VSEBINA – CONTENTS
PREGLEDNI ^LANKI – REVIEW ARTICLES
Physical simulation of metallurgical processes
Fizikalna simulacija metalur{kih procesov
S. T. Mandziej . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
IZVIRNI ZNANSTVENI ^LANKI – ORIGINAL SCIENTIFIC ARTICLES
Thermodynamic modeling for the alloy design of high speed steels and high chromium cast irons
Termodinami~no modeliranje na~rtovanja sestave hitroreznih jekel in litine z veliko vsebnostjo kroma
M. Pellizzari . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Use of Grey based Taguchi method in ball burnishing process for the optimization of surface roughness and microhardness of
AA 7075 aluminum alloy
Uporaba Grey-Taguchijeve metode pri procesu glajenja za optimizacijo povr{inske hrapavosti in mikrotrdote aluminijeve zlitine AA 7075
U. Esme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Effect of cerium additions on the AlSi17 casting alloy
Dodatek cerija livni zlitini AlSi17
S. Kores, M. Von~ina, B. Kosec, P. Mrvar, J. Medved . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
A generalized theory of plasticity
Posplo{ena teorija plasti~nosti
V. V. Chygyryns’kyy, K. A. Yakovlevich, I. Mamuzi}, B. A. Nikolaevna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Comparative modeling of wire electrical discharge machining (WEDM) process using back propagation (BPN) and general
regression neural networks (GRNN)
Primerjalno modeliranje elektroerozijske `i~ne obdelave (WEDM) z uporabo povratnosti (BPN) in spo{ne nevronske regresijske mre`e
(GRNN)
O. Guven, U. Esme, I. E. Kaya, Y. Kazancoglu, M. K. Kulekci, C. Boga . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Degradation of bacteria Escherichia coli by treatment with Ar ion beam and neutral oxygen atoms
Uni^evanje bakterij Escherichia coli s curkom ionov Ar in nevtralnih atomov kisika
K. Eler{i~, I. Junkar, A. [pes, N. Hauptman, M. Klanj{ek-Gunde, A. Vesel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
STROKOVNI ^LANKI – PROFESSIONAL ARTICLES
Effect of heat treatment on the microstructure and properties of Cr-V ledeburitic steel
Vpliv toplotne obdelave na mikrostrukturo in lastnosti ledeburitnega jekla Cr-V
S. Krum, J. Sobotová, P. Jur~i . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
18. MEDNARODNA KONFERENCA O MATERIALIH IN TEHNOLOGIJAH, 15. – 17. november, 2010, Portoro`, Slovenija
18th INTERNATIONAL CONFERENCE ON MATERIALS AND TECHNOLOGY, 15–17 November, 2010, Portoro`, Slovenia . . . . . . 163
ISSN 1580-2949
UDK 669+666+678+53 MTAEC9, 44(3)103–161(2010)
MATER.
TEHNOL.
LETNIK
VOLUME 44
[TEV.
NO. 3
STR.
P. 103–161
LJUBLJANA
SLOVENIJA
MAY–JUN.
2010

S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
PHYSICAL SIMULATION OF METALLURGICAL
PROCESSES
FIZIKALNA SIMULACIJA METALUR[KIH PROCESOV
Stan T. Mandziej
Advanced Materials Analysis, P. O. Box 3751, NL–7500DT Enschede, Nederland
stanamanl@cs.com
Prejem rokopisa – received: 2009-11-16; sprejem za objavo – accepted for publication: 2009-12-08
Worldwide demand of better and more efficient metallurgical processes, leading to low costs of their products, stimulates
intensive research to reach these goals. In this respect, any full-scale industrial experiments appear non-acceptable. Cutting off
the R&D costs and fast introducing of new technologies is possible when physical and numerical simulations are used. The
computer simulation can be only correct when exact data of materials behaviour at processing conditions are known. To obtain
the data, physical simulation is needed and it must be executed on multi-purpose thermal-mechanical testing devices accurately
reproducing the real industrial processing conditions. For continuous casting or metal forming, individual phases of processes or
multi-step operations must be followed, characterized by their time, temperature, and by applied forces, strains and strain rates.
Actually the physical simulation, as compared with full-scale industrial testing, allows in a fraction of time for a fraction of cost
an improvement of existing technology or development of a new one for modern materials and products. It can be used for
solving production problems due to solidification phenomena or deformability limits, which result in hot cracking and rejection
of the product. In this paper several examples of physically simulated procedures are given and their physical background
discussed.
