High Temperature Tribology in Hot...

59
High Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's level (120 credits) 2019 Luleå University of Technology Department of Engineering Sciences and Mathematics

Transcript of High Temperature Tribology in Hot...

Page 1: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

High Temperature Tribology

in Hot Stamping

Evan Kurnia

Mechanical Engineering, master's level (120 credits)

2019

Luleå University of Technology

Department of Engineering Sciences and Mathematics

Page 2: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

i

Preface

The work presented in this master thesis has been carried out at the Department of Engineering

Sciences and Mathematics, Division of Machine Elements, Luleå University of Technology. This work has

been a part of projects in cooperation with Gestamp Hardtech AB and Volvo Car Corporation.

First of all, I would like to thank my supervisor Associate Professor Jens Hardell for his guidance

and in-depth discussions. I would like to thank Dr. Leonardo Pelcastre for the training and supervision with

hot-strip tribometer and profilometer. I also thank Justine Decrozant-Triquenaux who listened to my

progress of this work. To Daniel Strömbergsson, I thank you for your training and discussions regarding

acoustic emission, the equipment, and Matlab coding for signal processing.

Finally, for my parents, I really thank you for all your support. I really appreciate that you listened

and even discussed my work although you are not an engineer. I thank you for listening to me, both good

and bad news. Your support is more than just the scope of this work, it is a life lesson which I know will

be useful in a bigger scope of life. I will always remember that you said: “You can find many smart people,

but you will find very few trustworthy people”. I will always remember to remain a gentleman no matter

how hard the challenge will be. Again, thank you very much for being great and supporting parents. Terima

kasih banyak.

Evan Kurnia, Luleå, July 2019

Page 3: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

ii

Abstract

Many automotive components are made of Al-Si coated ultra-high strength boron steel (UHSS)

and are produced by hot stamping process. In this process, the workpiece is heated to an austenitizing

temperature and is then formed and quenched simultaneously between the tools to achieve the desired

shape and high strength. During hot stamping process, friction and wear occur which affect formability

and maintenance intervals for tool replacement and repair. To repair worn tools, metal is deposited by

fusion welding technique. The tribological behaviour of repair welded tool steel sliding against Al-Si coated

UHSS has not been studied in detail and there is a need to investigate if the modified tool surface will

affect friction and wear.

Hot stamping, similar to many manufacturing processes, is affected by the global mega trend of

digitalization and Industry 4.0. To monitor the process and optimize the control and operation are the

main aims. In view of this, tribological condition monitoring is a promising approach that can allow

measurement of physical properties such as vibrations, temperatures, and acoustic emission to be

coupled to the tribological response of the system. The aim is to monitor the hot stamping process and

enable early detection of changes in friction and wear which can be used for e.g. optimized maintenance

and minimized scrap.

The aim of this M.Sc. thesis was to improve the robustness of hot forming processes by studying

the tribological behaviour of repair welded tool steel sliding against Al-Si coated UHSS under conditions

relevant for hot stamping. Another aim was to obtain more predictable tool maintenance by the

implementation of acoustic emission measurement system on a hot-strip tribometer and correlating

condition monitoring signals to friction and wear phenomena.

The tribological tests were carried out using a hot-strip tribometer in conditions representative of

a hot stamping process of automotive components. Acoustic emission during sliding between hot work

tool steel and different automotive component material surfaces was measured at room temperature in

the same strip drawing tribometer and correlated to friction and wear of the surfaces to get more

predictable maintenance intervals.

Tool steel specimens were welded with the same material as the base material QRO90. Before

conducting the tribological test, the repair welded tool steel pin cross-section was polished, etched, and

observed under optical microscope and SEM to analyze the effect of Tungsten Inert Gas (TIG) welding

process on the microstructure. The analysis was completed with EDS to study the elements in the

microstructure. Microhardness was measured to obtain the microhardness profile from the repair welded

tool steel pin surface to the bulk in order to study the effect of different microstructures on the mechanical

properties. The weight and surface roughness of the pins were measured before the tribological test. After

the test was finished, the weight of the pins was measured to calculate the weight difference. The sliding

surface of the pins and the strips were photographed. The sliding surface of the pins was also observed

and analyzed using SEM and EDS after the test to study wear characteristic of the repair welded tool steel

at high temperatures.

Acoustic emission signal from the sliding was studied using Toolox44 pins with surface roughness

300-400 nm and with lay direction parallel and perpendicular to sliding direction. Toolox44 pins were

sliding against uncoated UHSS, as-delivered Al-Si coated UHSS, and heat-treated Al-Si coated UHSS strips.

Acoustic emission was measured during the sliding at the same time as COF measurement. Weight of the

Page 4: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

iii

pins was measured before and after the test and the wear damage on both surfaces was photographed.

COF, AE signals in the time and frequency domain, and wear damage were compared and analyzed.

It is found that repair welded tool steel has similar COF compared to the original hot work tool

steel with the largest weight gain from the test at 700 ⁰C due to compaction galling mechanism with

slower lump formation and the presence of wear particles, transfer layer, and formation of lumps. The

weight gain is smaller from the test at 750 ⁰C due to faster lump formation. The weight loss from the test

at 600 ⁰C is due to abrasive wear mechanism. SEM micrographs revealed that the repair welded tool steel

surface and transfer layers can be found beneath a transfer layer. Wear particles adhered on the repair

welded tool steel surface come from broken transfer layer or directly from Al-Si coated UHSS.

A change in wear mechanism is indicated by acoustic emission burst signals or gradual amplitude

change in the time domain. Frequency analysis of AE signals revealed a change in wear mechanism due

to the formation of transferred material in the form of a lump causes AE signals with peaks at higher

frequencies above 0.3 MHz to shorten.

Page 5: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

iv

Contents

1. Introduction ............................................................................................................................... 1

1.1. Hot stamping .................................................................................................................................... 2

1.2. Tribological behaviour of tool steel sliding against Al-Si coated UHSS ............................................ 3

1.3. Hardfacing ........................................................................................................................................ 6

1.4. Acoustic emission ............................................................................................................................. 8

1.5. Signal processing ............................................................................................................................11

1.6. Research gaps ................................................................................................................................11

2. Aim and objectives ................................................................................................................... 12

3. Experimental work ................................................................................................................... 13

3.1. Test materials .................................................................................................................................13

3.2. Experimental technique .................................................................................................................14

3.2.1. Tribological behaviour of repair welded tool steel against Al-Si coated UHSS ...................14

3.2.2. Acoustic emission during sliding between hot work tool steel and UHSS..........................17

3.3. Analysis technique .........................................................................................................................19

3.3.1. Tribological behaviour of repair welded tool steel against Al-Si coated UHSS ...................19

3.3.2. Acoustic emission during sliding between hot work tool steel and UHSS..........................20

4. Results and discussion ............................................................................................................. 21

4.1. Tribological behaviour of repair welded tool steel against Al-Si coated UHSS ..............................21

4.1.1. Pre-test analysis ..................................................................................................................21

4.1.2. Tribological test result ........................................................................................................24

4.2. Acoustic emission during sliding between hot work tool steel and UHSS .....................................34

5. Conclusions .............................................................................................................................. 46

6. References ............................................................................................................................... 47

Appendix A: Special pin holder technical drawing ...................................................................... 50

Appendix B: Repair welded tool steel pin microhardness data................................................... 51

Appendix C: Repair welded tool steel pin surface roughness and weight .................................. 52

Appendix D: Toolox44 pin surface roughness and weight ............................................................ 53

Appendix E: Matlab code for AE signal conversion into the frequency domain ......................... 54

Page 6: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

1

1. Introduction

Automotive industry aims to reduce weight and improve the safety in automotive applications. In

order to achieve the targets, automotive industry needs the components to be made of a material with a

high strength-to-weight ratio such as ultra-high strength steel (UHSS) which explains the reason of

increasing demand for this class of material. UHSS automotive components such as B-pillar, bumper, and

tunnel shown in Figure 1.1. are manufactured by hot stamping process [1][2]. Like other metal forming

processes, friction and wear occur between the tools and the workpiece. Friction affects the formability

of the workpiece. Very high friction causes residual stresses and defects due to retarded metal flow. If the

workpiece is successfully produced, the operation requires higher force and power to overcome friction.

The tools are worn out after several forming cycles which causes product dimensional inaccuracy [3].

Worn tools are replaced with new tools while they are repaired as part of maintenance activities.

Repairing worn tools by hardfacing process, which is a metal deposition by fusion welding techniques,

shall enhance the wear resistance of the tools [4], extend the time interval for tool replacement and

repair, and keep the production running for a longer time. Tool replacement schedule is also important

to prevent defective production due to worn tools, but not too early as there will be a high number of tool

replacements and costs. A condition monitoring technique can be applied on the tools to capture the

signal which indicates the time to replace the tools and predict the optimum schedule.

Figure 1.1. Hot stamped automotive components [1].

Page 7: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

2

1.1. Hot stamping

UHSS automotive components are manufactured by hot stamping process. The process has two

variants; direct and indirect hot stamping. In direct hot stamping, the workpiece is heated up in the

furnace above the austenitization temperature (950 ⁰C) and is then transferred to the tools. The

workpiece is pressed and formed between the tools. While the tools are closed, the workpiece is

quenched by flowing water through the tools or directly to the workpiece. Quenching results in workpiece

transformation from austenite into martensite that gives material properties to meet the demands from

automotive industry for an ultimate strength reaching 1500 MPa. UHSS such as manganese-boron steel

(22MnB5) is a common material to undergo hot stamping process because boron is able to suppress the

formation of ferrite and perlite and promote martensite formation. In indirect hot stamping, which is

usually used for very complex geometries, the workpiece is formed at room temperature before it is

heated up in the furnace. The next steps of indirect hot stamping are the same to direct variant. In the

end of the process, the workpiece is cut to achieve the final component dimension and cleaned [1][5].

The steps of direct and indirect hot stamping processes are shown in Figure 1.2.

Figure 1.2. Illustration of hot stamping process variants [6].

During heating in the furnace, decarburization and oxidation occur on workpiece surface if the

furnace atmosphere is uncontrolled. Decarburization alters hardenability and mechanical property of the

steel because of the reaction between carbon in the steel and oxygen in the air. Oxidation also occurs

during transfer from the furnace into hot stamping press. Oxide scales slow down the heat transfer

between the workpiece and the tools during quenching and cause weaker strength of the workpiece

material [7]. Oxide scales also need to be cleaned by shot blasting or shot peening prior to further

processes such as corrosion protection, painting, and welding. To prevent decarburization and oxidation,

the workpiece is hot-dipped to apply a protective layer such as Al-Si coating. In the furnace, Al-Si coating

melts and aluminium diffuses to the UHSS substrate while iron from UHSS substrate diffuses into the

coating and results in hard intermetallic layers. These layers provide good barrier protection against

corrosion, spot weldability and suitable surface for painting. A disadvantage of Al-Si coated UHSS is that

Page 8: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

3

the coating sticks to furnace roller during heating [5]. Al-Si coating is usually applied on the workpiece in

direct hot stamping. For indirect variant, zinc coating is used because of limited formability of Al-Si coating

at room temperature [1]. Some authors [8][9][2] studied the use of zinc coating in direct hot stamping

process because zinc coating offers cathodic protection instead of barrier protection against corrosion.

Not only by hot-dipping, the coating can also be applied by electroplating such as Zn-Ni coating and its

tribological behaviour was studied by Kondratiuk et al [10].

Since Al-Si coating is commonly applied on UHSS components in hot stamping, many authors

studied tribological behaviour of tool steel sliding against Al-Si coated UHSS. Severe wear damage in this

tribological system is indicated by severe material transfer or galling on the tools. Definition of galling

according to ASTM G40 (2010) [11] is:

“A form of surface damage arising between sliding solids, distinguished by macroscopic, usually localized,

roughening and creation of protrusions above the original surface; it often includes plastic flow, material

transfer or both.“

1.2. Tribological behaviour of tool steel sliding against Al-Si coated UHSS

Pelcastre et al. [7] has studied the mechanism of galling in hot stamping which consists of four

stages. He also divided galling into adhesive and compaction galling where the former has a lower rate.

The schematic of the galling mechanism is shown in Figure 1.3. In stage I.I, wear debris of intermetallic

layers and oxidized tool steel are accumulated in the surface valleys or defects when the surface

roughness is higher. Further accumulation causes compaction of wear debris forming conglomerates in

stage I.II. In stage III, these conglomerates become an obstacle for wear debris between the surfaces to

move out so wear debris is further compacted and in stage IV severe galling has occurred as a big lump.

In case of a smooth surface, adhesive galling occurs and begins in stage II.I where material transfer starts

on regions with high contact pressure such as at the asperity. Further material transfer leads to the

formation of transfer layer in stage II.II. When the transfer layer starts to act as an obstacle, stage III begins

and the mechanism ends with severe galling in stage IV in the same way as explained before.

Figure 1.3. Schematic of the galling mechanism between tool steel and Al-Si coated UHSS [7].

Page 9: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

4

In [7], the study focused on parameters affecting galling such as contact pressure, surface

roughness, tool hardness, and heating conditions on Al-Si coating. Increasing contact pressure causes

more material transfer. The effect of surface roughness becomes dominant at lower contact pressure

while the effect of hardness is dominant at higher contact pressure. At lower contact pressure, higher

surface roughness provides a place to accumulate more wear debris to start compaction galling

mechanism and causes more severe galling. At higher contact pressure, the surface with lower hardness

has the asperities deformed causing larger real contact area for severe adhesive galling while the surface

with higher hardness has less deformed asperities and less wear debris causing less severe galling. Other

authors [12] found that increasing tool hardness leads to improved resistance against adhesive wear

because of reduced elastic deformation of asperities and reduced real contact area although similar

coefficient of friction (COF) are obtained. Related to the surface roughness, Pelcastre et al. found that a

surface roughness lay direction parallel to the sliding direction reduces galling severity [7].

