2003, Materials Science and Technology. 19
1513 - 1518

[22] Ulysse, P., 2002, International Journal of Machine Tools & Manufacture. 42 1549–
1557

[23] Johnson W., Kudo H.K., The Mechanics of Extrusion., 1962, Manchester
University Press, UK P.40

[24] FIDAP, Fluid Dynamic Analysis Package, version 7.6, Fluid Dynamics
International, Evanston, IL. 2002

[25] Reynolds A P, Deng X, Seidel T and Xu S., 2000, Proc. Joining of Advanced and
Specialty Materials (St Louis, MO, ASM International) pp 172–177

[26] Nunes A.C., Bernstien, E.L., and McClure, J.C., A rotating Plug Model for Friction
Stir Welding. Paper presented at the 81st American Welding Society annual Convention,
Chicago, IL, April 26-28, 2000

[27] Mitchell, Jason. Experimental Thermo-mechanics of Friction Stir Welding. M.S.

Thesis, Vanderbilt University, 2002

[28] Guerra M., Schmidt C., McClure J.C., Murr, L.E., Nunes A.C., 2003, Materials
Characterization. 49 95-101

[29] Colligan, K., 1999, Welding Journal, 78 229 –237.

[30] Colegrove P. 2000, 3-Dimensional Flow and Thermal Modeling of the Friction Stir
Welding Process. Paper presented at the 2nd Int. Symp. on Friction Stir Welding,
Gothenburg, Sweden, June 27-29, 2000

100
[31] Shercliff H. R., and Colgrove P. A., 2002, Math. Modelling Weld. Phenom. 6 927–
974

[32] Chen C.M., Kovacevic R., 2003, International Journal of Machine Tools &
Manufacturing 43 1319–1326

[33] http://www.memagazine.org/medes03/coolweld/coolweld.html

[34] Thomas, W.M., Threadgill P.L., and Nicolas E.D., 1999, Science and Technology of
Welding and Joining. 4(6) 365-372

[35] Mills, K.C., ‘Recommended Values of thermophysical properties for commercial

alloys’, 68-71; 2002, Woodhead Publishing Ltd., Cambridge, UK

[36] North T.H. North, G.J. Bendzsak and C. Smith. Material Properties Relevant to 3-D
FSW Modeling. Paper presented at the 2nd Int. Symp. on Friction Stir Welding,
Gothenburg, Sweden, June 27-29, 2000

[37] http://scienceworld.wolfram.com/physics/RayleighCriterion.html

[38] G.F. Braun and W.J. Arbegast, "Aluminum Flow Stress Determinations Using A
Gleeble System", presented at the 11th Annual Advanced Aerospace Materials and
Processes Conference and Symposium, ASM International,
Seattle WA, March 14-15, 2000

[39] Chao, Y. J., and Qi, X., 1999, 1st Int. Symp. on Friction Stir Welding (Thousand
Oaks, CA, USA)

[40] Russell M J and Shercliff H R 1999 1st Int. Symp. on Friction Stir Welding
(Thousand Oaks, California, USA)

[41] Schmidt, H., Hattel, J., Wert, J. 2004, Modeling and Simulation in Materials Science
and Engineering. 12 143–157

[42] Sheppard T., Jackson A. 1997, Mater. Sci. Technol.13 203-209.

[43] FLUENT, Fluid Dynamic Analysis Package, version 6.122 Fluid Dynamics
International, Evanston, IL. 2004

[44] Song M., and Kovacevic R. 2003, Mach. Tools Manuf. 43 605–15

[45] White, F., Kudo H.K., Viscous Fluid Flow, 1991, McGraw-Hill P.112

101