Heating of Al-Si coated UHSS influences galling severity through its microstructure. Al-Si coating

prior heating has a low hardness of 55 HV. After heating, Al-Si coating forms intermetallic layers with

different hardness for each layer. Formation of layers depends on heating temperature and holding time.

Holding the Al-Si coated UHSS at 700 ⁰C for 0 minutes still leaves Al-Si coating on top of the surface and

forms other layers below such as FeAl2/Fe2Al5, Fe2Al2Si, and unstable Fe2Al7-8Si. Holding it for a longer time

at the same temperature leaves Fe2Al7-8Si on the surface. Similar microstructure appears for holding time

0 minutes at 900 ⁰C. These heating conditions causes unstable COF. With higher temperature and longer

holding time, intermetallic layers consist of a diffusion layer, FeAl2/Fe2Al5 and Fe2Al2Si. Galling is reduced

after heating Al-Si coated steel at higher temperatures for longer holding time because of a hard and

stable phase FeAl2/Fe2Al5 at the surface reaching 800 – 850 HV and results in stable COF. High and unstable

COF is usually found after a lump of transferred material is formed on the tool steel surface. Uniform

transfer layer results in more stable and lower COF. In [13] it was also found that surface topography of

Al-Si coating does not affect tribological performances significantly when sliding against tool steel.

Another study [10] shows that increasing the austenitizing temperature for Al-Si coated UHSS to 920 ⁰C

from 880 ⁰C reduces adhesive wear where the authors held both temperatures for 6 minutes.

Further studies of tool steel sliding against coated steel in hot stamping pushed the development

of strip-drawing type tribometer that more accurately simulates the tribological conditions in the hot

stamping process where the tools slide repeatedly against fresh Al-Si coated UHSS surface. According to

[14], tribological test conditions with a strip drawing tribometer which represent hot stamping has contact

pressure 5, 10, and 15 MPa at temperatures of 600, 700, and 750 ⁰C and sliding at a speed of 100 mm/s.

The reason for a single test speed is because the speed only has a small influence on galling [15]. From

strip-drawing tribometer tests, it was found that galling occurs in the middle of the leading edge of the

contact area where contact pressure is lower [14]. They also found that test temperature during sliding

plays a significant role in the severity of galling. Severe galling occurs on the tool at higher temperatures

while mild galling and abrasive wear occur at lower temperatures. They correlated the result with contact

pressure and sintering that builds up the material on the tools. Lower contact pressure would not be

enough to sinter a transfer layer while higher contact pressure would build up the material at a higher

growth rate. Abrasive wear correlates to higher COF. COF is lower when the applied temperature and

contact pressure manage to form a transfer layer. A stable transfer layer is indicated by stable COF. When

galling occurs at the leading edge, COF rises up.

Page 10: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

5

The opposite trend for galling on the tool was found in [16] where adhesive wear occurs at lower

temperatures while abrasive wear mainly occurs at higher temperatures. Abrasive wear is caused by

embedded wear debris in the sheet or trapped debris between the surfaces. Although abrasive wear is

not dominant at lower temperatures, compaction galling occurs and is made of smaller size wear debris

compared to wear debris at higher temperatures. This result appears after sliding for longer distances

which shows the importance of sliding distance. During the first slides, abrasive wear is dominant at lower

temperatures scratching the tool surface and then is covered by transferred material after farther sliding

distance. The scatter of the coefficient of friction is related to tool wear evolution. The first slides have

the highest COF in every temperature compared to the next slides and these values are temperature

dependent. After farther sliding distances, temperature dependency is reduced because the interacting

surfaces change from tool steel-coating material to adhered coating material-coating material. Other

authors [12] also evaluated COF after sliding on several strips and explained that COF decreases because

adhered Al-Si layer on the tool gives a lubricating effect until the layer reaches a specific thickness. This

COF trend is also found on AlCrN coated tools but not for DLC coated tools which has stable COF and very

thin adhered material only at the leading edge after sliding against nine strips [17].

Thermal conductivity and heat transfer are two other parameters influencing galling. In [18] it was

found that at 600 ⁰C a tool steel with lower thermal conductivity has adhesive wear as a dominant

mechanism because this thermal property causes higher interface temperature and softens Al-Si coating.

Abrasive wear is dominant for higher thermal conductivity tool steel because the tool steel softens and is

scratched after it extracts more heat from the Al-Si coated UHSS. Effect of heat transfer also comes from

the material coating. In [9] it was found that a Zn coating causes more abrasive wear on the tools

compared to Al-Si coating because Zn coating has higher heat transfer coefficient which transfers the heat

to the tool steel, softens, and scratches it. Another study [12] shows small differences in material transfer

between smaller and lower thermal conductivity steels. Higher thermal conductivity tool steel can be

obtained when the alloying elements are precipitated as carbides instead forming a solid solution state in

the martensite phase. In [18], tool steel with larger Mo precipitates has similar COF compared to tool steel

with dispersed Mo.

The effect of friction on Al-Si coated UHSS formability is shown in [19] where lubricants were

developed and evaluated to reduce the coefficient of friction between the tools and Al-Si coated UHSS.

The lubricants were evaluated using hot deep drawing test machine where a flange is formed from a blank.

A flange is successfully formed under lubricated condition with lower COF but fractures under dry

condition with higher COF. Lower COF also results in longer draw-in length indicating better Al-Si coated

UHSS formability. Azushima et al. [20] found that tool surface roughness affects COF in lubricated

condition by serving as a lubricant trap and contacting asperities. Lubrication results in thinner adhered

layer on the tools. The surface roughness does not affect COF in dry condition. This finding perhaps is

caused by applying higher contact load leading to higher contact pressure which reduces the effect of

surface roughness. Applying a coating on the tools is another way to reduce COF.

Page 11: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

6

1.3. Hardfacing

Worn hot work tool steels are commonly repaired by a hardfacing process which is a deposition

of wear resistant metal on the tool surface by fusion welding techniques [4]. Improved wear resistance

can also be achieved by other materials and methods such as deposition of ceramics by PVD and CVD

processes but they can fail in mechanical presses due to high impact loading. Hardfacing is able to provide

a stronger bond between the weld metal and the substrate and can, therefore, be a better choice for

metal forming in mechanical press [21]. Welding technique for hardfacing process is selected based on

several factors such as weld metal, base metal, size and shape of the component, etc. Some welding

techniques for hardfacing are metal arc welding, Tungsten Inert Gas (TIG) welding, and plasma transferred

arc welding. Weld metal for hardfacing can be nickel, cobalt, or iron-based alloys with different

compositions contributing to the wear resistance. Commonly used weld metal is Fe-Cr-C steel because of

good abrasive wear resistance. Abrasive wear resistance increases with higher carbide volume fraction

until it will not improve further after reaching a certain limit. Carbides contribute to abrasive wear

resistance by disrupting abrasive particles to scratch the matrix. This occurs if carbides have higher

hardness than abrasive particles. Not only the volume fraction, distribution and the size of carbides also

contributes to the wear resistance [22].

Lee et al. [23] studied the wear resistance of iron-based weld metal and found that it is influenced

by the sliding distance and temperature, which also determine wear mechanism. At room temperature,

the wear resistance of Fe-Cr-C-Si steel comes from an increase in weld surface hardness because of strain-

induced transformation (SIT) from austenitic to martensitic phase. An increase in hardness reduces wear

loss due to adhesion. At higher temperature, inhibition of SIT reduces the martensitic transformation and

the hardness. This inhibition causes an increase in wear rate. Wear rate can be reduced after further

sliding distance at higher temperatures due to the formation of oxide layers which protects metal-to-

metal contact causing adhesive wear.

Lee et al. [24] studied the effect of different compositions to the wear resistance. Increasing

chromium and carbon content higher than the eutectic point forms a hypereutectic structure with a lot

of carbides and coarse microstructure. In this structure, insufficient carbon content forms a ferrite matrix.

Less effective work-hardening of the ferrite phase than the austenite phase causes severe adhesive wear

on the matrix. Too much alloying elements can also cause severe adhesive wear. Although the addition of

boron can harden the matrix by grain size refinement, too much boron causes the matrix to have no

austenitic phase required for martensitic transformation and formation of FeB carbides which is brittle

and is easily removed from the surface [25]. Martensite in lathe form is not only available in α’ phase with

BCT crystal structure but also in ε phase with HCP structure [26]. High manganese steel is able to transform

ε martensite at higher temperatures where α’ martensite transformation is inhibited. This results in higher

hardness and adhesive wear reduction at higher temperatures.

Oxide layers and ε martensite phase are formed in cobalt-based alloys such as Stellite 21 during

sliding depending on the temperature. At room temperature, oxides are not yet formed but full SIT into ε

martensite occurs. High COF is found and wear weight-loss is maximum at 200 ⁰C because oxide particles

have been formed but unable to form glaze layer to protect against wear. These oxide particles act as the

third body in abrasive wear mechanism which causes the highest wear. At higher temperatures, oxide

particles form a glaze layer inside the scratches and results in lower COF and better protection against

wear. At higher temperatures, wear protection also comes from incomplete transformation of austenite

Page 12: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

7

to ε martensite phase on the worn surface and work hardened residual austenite sublayer which supports

the formation of glaze layer and acts as good absorber of impact loading [27].

Hot forging tool surface temperature can reach 600 - 800 ⁰C, which is higher than the temperature

in wear test 550 ⁰C. Plastic deformation on Stellite 21 surface after hot forging is less severe because of

recrystallization that formed uniaxial grain. New carbides precipitate near Stellite 21 surface after hot

forging in the interdendritic area and near initial carbide. Compared to nickel-based weld metal such as

Inconel 625, its surface microstructure does not have carbide because its precipitation is very slow in

temperature higher than 650 ⁰C. Inconel 625 has a lower hardness compared to Stellite 21 because it has

increasing hardness only from work hardening. In hot forging process, glaze layer on Inconel 625 was

found unstable and this resulted in lower wear resistance compared to Stellite 21 [21].

Kashani et al. [28] studied the effect of different composition between weld and base material on

hardness and wear behaviour. Fusion line is the interface between substrate and weld overlay. From the

weld surface to the fusion line, hardness difference appears if weld and base material has a big difference

in carbon composition. Weld material with higher carbon content than base material will have decreasing

hardness near the fusion line because of carbon diffusion to base material with lower carbon

concentration causing less carbide formation. The opposite result occurs for weld material with lower

carbon concentration than base material while similar carbon content between weld and base materials

do not cause much difference along the weld overlay. The hardness of weld overlay near the fusion line

also increases after a wear test because of heat transfer from the surface during the test resulting in

carbide precipitation and higher hardness. Dilution level also affects weld material hardness in a similar

manner. For higher carbon weld material, high dilution causes worse wear resistance compared to low

dilution because of carbon diffusion to base material with lower carbon content causing less carbide

precipitation and lower hardness. The opposite result occurs for low dilution weld material.

Another parameter affecting the wear resistance is welding heat input. For weld steel with high

chromium content such as AISI H13, welding with high heat input causes decreasing chromium content in

the matrix, larger carbide formation, and worse oxidative wear resistance because the steel is unable to

form chromium oxide. Similar oxidative wear resistance occurs in high chromium weld steel after post-

weld heat treatment although it has higher hardness due to secondary hardening. Retained austenite and

martensite are available in dendritic structure while the interdendritic area contains more alloying

elements. The dendritic microstructure becomes more refined with lower heat input [29]. Compared to

nickel and cobalt-based weld metal, as-weld Inconel 625 microstructure consisted of a solid solution single

phase while Stellite 6 consisted of a dendritic supersaturated cobalt solution in austenite phase and

interdendritic eutectic of chromium carbides and supersaturated cobalt solution. The carbides in Stellite

are made of chromium and a small amount of molybdenum or tungsten and cobalt. Lower heat input and

faster cooling rate result in smaller dendritic, shorter dendritic arm spacing, and lower wear weight loss

compared to higher heat input and slower cooling rate [30].

Page 13: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

8

1.4. Acoustic emission

Some examples of condition monitoring techniques for tools in metalworking operation are

vibration, audio signal, and acoustic emission (AE). Each technique has its disadvantages, for examples

vibration signals are easily affected by environmental noise and vibrations of nearby machines. It is also

affected by sensor position, tool geometry, workpiece material, and punch speed. A change in process

parameters can change AE energy [31]. Audio signals are also affected other manufacturing process

working at the same time and require other signal extraction algorithm such as semi-blind technique to

extract audio associated with the monitored process [32]. Despite its disadvantage, many researchers use

AE to study tool wear in detail as a mean to monitor tool condition and also friction against workpiece

surface. This is due to high sensitivity, large frequency range, and without further complex signal

processing and algorithm [33].

According to [34], acoustic emission (AE) is defined as:

“...transient elastic waves generated from a rapid release of strain energy caused by a deformation or

damage within or on the surface of a material.”

In a smaller scale, the dynamic local rearrangement of material structure results in these elastic waves.

Elastic waves are stress waves in an elastic body. Depending on the material in which stress waves

propagate, they can also be plastic, nonlinear elastic, and even visco-elastic waves. Stress waves are

formed due to the inertia of particles of a media. When the surface is externally loaded, surface particles

will move to a new equilibrium position while neighbouring particles stay in the initial position causing

stresses between particles. The movement of neighbouring particles to new equilibrium follows later as

the cause of inertia. The interactions continue farther from the surface which become stress waves [35].

AE signals can be divided into continuous and discrete (or burst). Continuous AE consists of a

number of elementary events that emits small energy as elastic waves while discrete AE consists of a huge

number of elementary events in a short time emitting bigger energy than continuous AE. These two

acoustic emissions appear at the same time during monitoring the progress of material damage. Discrete

AE is used to monitor the appearance of cracks while continuous AE is used to monitor plastic

deformation, corrosion, and other physical phenomena.

The source of AE in metals can be related to different mechanisms. The mechanism leading to

plastic deformation including dislocation movement and grain boundary slip is one of the AE sources.

Other mechanisms as AE sources are phase transformation such as martensitic transformation and

mechanism related to damage such as crack initiation and growth [36]. During sliding, the collision

between surface asperities leads to deformation and fracture of asperities which generate AE [37][38]. In

adhesive wear, plastic deformation and fracture are caused by the formation of asperity junction and

asperity breakage leading to material transfer onto asperities of one or both of sliding surfaces. In abrasive

wear, plastic deformation and fracture of the surfaces are caused by ploughing and cutting action [39].

Acoustic emissions are captured, recorded and presented as signals. These AE signals have

different parameters useful for the analysis of material damage. One of the parameters is the frequency

spectrum, which tells the nature of AE source itself. Other parameters such as AE amplitude can be useful

to determine the energy of an AE source while the count rate tells the rate of defect growth [36]. There

are many AE parameters, more of them are shown in Figure 1.4. Analyzing AE signals based on relevant

Page 14: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

9

and important parameters from AE raw signal in the time domain is called the parameter-based method

which uses AE counts, cumulative AE hits, cumulative AE energy, and AE energy distribution. Another

method is called the frequency-based AE method which is based on power spectral analysis. This method

converts the signal into the frequency domain using the Fast Fourier Transform (FFT). The signal can also

be converted by other processes such as Wavelet Transform (WT) for time-frequency analysis. In this

method, dominant frequencies and the magnitudes are used to characterize AE signal [40]. Frequency-

based AE method will be used in this work to analyze AE signals extracted from the tribological test.

Figure 1.4. Parameters of an AE signal [41].

Tian et al. [38] showed that sliding on one surface or between two surfaces with low hardness

causes negative relationship between AE strength (in RMS) and COF. AE strength from sliding on low

hardness surfaces is not much affected by the load. The opposite trend applies to sliding on hard surfaces.

For both surface hardness, higher sliding speed results in higher AE amplitude. In the case of a positive

relationship between AE signal strength and COF, Benabdallah [42] found a linear relationship between

AE energy and friction work for plastics. In the same case, it was shown in [43] that it is possible to predict

COF from RMS of AE signal. Asamene and Sundaresan [37] studied the effect of surface roughness, sliding

speed, and contact pressure between two hard surfaces on AE signal frequency components. Smoother

surfaces tend to have signals with high frequency components while rougher surfaces tend to have low

frequency components. As the sliding speed increases, the magnitude at high frequencies increases and

the opposite applies at low frequencies. Change in contact pressure does not affect frequency

components but only affect signal strength which goes stronger with increasing contact pressure [38]. AE

signal amplitude increases with wear severity. Not only that, according to [44], AE signal amplitude can

Page 15: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

10

gradually increase as continuous AE when there is a change from elastic to plastic deformation and

accompanied by a decrease in median frequency.

According to [39], abrasive wear is characterized by distributed peaks in range 0.25 – 1 MHz while

adhesive wear is characterized by frequency peak around 1.1 MHz due to the generation of wear elements

and transfer particles. The peak at 1.1 MHz, not only appears from reciprocating tribometer but also

appears in adhesive wear on journal bearing. In this case, peaks around 0.5 MHz corresponds to plastic

deformation of the surfaces and formation of transfer particle while peaks around 0.3 MHz, which

occurred before seizure, corresponded to rolling and collision of transfer particles [45]. Distributed peaks

of abrasive wear are also found in the frequency range 0.3-0.9 MHz [40]. In the case of three body abrasion

on ceramic and metal mixed phase coating, abrasive wear mechanism accelerates crack propagation to

the surface causing the formation of interconnected micropits. The peaks are distributed in a smaller

frequency range between 0.2 – 0.25 MHz [46]. In [47] it was found that the sources of AE in that study

come from adhesion between asperities generating Schallamach ridges and ploughing causing scratches

with corresponding peaks around 1 MHz and 0.2 MHz. The location of the peaks due to adhesion is close

to that shown in [39] but the ploughing action did not result in widely distributed peaks. Kolubaev et al.

[48] studied that increasing COF due to adhesion is accompanied by the appearance of high frequency

peaks. Although Kolubaev et al. [48] mentioned the peaks to be at high frequencies, the range is only up

to 0.4 MHz which is lower than found in [39] for adhesive wear. Sindi et al. [49] found that adhesive wear

has distributed peaks in specific ranges which the authors divided based on three stages of wear. In the

early stage, the peaks are in the frequency range of 65 - 126 kHz due to material detachment of a surface.

The next stage has peaks in frequency ranges 190 - 375 kHz due to plastic deformation of a surface and

material transfer. In the last stage, the peaks are in the frequency range of 250 - 310 kHz due to fracture

of friction junctions. This distributed peaks also occur in [50] to the point where scuffing occurs.

Wear can be divided into three stages, these are running-in, steady state, and catastrophic wear

stage [36]. Analysis of AE signals can be based on these three stages such as in [40], but not always. AE

signal analysis can also be based on the progress of crack in the material consisting of four stages as shown

in [46]. The stage to analyze AE signals can also be self-determined to see the transition stage such as

determination of pre-scuffing in [50]. In this work, AE signals to detect wear progress before the

occurrence of galling shall be analyzed based on stages of galling mechanism in [7].

Page 16: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

11

1.5. Signal processing

Acoustic emission signals are recorded as data points. These data points are interpolated by a

finite linear combination of sines and cosines which is called as a Discrete Fourier Transform (DFT). This

interpolation uses matrixes to convert discreet AE signal points into its frequency components. Nyquist

frequency is the highest available frequency component of a signal captured at a specific sampling rate.

Nyquist frequency is equal to half of the sampling rate. Beyond Nyquist frequency, the values of frequency

components are negatives of those below Nyquist frequency. Symmetries of matrixes used in DFT are

used to avoid redundancies by an algorithm called Fast Fourier Transform (FFT) [51].

DFT considers the signal to be analyzed as a periodic sequence. A period of a sequence has the

missing end point, so the beginning of the next period can be considered as the end point of the previous

sequence. Raw signals to be processed by DFT are often asymmetric because of abrupt change at the end

of a period. This abrupt change causes the appearance of new peaks in the frequency domain known as

spectral leakages. To reduce spectral leakages, a weighting function known as a window is applied to data

before converting it into the frequency domain. One of the well-known windows is the Hann window

with its function weighting across a period of a signal [52]. The function in the Hann window is shown in

Figure 1.5.

Figure 1.5. The function of the Hann window which is applied on the signal to reduce spectral leakage [52].

1.6. Research gaps

Galling mechanism in hot stamping has been studied in detail but not for repair welded tool steel

sliding against Al-Si coated UHSS. Although there are many studies correlating condition monitoring

signals to wear mechanisms, the studies to correlate condition monitoring signals to galling mechanisms

are scarce. Tribological behaviour of repair welded tool steel sliding against Al-Si coated UHSS and

tribological condition monitoring in hot stamping process are important to ensure that worn tools can be

replaced at the optimum time and be repaired to continue the hot stamping process for an extended time.

Page 17: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

12

2. Aim and objectives

The aim of this work is to enable improved robustness of hot forming processes and more

predictable tool maintenance. The objectives of this work are:

• To understand the friction and wear mechanisms of repair welded tool steel sliding against Al-Si

coated UHSS at high temperatures.

• To implement acoustic emission measurements in a hot strip drawing tribometer

• To correlate the condition monitoring signals to friction and wear behaviour between hot work

tool steel and different automotive component material surfaces.

Page 18: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

13

3. Experimental work

Test materials and techniques for experiment and analysis are explained in this chapter.

Experimental and analysis techniques are divided into two sub-chapters for tribological behaviour and AE

studies. Tribological behaviour of repair welded tool steel sliding against Al-Si coated UHSS was studied

by tribological test in conditions representing hot stamping process at three temperatures. AE of sliding

surfaces was studied using different surface materials and different directions of the surface topography

lay.

3.1. Test materials

For the study of tribological behaviour of repair welded tool steel, pin specimens were made of

hot work tool steel QRO90 as the base material and the same material was used as the weld material.

These pins were slid against 1 m long and 15 mm wide UHSS strip made of 22MnB5 boron steel coated

with Al-Si with a strip thickness of 1.5 mm. The compositions of QRO90 [53] and 22MnB5 [5] are presented

in Table 3.1. The pins were TIG welded by Volvo using a welding process proprietary to Volvo. The excess

weld material on each pin was removed by wire-EDM to obtain 1.5 mm radii at the leading and trailing

edges and a flat surface with width 10 mm and length 7 mm where the sliding occurs. Repair welded pins

and Al-Si coated strips are shown in Figure 3.1 and Figure 3.2 with their respective dimensions.

Table 3.1. Composition of test pins and strips [53][5].

Figure 3.1. The dimension of test pins (left) [2] and the repair welded pins (right).

Steel Composition, wt%

C Si Mn P S Cr B Mo V Ni Fe

QRO90 0.38 0.3 0.75 - - 2.6 - 2.25 0.9 - Rem

Toolox44 0.32 0.6-

1.1

0.8 max

0.01

max

0.003

1.35 - 0.8 0.14 max 1 Rem

22MnB5 0.20-

0.25

0.20-

0.35

1.0-

1.3

max

0.03

max

0.01

0.14-

0.26

0.005 - - - Rem

Sliding surface

Page 19: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

14

Figure 3.2. The dimension of test strips (top) [2] and Al-Si coated UHSS (bottom).

For AE study, pre-hardened hot work tool steel (Toolox44) pins with the same dimension as shown

in Figure 3.1 were used. The composition of Toolox44 [5] is also shown in Table 3.1. The counter surfaces

of Toolox44 pins were Al-Si coated and uncoated UHSS strips made of 22MnB5 steel with the same

dimension as shown in Figure 3.2.

3.2. Experimental technique

3.2.1. Tribological behaviour of repair welded tool steel against Al-Si coated UHSS

The side of three repair welded pins was polished and etched with Nital 3% for 1 minute and 30

seconds to see the microstructure using a Nikon MA200 optical microscope at 25X and 500X

magnification. The microhardness of etched surfaces was measured to obtain the microhardness profile

based on 30 measurement points starting with the first point at 0.025 mm from the sliding surface. The

following points were separated by 0.075 mm until reaching 0.325 mm from the sliding surface. It was

continued with the next points separated by 0.3 mm until reaching 6.025 mm from the sliding surface and

was continued with the next points separated by 1 mm until reaching the final value at 12.025 mm from

the sliding surface. The measurement of 30 points was repeated three times for an etched surface of three

repair welded pins. The microhardness was measured by 100gf load. The side of the other two repair

welded pins was polished and etched with Nital 3% for 2 minutes to see the microstructure under JEOL

JSM-IT300LV SEM microscope and for elemental analysis using EDS. The purpose of microstructure

observation and microhardness measurement is to investigate the effect of welding process on

mechanical properties of repair welded pins as pre-test analysis.

Page 20: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

15

After that, the sliding surface of repair welded pins were ground by grinding stone in the sliding

direction and a surface roughness Ra in the range 400-500 nm was obtained. The surface roughness was

measured using a Wyko NT1100 3D optical profilometer with the parameters in Table 3.2 in the middle of

the trailing edge as position 1, in the middle of the leading edge as position 2, and in-between position 1

and 2 as position 3 after the sliding surface of each repair welded pins was wiped with ethanol-soaked

tissue. Before the repair welded pins were tested, they were cleaned in heptane in an ultrasonic cleaner

for 5 minutes at room temperature. After that, the repair welded pins were rinsed by acetone, dried, and

weighed three times and their average weight was calculated. Al-Si coated strips were also wiped with

heptane-soaked tissue and followed by acetone-soaked tissue before the test.

Table 3.2. Parameters of Wyko NT1100 profilometers.

Parameters Option / Value

Measurement type VSI

Resolution Full

Objective 2.5x

Field of view 1x

Scan speed 1x

Scan option Single scan

Averaging number 2

Backscan 15 μm

Length 30 μm

Modulation Threshold 2%

An unworn and cleaned pair of repair welded pins were assembled on hot-strip tribometer and

the contact between repair welded pins and the strip was aligned using shims under the pin holders and

pressure sensitive film which was loaded with 1200 N to check the contact patch. After contact alignment,

a clean Al-Si coated UHSS strip was assembled and was given a pre-tension. The electric current was

passed through the strip to heat it in a sequence by controlling the voltage. The voltage 1.35 V was held

for 5 minutes which brought the strip temperature to 600 ⁰C. The heating continued by holding the voltage

2.2 V for 2 minutes and 30 seconds that brought the strip temperature to 700 ⁰C. After reaching 700 ⁰C,

the voltage was brought up to 3.7 V and was held for 40 seconds which increased the strip temperature

up to 920 ⁰C. After that, the strip temperature was brought down to the test temperature by selecting the

corresponding voltage as presented in Table 3.3.

Table 3.3. The heating sequence with the corresponding voltage.

Step Voltage (V) Temperature (⁰C) Duration / Note

1 1.35 600 5 minutes

2 2.20 700 2 minutes 30 seconds

3 3.70 920 40 seconds

4

1.50 600 Test temperature held for approximately 45 seconds for sliding

preparation and sliding time.

2.15 700

2.40 750

Page 21: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

16

A contact pressure of 10 MPa was applied and the repair welded tool steel pins slid at a speed of

100 mm/s for 200 mm sliding distance for COF measurement. Tribological test conditions are listed in

Table 3.4 and the illustration of the sliding between a pair of pins and a strip is shown in Figure 3.3 where

side A and side B of the strip and the pins are described. After COF was measured, the repair welded pins

and Al-Si coated strip were separated and cooled by the air blast. If severe galling occurred, which was

indicated by severe scratches on both sides of the strip, then the test was repeated with new repair

welded pins and a new strip. If galling did not occur, the second clean strip was mounted and the same

repair welded pins were used again. This continued up to the third strip if galling still did not occur on the

second strip and then proceeded with new repair welded pins and a new strip for test repetition or new

test at different temperatures.

Table 3.4. Test conditions in tribological behaviour study.

Parameters Value

Contact pressure 10 MPa

Load 700 N

Speed 100 mm/s

Sliding distance 200 mm

Sampling rate 200 Hz

Temperature (strip)

600 ⁰C

700 ⁰C

750 ⁰C

Figure 3.3. Main features of hot-strip tribometer [14] (left) and the illustration of the pins sliding on the strip [2]

(right). Side A is the side with a moveable test pin for applying the normal load and side B is the side with a

stationary test pin [2].

Side A Side B

Page 22: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

17

Repair welded pins were weighted three times after the test and finally, the worn surfaces of

repair welded pins were examined under SEM and analyzed by EDS. For post-test analysis, a total of nine

SEM images were taken on each of the worn repair welded pin surface. Three images at the leading edge,

three images at the trailing edge, and three images between the edges. SEM images at special locations

were taken if there were interesting details. SEM images were taken using the parameters listed in Table

3.5. EDS analyses were taken in the middle of the leading edge, in the middle between the edges, and

special locations.

Table 3.5. Parameters for SEM images

Parameters Value / Option

Magnification 100X and 800X

Working distance 9.4-9.5 mm

Current voltage 15 kV

Probe current High, 40% and 70%*

Vacuum High

* 70% power for high probe current was used for some images due to SEM usage after maintenance.

3.2.2. Acoustic emission during sliding between hot work tool steel and UHSS

Toolox44 pins were ground by abrasive paper P240 to obtain a surface roughness Ra in the range

300-400 nm based on the measurement on the same three locations as in repair welded pins after they

were wiped with ethanol-soaked tissue. The surface roughness of uncoated UHSS strips is 1780 ± 120 nm

[5]. Eight Toolox44 pins were ground in the same direction as repair welded pins which were parallel to the

sliding direction. Four other Toolox44 pins were ground in the perpendicular direction to the sliding

direction. Toolox44 pins, Al-Si coated, and uncoated UHSS strips were cleaned following the same

procedure as in the tribological study of repair welded tool steel. Toolox44 pins were also weighted three

times before the test and the average weight was calculated. An unworn and cleaned pair of Toolox44 pins

was mounted on the hot-strip tribometer with a special pin holder for the pin on side B. Technical drawing

of this special pin holder is presented in Appendix A. The contact to the strip was aligned following the

same procedure as previously described. After the alignment, the AE sensor Physical Acoustics WSα was

dipped into the molten beeswax which was heated to 75 ⁰C [54]. After the AE sensor surface was covered

by beeswax with uniform thickness, it was attached immediately to the side of the pin mounted in the

special pin holder and was supported by a rubber band. The process of AE sensor attachment is shown in

Figure 3.4.

Figure 3.4. The process of attaching AE sensor, a) a clean AE sensor was ready, b) dip AE sensor surface on the

molten beeswax, c) lift AE sensor, d) attach AE sensor on the pin on side B.

a) b) c) d)

Page 23: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

18

The AE sensor was connected to a pre-amplifier Physical Acoustics 2/4/6C with a built-in bandpass

filter of 100 – 1200 kHz which amplified the AE signal 40 dB. For the sliding between Toolox44 pins and a

heat-treated Al-Si coated UHSS strip, it was set to amplify the signal by only 20 dB due to stronger AE

signal from this sliding compared to the others shown in Table 3.6. Choosing higher amplification causes

AE equipment to stop AE signal acquisition due to Channel Protection Fault. The pre-amplifier was

connected to a computer installed with 14-bit PCI-digitizer (GaGe OSC-432-007) and GageScope software

where the AE signals were saved. The assembly of AE equipment is shown in Figure 3.5.

Table 3.6. List of sliding surfaces in AE study.

No. Lay Direction of Toolox 44 Pin

Strip Material Number of tests

Amplification (dB)

1 Parallel Uncoated UHSS 2 40

2 Perpendicular Uncoated UHSS 2 40

3 Parallel As-delivered Al-Si coated UHSS 1 40

4 Parallel Heat-treated Al-Si coated UHSS* 1 20

* The strip went through heat sequence in Table 3.3 with 700 ⁰C as step 4

Figure 3.5. The assembly of AE equipment.

Page 24: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

19

A clean strip was mounted after AE equipment assembly and was given a pre-tension. The AE

signal was captured with a sampling rate of 5 MHz to obtain Nyquist frequency 2.5 MHz and to ensure it

was above maximum frequency which can be captured by the AE sensor (1 MHz). AE signal acquisition

was started approximately 1 second before the sliding of the Toolox44 began. The reason for conducting

the AE study at room temperature was due to the melting temperature of the beeswax 75 ⁰C. If the wax

melts, AE sensor risks to be damaged by the current from the strip and to be exposed to a higher

temperature than AE sensor maximum working temperature (180 ⁰C). AE acquisition duration was 6

seconds to prevent incomplete acquisition because the acquisition stops before the sliding ends.

Parameters of AE signal acquisition are listed in Table 3.7. After successful COF measurement and AE signal

recording, the AE sensor was removed and cleaned from the beeswax with ethanol-soaked tissue. Toolox44

pins were weighted after the test three times and the average weight after the test was calculated. The

test was repeated with different strips and different pin surface topography lay direction according to

Table 3.6. The photo of worn repair welded and Toolox44 pin surfaces and the strips were taken by a

camera.

Table 3.7. AE signal acquisition parameters.

Parameters Value

Sampling rate 5 MHz

Data points (post trigger) 30 000 000

Duration 6 seconds

Amplification 20 & 40 dB

3.3. Analysis technique

3.3.1. Tribological behaviour of repair welded tool steel against Al-Si coated UHSS

The microstructure of repair welded pins under the optical microscope with 25X magnification

reveals four different regions with different microstructures on each repair welded pin. The photos of

these microstructures were taken and stitched to present them as a whole repair welded pin surface. The

detail of each microstructure was observed at 500X magnification. The microhardness profile representing

repair welded pins was obtained by averaging nine microhardness profile from three rows microhardness

measurement on each repair welded pin with a total of three repair welded pins being measured. The

result of microhardness measurement and average microhardness calculation are presented in Appendix

B. The microstructure observed by SEM at 800X magnification was to see different phases in detail and to

analyze the element of different phases by observing different peaks in EDS spectrum. SEM micrographs

were also taken for each microstructure. Correlation between repair welded pin microstructures and

microhardness profile was made to finish the pre-test analysis.

The graph of COF over time was obtained after each test. The graph was converted to represent

COF over sliding distance by multiplying the time with the sliding speed 100 mm/s. The early part of the

curve where COF is zero was cut by looking at COF and traction force values. When COF was rising, the

corresponding traction force with a higher value than the traction force in previous data points was

Page 25: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

20

chosen as the starting point when the sliding started. The chosen COF and traction force as the starting

point must have positive values. When the sliding distance reached 200 mm from the starting point, the

remaining part of the curve was removed. COF data in the stable region were averaged to obtain COF

from a test on a strip. COF from the next strips were also calculated and all COF from all strips using the

same pins were averaged giving the average COF of a test. Average COF was again averaged with its

repetition to obtain average COF at a specific temperature.

The weight difference was calculated by subtracting the weight average of three pins before the

test by the weight average after the test. After weight difference was calculated for a test, it was averaged

with weight difference from the repetition to represent the weight difference at a specific temperature.

The measured weight, weight difference, and surface roughness data of repair welded tool steel pins are

provided in Appendix C. SEM micrographs at 100X and 800X magnification for post-test analysis were

obtained to correlate wear particles, transferred layers, and lumps to galling mechanisms. Repair welded

pin sliding surfaces were further analyzed by EDS to confirm different surfaces by detecting different main

elements, aluminium or iron.

3.3.2. Acoustic emission during sliding between hot work tool steel and UHSS

The weight difference of Toolox44 pins was calculated in the same way as described above, except

for conditions which had no repetition, no further averaging with another weight difference from the

repetition was made. The result of Toolox44 pin weight difference calculation is presented in Appendix D

with the measured weight and surface roughness. COF curves were also cut by the same method as before

but the graph is also presented with time along with the sliding distance to compare it with raw AE signals

in the time domain. Raw AE signals before the sliding started were cut. The time at which the sliding

started was determined by the time when the amplitude started to rise because of the sliding and it was

observed by zooming-in raw AE signals. AE signals after two seconds of sliding were cut. AE signals in the

time domain were compared to corresponding COF curves and worn strip surfaces and were divided based

on the amplitudes or appearances of burst signals. Each region was analyzed further by conversion of AE

signals in the time domain into the frequency domain. AE signals in the frequency domain from different

regions and different sliding combinations were compared to characterize the wear behaviour of the

tribological system.

Matlab was used to convert raw AE signals into the frequency domain using the code in Appendix

E. The duration of converted raw AE signals was only 0.001 seconds using 5000 data points. These selected

5000 points had the amplitude multiplied by Hann window function for the same number of data points.

FFT algorithm was applied on AE signals of 5000 data points. The result of the FFT algorithm was made

absolute and divided by the length of converted AE signals which was 5000 points. Only the first half of

the result is presented because the other half is the mirroring from the FFT. FFT result was plotted in the

frequency domain up to 1 MHz, the maximum frequency component which can be captured by AE sensor,

Physical Acoustics WSα. The frequency domain was actually up to Nyquist frequency 2.5 MHz, half of the

sampling rate to remove the mirroring.

Page 26: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

21

0

100

200

300

400

500

600

700

800

0 2 4 6 8 10 12 14

Mic

roh

ard

nes

s (H

V)

Depth from surface (mm)

Average repair welded tool steel pin microhardness

4. Results and discussion

Results and discussion are divided into two sub-chapters for tribological behaviour and AE studies.

Discussion of tribological behaviour study starts with the microstructures and microhardness profile of

repair welded tool steel pins to investigate the effect of TIG welding on the pins. Then, the COF evolution

over sliding distance and corresponding worn surfaces at different test temperatures are discussed.

Weight differences of the pins are discussed in the end. Discussion of AE study is about the comparison

between COF evolution with sliding distance and time, raw AE signals, and worn strip surfaces. This

comparison is to detect and understand galling progress by AE signals for tool maintenance purpose.

4.1. Tribological behaviour of repair welded tool steel against Al-Si coated UHSS

4.1.1. Pre-test analysis

The TIG welding process results in four different regions with different microstructures and

microhardness profile as shown in Figure 4.1 and Figure 4.2. Looking at these microstructures into more

detail in Figure 4.3, region 1 near the sliding surface of the pin has a dendritic microstructure but in a

darker colour than in region 2. At higher magnification, the dendritic area in region 1 does not have as

much martensite lathes as in region 2. The difference in martensite in the dendritic area explains the rising

microhardness shown in Figure 4.2 from 0.025 mm to 0.325 mm from the sliding surface. Presence of

martensite lathes in the dendritic area in region 1 and 2 is according to [29]. Interdendriric area in region

1 and 2 is expected to have ferrite phase because of low chromium content in QRO90. Going deeper into

the bulk, clear grain boundaries are found and carbides (white particles) start to form in region 3.

Martensite lathes are fully distributed in the grains which explains the highest microhardness at 3.925

mm from the sliding surface. The grain size is larger in region 3 near region 2 but reduces as it goes into

region 4. In region 4, martensite lathes are less than in region 3 but more carbides are formed. The lathes

in region 4 are possibly tempered martensite. Because of this, the microhardness drops to 484 HV at the

end of the microhardness profile.

3

2

1

4

Figure 4.1. The microstructure of a

repair welded tool steel pin at 25x

magnification divided into four regions.

Figure 4.2. Repair welded tool steel pin microhardness profile.

Horizontal bars represent the standard deviation of microhardness

measurement at each point.

Page 27: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

22

Figure 4.3. Microstructures of a repair welded tool steel pin at 500X magnification (left) by optical microscope and

800X magnification (right) by SEM on each region.

Region 1

Region 2

Region 3

Region 4

Region 1

Region 2

Region 3

Region 4

Sliding surface

Page 28: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

23

The EDS spectra obtained from different locations show the elements in those particular

locations. EDS spectrum 1 and spectrum 8 from different locations are shown in Figure 4.4. Spectrum 1

shows that the elements in the matrix are mostly iron characterized by two separate peaks. EDS analysis

on martensite lathe in different regions results in a similar spectrum as spectrum 1. Spectrum 8 shows

that the carbides are made of alloying elements Vanadium, Molybdenum, and Chromium. Looking at the

weight% of spectrum 8, the alloying element with the most content is Vanadium and Chromium is the

least. Based on the peaks and weight% of spectrum 8, Vanadium carbides are present with small amount

of Molybdenum and Chromium forming the carbide particles but this needs to be investigated further.

EDS analysis of carbides in region 3 also results in a similar spectrum as spectrum 8.

Dendritic microstructure in region 1 and region 2 is the result of solidification after the welding

process. Region 3 and region 4 stayed solid during the welding process but they were affected by heat.

The heat from the welding process caused grain growth and diffusion of alloying elements into the solid

solution in region 3. Region 4 was less affected by the heat so the alloying elements were still in the form

of carbides. Less heat effect caused slower grain growth and this was also hindered by the presence of

carbides. During the welding process, the grains in region 3 were transformed into austenite phase

because of the heat. When the welding finished, rapid cooling occurred and transformed austenite into

martensite with possible retained austenite along with martensite lathes. Martensitic transformation in

region 3 occurred because of alloying elements in the solid solution that stabilized austenite phase during

rapid cooling and brought down temperature to start martensite transformation.

Figure 4.4. Spectrum 1 (bottom left) from EDS analysis on the matrix in region 4 and spectrum 8 (bottom right) from

a carbide.

Page 29: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

24

4.1.2. Tribological test result

Sliding between repair welded tool steel and Al-Si coated UHSS at temperature 600 ⁰C has a trend

of rising COF as shown in Figure 4.5 together with worn strips and pins. Sudden COF increase after a short

sliding distance is from the acceleration at the start of sliding. Wear damage on the strip is not visible at

the start of sliding but starts to be visible after some distance where COF is more stable. This characteristic

also appears after a test at different temperatures. After a sudden increase, COF increases gradually to

the end of sliding for every strip including in a test repetition and results in an average COF of 0.39 at test

temperature 600 ⁰C. Wear damage on the strips shows dominant abrasive wear to the end of the sliding

and is supported by no visible material transfer found on the pins by the naked eye.

Figure 4.5. COF as a function of sliding distance (top) with corresponding worn strips (middle) and worn pins

(bottom) after the tribological test at 600 ⁰C. The arrow indicates sliding direction.

Strip number:

3

2

1

Side A

Side A Side B

0

0.1

0.2

0.3

0.4

0.5

0.6

0 20 40 60 80 100 120 140 160 180 200

Co

effi

cien

t o

f Fr

icti

on

Sliding distance (mm)

Strip 1 Strip 2 Strip 3

Page 30: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

25

Looking at the pin surface in more detail under SEM, patches of aluminium from the strip are

present near the leading edge as shown in Figure 4.6. Aluminium patches also appear with a smooth

surface, indicating the patches were load bearing. Wear particles appear more in the centre and near the

trailing edge.

Figure 4.6. SEM micrographs of the pin on side A after the test at temperature 600 ⁰C at location a) at top near the

trailing edge, b) at top in the centre, c) at top near the leading edge, d) in the middle near the trailing edge, e) in the

middle and centre, f) in the middle near the leading edge, g) at bottom near the trailing edge, h) at bottom in the

centre, i) at bottom near the leading edge at 100X magnification.

a) b) c)

d) e) f)

g) h) i)

Page 31: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

26

The SEM micrograph at 800X magnification (Figure 4.7) shows an aluminium patch that appears

to break into smaller wear particles exposing repair welded tool steel surface beneath it, so wear particles

in the centre and near the trailing edge might be from the broken aluminium patches near the leading

edge. This also indicates that the patches were formed by the compaction of wear particles instead of the

adhesion of the transferred material. An EDS spectrum showing that the patches and wear particles are

made of aluminium is shown in Figure 4.7 together with EDS spectrum of repair welded tool steel surface.

Scratches on the repair welded tool steel surface support the hypothesis of abrasive wear mechanism.

EDS spectra for repair welded tool steel surface and transferred material including wear particles, patches,

and lumps after the test at different temperatures are similar to the spectra in Figure 4.7.

Figure 4.7. EDS spectrum of broken aluminium patches (bottom left) and repair welded tool steel surface (bottom

right).

The results of sliding between repair welded tool steel and Al-Si coated UHSS at 700 ⁰C are similar

to the results at 600 ⁰C with slower COF increase in the stable region, except for the result after sliding on

the third strip as shown in Figure 4.8 together with worn strips and pins. Sliding at 700 ⁰C results in an

average COF of 0.38 from the first and second strip in the stable region. The COF reaches around 0.6 after

sliding for 40 mm on the third strip which corresponds to a change in wear mechanism on side A of the

third strip. The mechanism changes into severe ploughing due to the formation of a lump of transferred

material on the pin on side A. The COF rises between 20 mm – 40 mm sliding distance which indicates

Page 32: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

27

that the lump was formed after sliding for 20 mm. The scratch only on side A results in a COF of 0.6 to the

end of the sliding. The strip on side B, however, only shows ploughing damage that causes COF rising to

around 0.8 after reaching around 70 mm sliding distance then the ploughing disappears. This indicates

that a lump was formed on the pin on side B and then was broken off as it continued sliding. This resulted

in the wear mechanism to return to that of the first and second strip. A smaller lump on the pin on side B

is visible in Figure 4.8 which might indicate the broken lump.

Figure 4.8. COF as a function of sliding distance (top) with corresponding worn strips (middle) and worn pins

(bottom) after the tribological test at 700 ⁰C. The arrow indicates sliding direction.

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80 100 120 140 160 180 200

Co

effi

cien

t o

f Fr

icti

on

Sliding distance (mm)

Strip 1 Strip 2 Strip 3

Side A Side B

Side A

Side B

Strip number:

3

2

1

3

2

1

Page 33: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

28

SEM micrographs of the pin on side A in Figure 4.9 show that the lump at the leading edge was

formed by the compaction of wear particles indicating the occurrence of compaction galling. A smoother

surface of the lump indicates that the surface was compacted further by flattening the agglomerated wear

particles similar to the patches in Figure 4.6. The amount of wear particles in the centre and near the

trailing edge are less than in Figure 4.6 because the lump formed at the leading edge was load bearing

during the sliding. Patches near the leading edge at the top and bottom might be formed earlier during

sliding on the first and second strip before a larger lump was formed during sliding on the third strip.

Figure 4.9. SEM micrographs of the pin on side A after the test at temperature 700 ⁰C at location a) at top near the

trailing edge, b) at top in the centre, c) at top near the leading edge, d) in the middle near the trailing edge, e) in the

middle and centre, f) in the middle near the leading edge, g) at bottom near the trailing edge, h) at bottom in the

centre, i) at bottom near the leading edge at 100X magnification, j) in the middle near the leading edge, and k) in the

middle at the leading edge at 800X magnification.

a) b) c)

d) e) f)

g) h) i)

j) k)

Page 34: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

29

A lump of transferred material is not always formed during tests at temperature 700 ⁰C. In Figure

4.10, SEM micrographs of another test pin from test repetition at 700 ⁰C show no lump. These SEM

micrographs show more wear particles across the surface compared to SEM micrographs in Figure 4.6

with the transferred aluminium layer near the leading edge. Unlike in Figure 4.7, there can be another

transferred aluminium layer or repair welded tool steel surface beneath the transferred aluminium layer.

Figure 4.10. SEM micrographs of another pin on side A after test repetition at temperature 700 ⁰C at location a) at

top near the trailing edge, b) at top in the centre, c) at top near the leading edge, d) in the middle near the trailing

edge, e) in the middle and centre, f) in the middle near the leading edge, g) at bottom near the trailing edge, h) at

bottom in the centre, i) at bottom near the leading edge at 100X magnification, in the middle near the leading edge

at the boundary of transferred aluminium layer exposing repair welded tool steel surface (j) and another transferred

aluminium layer (k) beneath it at 800X magnification.

a) b) c)

d) e) f)

g) h) i)

j) k)

Page 35: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

30

At 750 ⁰C, the transferred lump made of compacted aluminium wear particles from the strip

appeared after reaching sliding distance near 100 mm. Figure 4.11 supports this by the rising COF near

100 mm of sliding distance and a change in the strip damage on both sides into ploughing to the end of

the sliding. The COF before ploughing is 0.44 and reaches more than 1 at the end of the sliding. Test

repetition at 750 ⁰C shows a different location for the start of the ploughing which is near the end of the

sliding. This might be due to higher surface roughness for the pins in Figure 4.11.

Figure 4.11. COF as a function of sliding distance (top) with corresponding worn strips (middle) and worn pins

(bottom) after the tribological test at 750 ⁰C. The arrow indicates sliding direction.

SEM micrographs in Figure 4.12 show similar characteristic to Figure 4.9 which have few wear

particles in the centre and near the trailing edge due to the load bearing lump at the leading edge which

reduces the contact area. The lump in Figure 4.12 has smooth and rough surface interchangeably. The

0

0.2

0.4

0.6

0.8

1

1.2

0 20 40 60 80 100 120 140 160 180 200

Co

effi

cien

t o

f Fr

icti

on

Sliding distance (mm)

Strip 1

Side A Side B

Side A

Side B

Page 36: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

31

smooth surface might be from the broken part of brittle intermetallic lump while the rough surface is from

the compaction of wear particles. The mechanism of compaction galling having varying roughness of the

lump has to be studied further.

Figure 4.12. SEM micrographs of the pin on side A after the test at temperature 750 ⁰C at location a) at top near the

trailing edge, b) at top in the centre, c) at top near the leading edge, d) in the middle near the trailing edge, e) in the

middle and centre, f) in the middle near the leading edge, g) at bottom near the trailing edge, h) at bottom in the

centre, i) at bottom near the leading edge at 100X magnification, at the transition from rough to smooth compacted

lump (j), and at the transition from smooth to rough compacted lump (k) at the leading edge at 800X magnification.

In the test repetition at 750 ⁰C, slower lump formation than the first test caused larger pin surface

area to bear the load and wear particles to be distributed across the surface similar to Figure 4.10 with

a) b) c)

d) e) f)

g) h) i)

j) k)

Page 37: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

32

aluminium patches in the centre, transferred aluminium layer near the leading edge, and the lump after

sliding only on one strip indicating a combination of severe adhesion and compaction galling. SEM

micrographs of the pin in the test repetition are in Figure 4.13. Aluminium patches in the centre might be

formed by the compaction of wear particles directly from the Al-Si coated strip or from broken transferred

aluminium layer.

Figure 4.13. SEM micrographs of another pin on side A after test repetition at temperature 750 ⁰C at location a) at

top near the trailing edge, b) at top in the centre, c) at top near the leading edge, d) in the middle near the trailing

edge, e) in the middle and centre, f) in the middle near the leading edge, g) at bottom near the trailing edge, h) at

bottom in the centre, i) at bottom near the leading edge at 100X magnification, j) at the boundary of transferred

aluminium layer near the leading edge, and k) the lump at the leading edge at 800X magnification.

a) b) c)

d) e) f)

g) h) i)

j) k)

Page 38: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

33

The average COF and the pin weight difference after sliding at different temperatures are

presented in Figure 4.14 and Figure 4.15 respectively. The COF at 600 ⁰C and 700 ⁰C in the stable region

are similar because there was no severe adhesion. This is supported by no transferred layer on the pin

after the test at 600 ⁰C and the lump is not always formed after sliding on three strips at temperature 700

⁰C. The stable COF at 750 ⁰C is higher than at lower temperatures because of the severe adhesion leading

to the formation of a lump and ploughing after sliding on the first strip. The COF in Figure 4.14 are similar

to the result in [14] with a COF of 0.4 at temperature 600 ⁰C and 700 ⁰C and reaches 0.45 at temperature

750 ⁰C despite higher hardness due to the welding compared to the original QRO90 used in [14].

Figure 4.14. The average coefficient of friction of sliding between repair welded tool steel and Al-Si coated UHSS at

different temperatures.

Figure 4.15. Weight differences of the pins after tests at different temperatures.

The weight loss at 600 ⁰C is due to material removal caused by abrasive wear and reduction of

pick up of Al-Si coating onto the surface of the pin. The weight gain at 700 ⁰C is the largest because of the

presence of wear particles, transfer layer, and formation of lumps due to less severe adhesion resulting

in slower lump formation and longer contact duration with a larger contact area to adhere more wear

particles across the pins. More severe adhesion at 750 ⁰C causes faster lump formation which quickly

reduces the contact area to adhere wear particles on the pins, so the weight gain is smaller because it

mainly increased from the lump formation during sliding.

0.39 0.380.44

0.00

0.20

0.40

0.60

0.80

1.00

Co

effi

cien

t o

f Fr

icti

on

600 ⁰C 700 ⁰C 750 ⁰C

0.08

-4.12

-1.60

-12.00

-10.00

-8.00

-6.00

-4.00

-2.00

0.00

2.00

4.00

Wei

ght

Dif

fere

nce

(m

g)

600 ⁰C 700 ⁰C 750 ⁰C

Page 39: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

34

4.2. Acoustic emission during sliding between hot work tool steel and UHSS

Figure 4.16 shows the relationship between COF, AE signal in the time domain, and wear damage

of sliding between uncoated UHSS and Toolox44 with parallel surface roughness lay direction to the sliding

direction. Increase in COF up to around 0.2 is accompanied by a gradual increase of AE signal amplitude.

After 0.2 seconds sliding time, it reaches sliding distance 20 mm and up to this point, mild adhesive wear

damage appears on the strip along with the ploughing. Continued sliding caused severe ploughing on the

strip but results in stable COF. The change of wear mechanism where ploughing action is dominant is

indicated by the first burst of the AE signal near 0.2 seconds. The AE signal enters a different region where

burst signals appear after 0.2 seconds of sliding.

Figure 4.16. COF as a function of time (top) and corresponding raw AE signals (middle) from the sliding between

uncoated UHSS and Toolox44 with parallel surface roughness lay direction to the sliding direction. The ruler near the

worn strip is to measure wear progress on the strip (bottom).

0 20 40 60 80 100 120 140 160 180 200

0

0.1

0.2

0.3

0.4

0.5

0.6

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Sliding distance (mm)

Co

effi

cien

t o

f Fr

icti

on

Time (seconds)

Parallel 1 Parallel 2

Side B Parallel 1

Parallel 2 Side B

Parallel 1

Parallel 2

Page 40: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

35

Further analysis of the AE signal in the frequency domain reveals that the AE signal before the

first burst has higher peaks at frequencies above 0.3 MHz compared to the AE signal after the first burst.

After the first burst, the AE signal has more individual peaks that stand out compared to the others across

the frequency. Analysis in the frequency domain is shown in Figure 4.17. The result from test repetition

shows similar change of peaks to the result in Figure 4.17 but the change occurs after the burst around

0.8 seconds which is not the first burst. A change in wear mechanism is not always indicated by the

appearance of a burst signal, it can also be indicated by a change in continuous AE signal amplitude found

in the sliding of Toolox44 with perpendicular surface roughness lay direction.

Figure 4.17. Frequency analysis of AE signal between uncoated UHSS and Toolox44 with parallel surface roughness

lay direction to the sliding direction from the first test. AE signal analysis starts at 0.2 seconds (top) and 0.8 seconds

(bottom) of the sliding time.

Analysis at 0.2 seconds

Analysis at 0.8 seconds

Page 41: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

36

Perpendicular surface roughness lay direction to the sliding direction of Toolox44 on uncoated

UHSS has a relationship between COF, raw AE signals, and wear damage as shown in Figure 4.18. A test in

this condition gives severe and mild scratches. On one strip, both scratches appear up to a sliding distance

around 110 mm. This corresponds to gradually rising COF and an AE signal burst around 1.1 seconds after

sliding. Stable COF is obtained after continued sliding which damages the strip surface with only severe

scratching.

Figure 4.18. COF as a function of time (top) and corresponding raw AE signals (middle) from the sliding between

uncoated UHSS and Toolox44 with perpendicular surface roughness lay direction to the sliding direction. The ruler

near the worn strip is to measure wear progress on the strip (bottom).

Side B Perpendicular 1

Side B Perpendicular 2

0 20 40 60 80 100 120 140 160 180 200

0

0.1

0.2

0.3

0.4

0.5

0.6

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Sliding distance (mm)

Co

effi

cien

t o

f Fr

icti

on

Time (seconds)

Perpendicular 1 Perpendicular 2

Perpendicular 1

Perpendicular 2

Page 42: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

37

The formation of a lump causing severe scratching is detected by the AE signal in further analysis

into the frequency domain in Figure 4.19. The AE signal at 0.2 seconds has peaks at lower frequencies

(below 0.3 MHz) in magnitude lower than 3 x 10-4 Volt. These peaks usually appear in the AE signal before

0.4 seconds except around 0.04 seconds of sliding with high AE signal amplitude. Sliding for a longer time

causes an increase in peak magnitude for one or more peaks at lower frequencies as shown in frequency

analysis result at 1.8 seconds in Figure 4.19 while the peak magnitude at higher frequencies remains

similar during the whole sliding duration. The change of peaks in the frequency domain is correlated to

increasing AE signal amplitude after 0.4 seconds of sliding which is not due to acceleration. This does not

mean that after 0.4 seconds of sliding the frequency of AE signal always has peaks as in the analysis result

at 1.8 seconds. It only means that the higher magnitude peaks at lower frequencies start to appear much

more frequently after 0.4 seconds of sliding. Some frequency analyses after 0.4 seconds give peaks with

magnitude lower than 3 x 10-4 Volt at lower frequencies. This varying AE signal characteristic in the

frequency domain might be due to the lump evolution over sliding distance with a cycle of lump formation

and breakage starting after 0.4 seconds of sliding. AE signal with higher magnitude peaks at lower

frequencies around 0.04 seconds of sliding might be due to defects of perpendicular surface roughness

lay direction that plough in a similar way to lumps after sliding for a short distance. The surface

smoothened immediately and is characterized by decreasing AE amplitude until 0.4 seconds of sliding with

frequency characteristic having lower peak magnitude at lower frequencies.

Figure 4.19. Frequency analysis of AE signal between uncoated UHSS and Toolox44 with perpendicular surface

roughness lay direction to the sliding direction from the first test. AE signal analysis starts at 0.2 seconds (top) and

1.8 seconds (bottom) of the sliding time.

Analysis at 0.2 seconds

Analysis at 1.8 seconds

Page 43: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

38

Test repetition for this sliding does not result in a similar behaviour. Although peaks at lower

frequencies are usually higher than at higher frequencies, the height ratio between these peaks remains

similar across the frequencies as shown in Figure 4.20. There are some changes between these peaks but

it only remains for a short period and the peaks keep changing. This might be explained by a more uniform

load bearing pin surface compared to the first test. The load was carried by the lumps and pin surface area

between the lumps. This also explains the unworn strip surface between two scratches. The peaks in the

frequency domain keep changing depending on the dominant load bearing surface whether it is the lumps

or pin surface area.

Figure 4.20. Frequency analysis of AE signal between uncoated UHSS and Toolox44 with perpendicular surface

roughness lay direction to the sliding direction from the second test. AE signal analysis starts at 0.4 seconds (top)

and 1.8 seconds (bottom) of the sliding time.

Sliding on as-delivered Al-Si coated UHSS gives the highest stable COF 0.86 compared to other

tests in AE study. Although the COF increases during sliding for 20 mm, the AE signal amplitude remains

relatively flat after the acceleration as shown in Figure 4.21. Ploughing of the Al-Si coating started at the

same time as the sliding. The ploughing became more severe along the strip which is indicated by the

appearance of many bursts after 0.5 seconds but the COF remains stable. The burst region of the AE signal

might indicate a cycle of aluminium lump breakage and formation on the pins. Lump breakage is indicated

by a burst while lump formation is indicated by a continuous signal. The amplitude of a burst signal reaches

1 V, much higher than the sliding on uncoated UHSS. This might mean that the critical size for breakage

of the aluminium lump was bigger than in the case of the steel lump after sliding on uncoated UHSS. The

continuous signal comes from deformation of the aluminium coating and the transferred lump before

breakage.

Analysis at 0.4 seconds

Analysis at 1.8 seconds

Page 44: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

39

Figure 4.21. COF as a function of time (top) and corresponding raw AE signal (middle) from the sliding between as-

delivered Al-Si coated UHSS and Toolox44 with parallel surface roughness lay direction to the sliding direction. The

ruler near the worn strip is to measure wear progress on the strip (bottom).

The peaks in the frequency domain shown in Figure 4.22 do not change much and are similar between

the AE signal analyzed at different sliding times.

Figure 4.22. Frequency analysis of AE signal between as-delivered Al-Si coated UHSS and Toolox44 with parallel surface

roughness lay direction to the sliding direction. AE signal analysis starts at 0.5 seconds of the sliding time.

0 20 40 60 80 100 120 140 160 180 200

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Sliding distance (mm)

Co

effi

cien

t o

f Fr

icti

on

Time (seconds)

Analysis at 0.5 seconds

Page 45: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

40

Sliding on heat-treated Al-Si coated UHSS gives stable COF after sliding for 5.5 mm. Although the

COF is smaller than sliding on as-delivered Al-Si coated UHSS, it corresponds to the highest AE signal

amplitude compared to other sliding combinations in AE study as shown in Figure 4.23. The strongest AE

signal was also obtained with 20 dB amplification while other AE signals were amplified by 40 dB. The

strong AE signal comes from the ploughing of Toolox44 surface by hard asperities on intermetallic (FeAlSi)

coating due to harder heat-treated Al-Si coated UHSS. The FeAlSi intermetallics in the coating have

hardness levels around 800-850 HV [7] which is much higher compared to Toolox44 which has a hardness

of 410 - 475 HBW [55] (around 440 HV). Large deformation of Toolox44 surface due to ploughing generated

strong AE signal according to [36].

Figure 4.23. COF as a function of time (top) and corresponding raw AE signal (middle) from the sliding between heat-

treated Al-Si coated UHSS and Toolox44 with parallel surface roughness lay direction to the sliding direction. The ruler

near the worn strip is to measure wear progress on the strip (bottom).

0 20 40 60 80 100 120 140 160 180 200

0

0.1

0.2

0.3

0.4

0.5

0.6

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Sliding distance (mm)

Co

effi

cien

t o

f Fr

icti

on

Time (seconds)

Page 46: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

41

The characteristic of the AE signal is shown in Figure 4.24. Only from the sliding on heat-treated

Al-Si coated UHSS, the frequency characteristic of the AE signal during acceleration differs. During

acceleration, the peaks in Figure 4.24 are distributed across the frequency with comparable height. As

Toolox44 continued sliding, the height of the peaks goes lower as the frequency goes higher.

Figure 4.24. Frequency analysis of AE signal between heat-treated Al-Si coated UHSS and Toolox44 with parallel

surface roughness lay direction to the sliding direction. AE signal analysis starts at 0.05 seconds (top) and 1.2 seconds

(bottom) of the sliding time.

The result of frequency analysis is affected by the magnitude of noise which was measured during

stationary contact. Amplification of 40 dB has noise with comparable frequency components from 0.1 – 1

MHz, therefore the result of frequency analysis is not much affected by height differences between the

peaks. However, amplification of 20 dB causes some peaks to stand out as shown in Figure 4.25. Although

they are standing out, the frequency component of the AE signal from the sliding on heat-treated Al-Si

coated UHSS has much higher peak magnitude at lower frequencies. Peak magnitude difference between

the sliding and the noise is in the order of three meaning that the standing peaks due to noise do not

affect the frequency analysis in Figure 4.24 too much.

The effect of noise can also be evaluated in the time domain. Ideally, the AE signal amplitude is

zero in stationary contact but the presence of noise generated a small amplitude before the sliding

started. Figure 4.26 shows the AE signal in the time domain before the sliding started and during initial

sliding which is indicated by a gradual increase of the amplitude. Before sliding, the amplitude reaches

0.0014 V which is smaller than continuous signal amplitude around 0.01 V during initial sliding. From this

comparison in the time domain, the effect of noise in the result reaches 14%. The AE signal in Figure 4.26

is the same as in Figure 4.16 for the first test. This signal is used to evaluate the effect of noise because

the amplitude of the continuous signal is the smallest compared to other sliding signals. AE continuous

signals with larger amplitude from different sliding combinations give a smaller noise effect because the

noise amplitude is similar, which is around 0.0014 V.

Analysis at 0.05 seconds

Analysis at 1.2 seconds

Page 47: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

42

Figure 4.25. Noise in the frequency domain from AE measurement using pre-amplifier 40 dB (top) and 20 dB

(bottom).

Figure 4.26. Noise in the time domain from AE measurement using pre-amplifier 40 dB.

Page 48: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

43

From the test on Al-Si coated UHSS, the COF is not always proportional to AE signal strength.

Higher COF and smaller continuous AE signal amplitude is found during sliding on as-delivered Al-Si coated

UHSS while the sliding on heat-treated Al-Si coated UHSS gives the opposite result for COF and AE signal

strength. From the test on uncoated UHSS, the COF is proportional to AE signal strength. Figure 4.16 shows

that the second test has larger AE signal amplitude and higher stable COF of 0.49 than the AE signal

amplitude from the first test with a lower stable COF of 0.32. Figure 4.18 also shows larger AE signal

amplitude and higher stable COF during sliding for 0.8 seconds and after 1.4 seconds in the second test

compared to the first test. This result is similar to what was found in [38] where sliding on one or two soft

surfaces results in a negative relation between COF and AE RMS and sliding on both hard surfaces results

in a positive relation. Comparison of COF between different sliding combinations is presented in Figure

4.27.

Figure 4.27. COF of sliding between Toolox44 and UHSS. COF of sliding on as-delivered and heat-treated Al-Si coated

UHSS do not have standard deviation due to no test repetition.

The surface topography lay directions of Toolox44 with surface roughness 300 – 400 nm have a

small effect on COF and AE signal strength. This is due to the formation of lumps during the sliding that

scratch and control the COF and AE strength until the end of the sliding. Although continuous AE strength

is similar, the appearance of burst signals and gradual change in amplitude can mean a change in wear

mechanism. Connecting wear mechanism in this case to galling mechanism according to [7], a change

where peaks at higher frequencies go shorter or peaks at lower frequencies go higher indicates transition

from stage II to stage III. Stage II is the stage where conglomerates of wear debris are formed in case of

compaction galling or the stage where the transfer layer is formed in case of adhesive galling. Stage III is

the stage where the conglomerates or transfer layers become obstacles for wear particles to move out of

the contact. It is also illustrated in Figure 1.3 that a small lump has formed in stage III before severe galling

in stage IV. The transition is supported by examination of the worn strip surfaces in Figure 4.16 which

shows that transition of wear damage to only one scratch corresponds with the change of peaks at higher

frequencies. Transition of wear damage to one scratch indicates galling mechanism in stage III where the

lump accumulates wear particles and grows into a bigger lump. This also means that tribological condition

in the stamping process can be monitored by AE signals, both in the time and frequency domain. Thinking

0.400.36

0.86

0.45

0.00

0.20

0.40

0.60

0.80

1.00

Co

effi

cien

t o

f Fr

icti

on

Parallel Perpendicular

As-obtained Al-Si coated UHSS Heat treated Al-Si coated UHSS

Page 49: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

44

that a lump is a surface defect, shortened peaks at higher frequencies after the transition of wear damage

is according to [37]. The authors found that AE signals from smoother surfaces have higher frequency

components than that from rougher surfaces. A rough surface is a more defective surface than a smooth

surface, therefore peaks at higher frequencies shorten if the surface becomes more defective, whether

because of lump formation or rough surface.

Comparison between Toolox44 pin weight differences of different sliding combinations is

presented in Figure 4.28 and the images of worn Toolox44 pins are in Figure 4.29. Toolox44 pins sliding on

as-delivered Al-Si coated UHSS strip gained significant weight of 49.42 mg and is omitted in Figure 4.28

for a clearer comparison between other conditions.

Figure 4.28. Weight differences of Toolox44 pins of different sliding combinations. Toolox44 pin weight difference after

sliding against as-delivered Al-Si coated UHSS is not presented because it is too large compared to other sliding

combinations.

-0.55 -0.38 -0.24

-12.00

-10.00

-8.00

-6.00

-4.00

-2.00

0.00

2.00

4.00

Wei

ght

dif

fere

nce

(m

g)

Parallel Perpendicular Heat-treated AlSi UHSS

Page 50: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

45

The largest weight gain can be explained by attachment of a big aluminium lump on the pin in

Figure 4.29(e). The smallest weight gain from the sliding against heat-treated Al-Si coated UHSS can also

be explained in Figure 4.29(f) by Toolox44 pins with no lump. Figure 4.29(c) and 4.29(d) show surface

roughness in perpendicular lay direction acts as a wear particle trap, but somehow it does not form a

bigger lump compared to Figure 4.29(a) and 4.29(b). This is contradicting the result from [7] where parallel

surface roughness lay direction should reduce galling tendency. This perhaps is caused by a test condition

at room temperature while [7] was performed at high temperatures. As a result of lump formation,

Toolox44 pins with parallel surface roughness lay direction has higher weight gain than the pins with

perpendicular lay direction. Except Figure 4.29(e), the test did not form the lump at the leading edge as

expected in [14] which conducted the test at high temperatures.

Figure 4.29. The surface of the pins after the test for AE study. The arrow indicates sliding direction. The pins with

parallel surface roughness lay direction in (a) and (b) and with perpendicular surface roughness lay direction in (c)

and (d) sliding against uncoated UHSS strip. The pins sliding against as-delivered and heat-treated Al-Si coated UHSS

in (e) and (f).

a) b)

c) d)

e) f)

Side A Side B Side A Side B

Side A Side B Side A Side B

Side A Side B Side A Side B

Page 51: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

46

5. Conclusions

Tribological studies to investigate the effect of repair welded tool steel on friction and wear during

sliding against Al-Si coated ultra-high strength boron steel (UHSS) as well as studies using acoustic

emission measurements in a room temperature strip drawing test with tool steel, uncoated and Al-Si

coated UHSS has been carried out. The salient conclusions from this work are as follows:

The average COF during sliding between repair welded tool steel and Al-Si coated UHSS at a

temperature of 600 ⁰C and 700 ⁰C are 0.39 and 0.38 respectively, which are lower than the average COF

of 0.44 at 750 ⁰C. This is due to abrasive wear mechanism at 600 ⁰C and galling mechanism with slower

material transfer lump formation at 700 ⁰C. Because of these mechanisms, a lower stable COF is obtained

when a lump is not formed. At 750 ⁰C, severe adhesion causes a higher stable COF and galling mechanism

with faster lump formation leads to increasing COF due to the ploughing of the Al-Si coated UHSS surface.

Repair welded tool steel shows the highest weight gain at 700 ⁰C because of the presence of wear

particles, transfer layer, and formation of lumps. Less adhering wear particles after the sliding at 750 ⁰C

are because of the load bearing lump of material that prevents the contact of remaining repair welded

tool steel surface area. Abrasive wear at 600 ⁰C causes scratches on tool surface and does not cause

adhesion of many wear particles nor form transfer layers which results in a weight loss.

Between hot work tool steel and uncoated UHSS, a change in wear mechanism is indicated by an

acoustic emission burst signal or a gradual amplitude change in the time domain. A change in wear

mechanism due to the formation of a lump of transferred material generates AE signals with peaks at

higher frequencies above 0.3 MHz to shorten or peaks at lower frequencies below 0.3 MHz to rise.

Different lay directions on hot work tool steel do not affect the COF, AE signals, and wear damage on the

strip to a great extent. Hot work tool steel sliding on uncoated UHSS has a proportional relationship

between the COF and AE signal strength while sliding on Al-Si coated UHSS has an inversely proportional

relationship.

Repair welded tool steel can be used for hot stamping process with the same performance as the

original tool steel or even better. Acoustic emission measurement can be used to monitor tool condition

and detect a change of wear mechanism before causing severe ploughing of the Al-Si coated UHSS. Based

on the result of this work, it is suggested that the tools are replaced once a change in the AE signal is

detected to prevent further and deeper scratches on the workpiece and minimize scrap. A note has to be

taken that the result of AE study was obtained from tests at room temperature. Further study using a

small scale forming operation at high temperatures has to be conducted to confirm the application of

acoustic emission measurement in hot forming processes.

Page 52: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

47

6. References

[1] H. Karbasian and A. E. Tekkaya. A review on hot stamping. Journal of Materials Processing

Technology, 210(15):2103–2118, 2010.

[2] L. Pelcastre, J. Hardell, and B. Prakash. Tribological behaviour of Zn coated UHSS sliding against hot-

work tool steel at high temperatures. Wear, 376-377:423-432, 2017.

[3] Groover, M. P. 2010. Fundamentals of Modern Manufacturing: Materials, Processes, and Systems.

4th ed. Upper Saddle River, New Jersey: Pearson Prentice-Hall.

[4] Kalpakjian, S. and Schmid, S. R. 2010. Manufacturing Engineering and Technology. 6th ed. Upper

Saddle River, New Jersey: Pearson Prentice-Hall.

[5] S. Mozgovoy. High temperature friction and wear in press hardening. Licentiate thesis, Lulea

University of Technology, Lulea, Sweden, 2014.

[6] M. Geiger, M. Merkleinand J. Lechler. Determination of tribological conditions within hot stamping.

Production Engineering 2, 269-276, 2008.

[7] L. Pelcastre. High Temperature Galling Influencing Parameters and Mechanism. Ph.D. thesis, Luleå

University of Technology, Luleå, Sweden, 2013.

[8] A. Ghiotti, S. Bruschi, F. Sgarabotto, and P.F. Bariani. Tribological performances of Zn-based coating

in direct hot stamping. Tribology International, 78:142-151, 2014.

[9] A. Ghiotti, S. Bruschi, and F. Medea. Comparison of tribological and wear performances of AlSi and

Zn coatings in hot stamping of boron steel sheets. Wear, 332-333: 810-821, 2015.

[10] J. Kondratiuk and P. Kuhn. Tribological investigation on friction and wear behaviour of coatings for

hot sheet metal forming. Wear, 270(11-12):839–849, 2011.

[11] ASTM Standard G40. (2010), “Standard Terminology Relating to Wear and Erosion,”

[12] P. Schwingenschlögl, P. Niederhofer, and M. Merklein. Investigation on basic friction and wear

mechanisms within hot stamping considering the influence of tool steel and hardness. Wear, 426-

427:378-389, 2019.

[13] L. Pelcastre, J. Hardell, A. Rolland, and B. Prakash. Influence of microstructural evolution of Al-Si

coated UHSS on its tribological behaviour against tool steel at elevated temperatures. Journal of

Materials Processing Technology, 228:117-124, 2016.

[14] L. Deng, L. Pelcastre, J. Hardell, B. Prakash, and M. Oldenburg. Experimental Evaluation of Galling

Under Press Hardening Conditions. Tribology Letters, 66:93, 2018.

[15] S. Mozgovoy, J. Hardell, L. Deng, M. Oldenburg, and B. Prakash. Tribological behaviour of tool steel

under press hardening conditions using simulative tests. Journal of Tribology, 140:1-11, 2017.

[16] J. Venema, J. Hazrati, D.T.A. Matthews, R.A. Stegeman, and A.H. van den Boogard. The effects of

temperature on friction and wear mechanisms during direct press hardening of Al-Si coated ultra-high

strength steel. Wear, 406-407:149-155, 2018.

[17] A. Yanagida, T. Mukai, and K. Matsumoto. Adhesion behavior of aluminum-coated 22MnB5 steel in

hot flat drawing test under dry condition with coated tools. Procedia Manufacturing, 15: 1041-1046,

2018.

[18] A. Ghiotti, S. Bruschi, F. Medea, and A. Hamasaiid. Tribological behavior of high thermal conductivity

steels for hot stamping tools. Tribology International, 97:412-422, 2016.

[19] K. Uda, A. Azushima, and A. Yanagida. Development of new lubricants for hot stamping of Al-coated

22MnB5 steel. Journal of Materials Processing Technology, 228: 112-116, 2016.

Page 53: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

48

[20] A. Azushima, K. Uda, and A. Yanagida. Friction behavior of aluminum-coated 22MnB5 in hot stamping

under dry and lubricated conditions. Journal of Materials Processing Technology, 212(5): 1014-1021,

2012.

[21] H. Kashani, A. Amadeh, and M.R. Vatanara. Improvement of wear resistance of hot working tool steel

by hardfacing Part 2 – Case study. Materials Science and Technology, 24(3):356-360, 2008.

[22] B. Venkatesh, K. Sriker, and VSV Prabhakar. Wear characteristics of hardfacing alloys: state-of-the-art.

Procedia Materials Science, 10: 527-532, 2015.

[23] K. Lee, K. H. Ko, J. H. Kim, G. G. Kim, and S. Kim. Effects of temperature and sliding distance on the

wear behavior of austenitic Fe-Cr-C-Si hardfacing alloy. Tribology Letters, 26(2):131-135, 2007.

[24] K. Lee, S. Lee, Y. Kim, H.S. Hong, Y. Oh, and S. Kim. The effects of additive elements on the sliding wear

behavior of Fe-base hardfacing alloys. Wear, 255(1-6): 481-488, 2003.

[25] J. W. Yoo, S. H. Lee, C.S. Yoon, and S. J. Kim. The effect of boron on the wear behavior of iron-based

hardfacing alloys for nuclear power plants valves. Journal of Nuclear Materials, 352(1-3): 90-96, 2006.

[26] J. Kim, G. Kim, and S. Kim. The effect of manganese on the strain-induces martensitic transformation

and high temperature wear resistance of Fe-20Cr-1C-1Si hardfacing alloy. Journal of Nuclear

Materials, 289(3):263-269, 2001.

[27] H. Kashani, A. Amadeh, and A. Ohadizadeh. Effect of temperature on the strain induced γ→ε phase

transformation in Stellite 21 during wear test. Materials Science and Engineering A, 435-436: 474-477,

2006.

[28] H. Kashani, A. Amadeh and H. M. Ghasemi. Room and high temperature wear behaviours of nickel

and cobalt base weld overlay coatings on hot forging dies. Wear, 262(7-8):800-806, 2007.

[29] A. Gualco, H.G. Svoboda, E.S. Surian, and L.A. de Vedia. Effect of welding procedure on wear behaviour

of a modified martensitic tool steel hardfacing deposit. Materials and Design, 31: 4165-4173, 2010.

[30] H. Kashani, A. Amadeh, and M. Farhani. Improvement of wear resistance of hot working tool steel by

hardfacing Part 1 – Effect of microstructure and hardness, Materials Science and Technology,

23(2):165-170, 2007.

[31] I. Ubhayaratne. Audio-Based Signal Extraction Techniques For Stamping Tool Condition Monitoring.

Ph.D. thesis, Deakin University, Melbourne, Australia, 2017.

[32] I. Ubhayaratne, M. P. Pereira, Y. Xiang, and B.F. Rolfe. Audio signal analysis for tool wear monitoring

in sheet metal stamping. Mechanical System and Signal Processing, 85:809-826, 2017.

[33] V. V. Shanbhag, B. F. Rolfe, N. Arunachalam, and M. P. Pereira. Investigating galling wear behaviour in

sheet metal stamping using acoustic emissions. Wear, 414-415:31-42, 2018.

[34] J.R. Mathews. Acoustic emission. Number ISBN 0-8031-0116-3. Gordon and Breach Publishers Inc.,

1983.

[35] Wang, L. 2011. Foundation of Stress Waves. Kidlington, Oxford: Elsevier.

[36] Baranov, V., Kudryavtsev, E., Saryachev, G.A., and Schavelin, V.M. 2007. Acoustic Emission in Friction.

Kidlington, Oxford: Elsevier.

[37] K. Asamene and M. Sundaresan, Analysis of experimentally generated friction related acoustic

emission signals. Wear, 296(1-2): 607-618, 2012.

[38] P. Tian, Y. Tian, L. Shan, Y. Meng, and X. Zhang. A correlation analysis method for analyzing tribological

states using acoustic emission, frictional coefficient, and contact resistance signals. Friction, 3(1):36-

46, 2015.

[39] A. Hase, H. Mishina, and M. Wada. Correlation between features of acoustic emission signals and

mechanical wear mechanisms. Wear, 292-293:144-150, 2012.

Page 54: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

49

[40] D. Baccar and D. Söffker. Wear detection by means of wavelet-based acoustic emission analysis.

Mechanical Systems and Signal Processing, 60-61:198-207, 2015.

[41] B. Behrens, S. Hübner, and K. Wölki. Acoustic emission – A promising and challenging technique for

process monitoring in sheet metal forming. Journal of Manufacturing Processes, 29:281-288, 2017.

[42] H. Benabdallah. Friction wear and acoustic emissions of some plastics sliding against Si3N4. Wear,

264(1-2):152-156, 2008.

[43] H. S. Benabdallah & D. A. Aguilar. Acoustic Emission and Its Relationship with Friction and Wear for

Sliding Contact, Tribology Transactions, 51(6):738-747, 2008.

[44] V.E. Rubtsov, E.A. Kolubaev, A.V. Kolubaev, and V.L. Popov. Using Acoustic Emission for the Analysis

of Wear Wear Processes during Sliding Friction. Technical Physics Letter, 39(2):223-225, 2013.

[45] A. Hase, H. Mishina, and M. Wada. Fundamental study on early detection of seizure in journal bearing

by using acoustic emission technique. Wear, 346-347:132-139, 2016.

[46] L. Guolu, X. Zhonglin, D. Tianshun, W. Haidou, L. Jinhai, and K. Jiajie. Acoustic Emission Investigation

of Rolling/Sliding Contact Fatigue Failure of NiCr-Cr3C2 Coating. Journal of Thermal Spray Technology,

25(7)-1365-1372, 2016.

[47] M. Yahiaoui, F. Chabert, J.-Y. Paris, V. Nassiet, and J. Denape. Friction, acoustic emission, and wear

mechanisms of a PEKK polymer. Tribology International, 132:154-164, 2019.

[48] E.A. Kolubaev, A.V. Kolubaev, and O.V. Sizova. Analysis of acoustic emission during sliding friction of

manganese steel. Technical Physics Letters, 36(8):762-765, 2010.

[49] C.T. Sindi, M.A. Najafabadi, M. Salehi. Tribological Behavior of Sheet Metal Forming Process Using

Acoustic Emission Characteristics. Tribology Letters, 52(1):67-79, 2013.

[50] F. Saeidi, S.A. Shevchik, and K. Wasmer. Automatic detection of scuffing using acoustic emission.

Tribology International, 94:112-117, 2016.

[51] Heath, M.T. 1997. Scientific Computing, An Introduction Survey. 2nd ed. McGraw-Hill.

[52] F.J. Harris. On the Use of Windows for Harmonic Analysis with the Discrete Fourier Transform.

Proceedings of The IEEE, 66(1):51-83, 1978.

[53] QRO90 Catalog [54] S. Schnabel. High Frequency Elastic Wave Emission Caused by a Single Elastohydrodynamically

Lubricated Contact Fundamentals Sources and Principles. Doctoral Ph.D. thesis, Luleå University of Technology, Luleå, Sweden, 2016.

[55] Toolox 44 datasheet.

Page 55: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

50

Appendix A: Special pin holder technical drawing.

M4

Special pin holder Original pin holder

Page 56: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

51

Ave

rag

e

of 3

pin

s

Sta

nd

ard

De

via

tion

of 3

pin

s

Ro

w 1

Ro

w 2

Ro

w 3

Ave

rag

eS

tan

da

rd D

evia

tion

Ro

w 1

Ro

w 2

Ro

w 3

Ave

rag

eS

tan

da

rd D

evia

tion

Ro

w 1

Ro

w 2

Ro

w 3

Ave

rag

eS

tan

da

rd D

evia

tion

0.0

25

63

35

98

64

86

26

26

75

16

73

66

36

96

48

68

36

15

68

86

62

41

66

13

5

0.1

63

36

58

67

36

55

20

73

37

39

71

67

29

12

67

86

73

69

46

82

11

68

93

8

0.1

75

68

86

53

65

36

65

20

68

36

83

74

57

04

36

65

87

16

71

06

95

32

68

82

0

0.2

56

94

66

86

78

68

01

37

57

71

07

10

72

62

76

94

69

46

83

69

06

69

92

4

0.3

25

67

37

39

67

86

97

37

65

87

10

70

56

91

29

69

46

78

69

96

90

11

69

33

0.6

25

65

36

73

72

16

82

35

72

77

16

75

77

33

21

75

77

45

69

97

34

31

71

63

0

0.9

25

68

86

68

66

86

75

12

69

46

94

71

06

99

96

99

67

36

68

68

01

76

85

13

1.2

25

63

36

73

74

56

84

57

70

57

10

70

57

07

36

83

68

36

68

67

89

68

91

5

1.5

25

67

87

05

66

36

82

21

70

57

05

73

97

16

20

70

56

88

72

17

05

17

70

11

7

1.8

25

69

96

88

67

36

87

13

69

47

05

75

17

17

30

76

37

63

72

17

49

24

71

73

1

2.1

25

63

86

73

67

36

61

20

71

66

88

73

97

14

26

73

37

51

73

97

41

97

06

41

2.4

25

64

36

43

70

56

64

36

73

37

57

72

17

37

18

71

67

45

72

17

27

16

70

94

0

2.7

25

70

57

21

66

86

98

27

67

37

16

71

07

00

23

69

46

29

68

86

70

36

68

91

6

3.0

25

71

06

88

71

67

05

15

68

37

51

72

77

20

34

69

96

73

67

86

83

14

70

31

9

3.3

25

68

86

99

71

06

99

11

72

17

51

76

97

47

24

73

97

16

71

07

22

15

72

32

4

3.6

25

70

57

39

71

07

18

18

75

17

57

78

27

63

16

75

17

10

74

57

35

22

73

92

3

3.9

25

67

86

68

71

66

87

25

73

97

51

71

67

35

18

73

37

69

72

77

43

23

72

23

0

4.2

25

71

66

73

69

46

94

22

73

37

33

72

77

31

36

99

68

87

16

70

11

47

09

20

4.5

25

70

56

78

66

86

84

19

67

86

78

68

86

81

66

68

69

96

73

68

01

76

82

2

4.8

25

66

36

38

67

36

58

18

71

66

99

66

86

94

24

64

36

48

63

86

43

56

65

26

5.1

25

66

86

43

68

36

65

20

67

86

68

64

36

63

18

60

26

11

60

66

06

56

45

33

5.4

25

64

36

43

66

86

51

14

67

36

29

65

36

52

22

61

16

20

61

16

14

56

39

22

5.7

25

66

36

48

60

26

38

32

64

85

98

62

46

23

25

61

56

48

60

26

22

24

62

89

6.0

25

63

36

48

62

46

35

12

62

06

15

58

56

07

19

60

66

24

59

36

08

16

61

61

6

7.0

25

60

66

24

62

96

20

12

54

55

85

58

55

72

23

62

96

06

62

96

21

13

60

42

8

8.0

25

63

36

43

60

26

26

21

56

15

53

55

75

57

46

38

62

46

15

62

61

26

03

40

9.0

25

58

16

48

60

66

12

34

58

15

34

58

95

68

30

62

46

20

62

46

23

26

01

29

10

.02

55

45

57

75

45

55

61

85

38

54

55

69

55

11

65

98

56

95

89

58

51

55

64

19

11

.02

55

27

52

04

96

51

41

64

65

46

54

35

45

51

75

77

58

95

81

58

26

51

76

4

12

.02

54

87

51

04

60

48

62

54

90

46

84

77

47

81

14

90

50

34

74

48

91

54

84

5

Hard

ness (H

V)

Pin

1P

in 2

Pin

3

De

pth

Fro

m

Su

rfac

e

(mm

)

Appendix B: Repair welded tool steel pin microhardness data.

Page 57: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

52

14445.1614444.98

14340.0714340.04

14445.1414444.97

14340.0814340.04

14445.1514444.98

14340.0814340.02

14487.1614487.00

14445.1314445.20

14487.1814487.01

14445.1514445.18

14487.1714487.03

14445.1614445.20

14508.0114508.08

14603.8714603.93

14508.0114508.08

14603.8614603.93

14507.9914508.07

14603.8614603.96

14556.4814571.68

14559.6314560.77

14556.5314571.70

14559.6214560.77

14556.5214571.70

14559.6614560.79

13971.4813972.50

14206.7014208.32

13971.4613972.49

14206.7214208.31

13971.4913972.50

14206.7114208.33

15203.2315206.11

15208.1315209.03

15203.2315206.12

15208.1415209.03

15203.2115206.11

15208.1415209.03

0.02

0.010.01

0.02

0.030.01

0.020.01

14603.94

14560.78

14603.86

14559.64

Stand

ard

Deviatio

nA

fter TestA

verageStan

dard

Deviatio

n

14340.030.01

Stand

ard

Deviatio

nB

efore Test

Side A

Side B

Pin

We

ight (m

g)

Average

Stand

ard

Deviatio

nA

fter TestA

verageStan

dard

Deviatio

n

We

ight D

ifferen

ces

Side A

Side B

Average

Befo

re TestA

verage

0.08

-4.12

-1.600.91

7.39

0.10

-1.61

215203.22

15206.11-2.89

-0.89

0.020.01

0.010.01

0.010.01

0.010.00

14208.32

15209.03

750 ⁰C

113971.48

13972.50

-0.07

700 ⁰C

-1.0214206.71

15208.14

-0.08

214556.51

14571.69-15.18

-1.14

0.011

14508.0014508.08

0.170.04

214487.17

14487.010.16

-0.05

0.010.01

0.0114445.19

14340.08

14445.15

0.01

0.010.02

600 ⁰C

114445.15

14444.98

Temp

erature

Test

Nu

mb

er

1513.19

414.19

2486.85

485.55

3470.90

472.36

1441.44

436.99

2435.24

402.34

3449.50

469.44

1448.51

464.44

2516.00

497.25

3483.48

438.29

1474.61

445.89

2442.80

548.55

3583.07

472.23

1718.14

793.16

2781.63

716.24

3729.02

801.88

1490.83

476.54

2486.46

533.80

3481.68

565.18

47.13

44.95

21.36

7.15

33.75

73.54

33.95

4.58

29.54

53.32

442.06

457.37490.31

436.26

Surface R

ou

ghn

ess (nm

)

37.97

33.56

Side A

Average

Side A

Stand

ard D

eviation

Side A

Side B

Average

Side B

Stand

ard D

eviation

Side B

482.66

500.16

742.93

486.32

466.66

488.89

770.43

525.17

Temp

erature

Test

Nu

mb

erP

ositio

n

600 ⁰C

700 ⁰C

750 ⁰C

121212

Appendix C: Repair welded tool steel pin surface roughness and weight.

Page 58: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

53

14095.6814095.96

13712.7413713.18

14095.6914095.94

13712.7513713.20

14095.6814095.94

13712.7413713.18

14100.3014100.94

13930.0613930.91

14100.3114100.94

13930.0513930.90

14100.3214100.94

13930.0613930.90

13981.1513981.33

14050.8214051.11

13981.1313981.33

14050.8114051.11

13981.1313981.33

14050.8114051.13

14133.1114133.63

14275.0714275.58

14133.1014133.64

14275.0714275.57

14133.0914133.63

14275.0814275.58

14363.6414461.86

14383.9214384.52

14363.6314461.87

14383.9214384.53

14363.6514461.87

14383.9214384.53

14204.0114204.20

13648.2313648.54

14204.0214204.19

13648.2413648.53

14204.0214204.21

13648.2213648.52

0.01-0.18

-0.30

-49.4269.03

-0.240.08

0.01-98.23

-0.61

114204.02

0.0114204.20

0.0113648.23

0.0113648.53

14383.920.00

14384.531

14363.640.01

14461.870.01

As-d

elivered A

l-Si

coated

UH

SS

Heat-treated

Al-Si

coated

UH

SS

0.16

214133.10

0.0114133.63

0.0114275.07

0.0114275.58

0.01

0.0114051.12

0.01-0.19

-0.30

-0.38

-0.53-0.50

0.01-0.63

-0.85

Perpen

dicu

lar

113981.14

0.0113981.33

0.0014050.81

-0.550.25

214100.31

0.0114100.94

0.0013930.06

0.0113930.90

13712.740.01

13713.190.01

Co

nd

ition

Test

Nu

mb

er

Pin

We

ight (m

g)W

eigh

t Diffe

rences

Side A

Side B

Befo

re TestA

verageStan

dard

Deviatio

nA

fter TestStan

dard

Deviatio

nA

verage

-0.26-0.44

Parallel

114095.68

0.0114095.95

0.01

Stand

ard

Deviatio

nB

efore Test

Average

Stand

ard

Deviatio

nA

verageA

fter TestA

verageStan

dard

Deviatio

nSid

e ASid

e B

1386.04

317.39

2421.18

292.00

3389.29

319.76

1381.87

317.97

2346.76

326.12

3345.49

298.10

1348.63

307.01

2297.85

319.14

3339.47

286.69

1397.46

363.58

2375.88

372.91

3334.12

296.59

1352.23

382.11

2407.46

357.75

3334.89

363.19

1389.27

283.96

2334.28

363.61

3310.83

287.79

40.26311.79

44.92

As-d

elivered A

l-Si

coated

UH

SS

Heat-treated

Al-Si

coated

UH

SS1

344.79

1364.86

37.90367.68

12.79

16.40

2369.15

32.20

Parallel

1398.84

19.42309.72

15.39

344.3641.63

Perpen

dicu

lar

1328.65

27.06304.28

2358.04

20.65314.06

14.41

Co

nd

ition

Test

Nu

mb

erP

ositio

n

Surface R

ou

ghn

ess (nm

)

Side A

Average

Side A

Stand

ard D

eviation

Side A

Side B

Average

Side B

Stand

ard D

eviation

Side B

Appendix D: Toolox44 pin surface roughness and weight.

Page 59: High Temperature Tribology in Hot Stampingltu.diva-portal.org/smash/get/diva2:1345637/FULLTEXT01.pdfHigh Temperature Tribology in Hot Stamping Evan Kurnia Mechanical Engineering, master's

54

Appendix E: Matlab code for AE signal conversion into the frequency domain.

clear all;

close all;

samplef = 5000000;

nyqf = samplef/2;

signal = dlmread('AS_CH01-2.asc','',13,0);

%%

tc = 1.0284; %duration to cut early signal corresponding to the waiting for the sliding

fftp = 5000; %fft points

t = signal(:,1)-tc; %corrected time data, starts at the same time as sliding

ts = 0.8; %the time when the signal starts to be analyzed

tol = 6.67*10^(-8); %Tolerance to find the data with specific time value

danals = find(t > (ts - tol) & t < (ts + tol) ); %find the data corresponds to signal analysis

starting time

danale = danals - 1 + fftp; %find the data that corresponds to signal analysis ending time

metrix = [t,signal(:,2)]; %new data matrix with corrected time.

a = signal(danals:danale,2); %get the amplitude corresponds to analysis starting and ending time

f = linspace(0,nyqf,length(a)/2);%divided by 2 to match number of nodes to length of domain

H = hann(length(a));

signalfreq = fft(a.*H); %multiply hann window function to the signal amplitude

signalfreqabs = abs(signalfreq)/length(a);%divided by length a to compensate spectral energy

fftspec=signalfreqabs(1:length(a)/2);%keeping only first half to avoid mirroring

%%

figure(1)

subplot(2,1,1);

plot(metrix(:,1),metrix(:,2));

xlim(([0 2]));

ylim(([-0.3 0.3]));

xlabel('Time(s)');

ylabel('Voltage (V)');

set(gca,'FontSize',25);

subplot(2,1,2);

plot(f/1000000,fftspec);

xlim(([0.1 1]));

ylim(([0 4*10^(-4)]));

xlabel('Frequency (MHz)');

ylabel('Voltage (V)');

set(gca,'FontSize',25);

Sliding Combination Value of Variable “tc”

Parallel 1 1.0284

Parallel 2 1.2217

Perpendicular 1 1.1011

Perpendicular 2 1.1563

As-delivered Al-Si coated UHSS 1.4362

Heat-treated Al-Si coated UHSS 0.6710

The values in the table are to determine the start of the AE signal in the time domain that approximate

the starting time of the sliding. For evaluation of noise effect on the result in the frequency domain, tc =

0 and ts = 0.2 are used for AE signals from sliding combination Parallel 2 and Heat-treated Al-Si coated

UHSS.