Key words: physical simulation, Gleeble thermal simulator, deformation, hot cracking, welding, casting
Rast zahtev po bolj{ih in bolj u~inkovitih metalur{kih procesih stimulira po vsem svetu intenzivne raziskave, kako dose~i te
cilje, ker so za tak namen nesprejemljivi realni industrijski poizkusi. Zmanj{anje RR-stro{kov in hitro uvedbo novih tehnologij
dose`emo, ~e uporabimo fizikalno in numeri~no simulacijo procesov. Ra~unalni{ka simulacija je natan~na, ~e uporablja prave
podatke o vedenju materiala v procesnih razmerah. Da bi te podatke pridobili, je potrebna fizikalna simulacija, ki se izvr{i na
ve~namenski termomehani~ni preizkusni napravi, ki natan~no reproducira pogoje realnega industrijskega procesiranja. Za
neprekinjeno litje in oblikovanje kovin je treba upo{tevati razli~ne faze procesov, ki se odvijajo v ve~ stopnjah, ki jih
karakterizirajo ~as, temperatura in uporabljene sile, deformacije in hitrosti deformacije. Fizikalna simulacija, ~e jo primerjamo z
industrijskim preizkusom v polnem obsegu, omogo~i, da se v delcu ~asa za del stro{kov dose`e izbolj{anje sedanje tehnologije
ali razvoj nove za moderne materiale in proizvode. Uporabljamo jo lahko za re{itev proizvodnih problemov povezanih s
strjevalnimi pojavi ali z mejnimi deformacijami, ki povzro~ijo vro~o pokljivost in izme~ek proizvodov. V tem ~lanku
predstavljamo primere fizikalno simuliranih procedur in razpravljamo o njihovem fizikalnem ozadju.
Klju~ne besede: fizikalna simulacija, Gleeblejev termi~ni simulator, deformacija, vro~a pokljivost, varjenje, litje
1 INTRODUCTION – PHYSICAL SIMULATION
OF WELDING
Historically, welding was the first metallurgical
process successfully studied by the simulation technique,
in particular the use of simulated thermal cycles to
reproduce the situation occurring in parent plate material
affected by heat generated during welding. Further
development of this technique led to the introduction of
several testing procedures aiming to exactly replicate
various situations in the weld metal or between the
welded joint and adjacent material, in particular these
occurring due to the stress relaxation and plastic
accommodation processes during multiple heat cycling
characteristic of multi-pass welds. The necessity to apply
physical simulation for studying ductility and fracture
behaviour of heat-affected zones of welds, occurred just
after WW-II and was associated with low ductility of the
weld heat-affected zone in high-strength steels intro-
duced to shipbuilding 1. In 1949 the first weld thermal-
cycle simulator was built and utilized to generate
uniform heat-affected zone microstructures through the
CVN-sample for impact testing 2. At that time, using the
simulation, the results of impact strength appeared to be
often better than on real welds. This discrepancy
initiated more detailed studies on physical phenomena
standing behind the toughness of weldments, which
studies revealed the role of thermal gradients in affecting
phase transformations and generating crystallographic
lattice defects to plastically accommodate various
micro-strains. Accordingly, the weld thermal-cycle
simulator was equipped with a mechanical system
capable to deform specimens with adequate speed at
exact temperatures. When this was done, the Gleeble
thermal-mechanical simulator was born 3. Since 1957,
when the first commercial Gleeble was produced for the
purpose of weldability studies, it underwent several
metamorphoses and the recent dynamic thermal-mecha-
nical simulators allow the use of physical simulation
techniques not only for welding but also for other
industrial application processes, such as high strain rate
multi-step hot forming with all kinds of thermal-mecha-
nical processes, melting and controlled solidification for
the purpose of conventional and continuous casting as
well as semi-solid processing, checking for susceptibility
Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119 105
UDK 669.14.018.298:621.791:621.785 ISSN 1580-2949
Pregledni ~lanek/Review article MTAEC9, 44(3)105(2010)
to hot cracking and embrittling, studying sintering, stress
relaxation, accelerated creep, thermal-mechanical
fatigue, and others.
The main concept of the thermal-mechanical simu-
lator, however, has not been changed much since the
beginning, and the actual Gleeble system 4 comprises: an
AC electric resistance heating system as it was origi-
nally, a mechanical deformation system which earlier
was pneumatic and now is servo-hydraulic, a vacuum
and/or controlled atmosphere working chamber, and like
in all the most recent testing equipments the computer
control plus data acquisition and processing. For specific
applications it may comprise various working units and
attachments. And like at the beginning, the main aim of
the physical simulation is to exactly reproduce the ther-
mal and mechanical situation of the workpiece material
as it appears in real processing, and to obtain identical
microstructures as well as mechanical properties of the
physically simulated materials.
1.1 Weld heat-affected zone
In the historical case of welding and its vulnerable to
embrittlement and cracking heat-affected zone (HAZ),
during the electric arc action a short thermal cycle
appeared comprising a rapid heating, followed after
reaching peak temperature by a cooling due to heat flow
from the hot zone of the weld towards the cooler bulk of
the material. The heating and cooling rates result from
the balance between the energy input and the heat flow,
and are controlled by thermal gradients between the hot
zone of the weld and the parent plate. During the thermal
cycle the electric current flows through the HAZ and the
heat flows from the fusion surface between the weld and
the plate to form isothermal planes perpendicular to this
main direction of the heat flow, as schematically
presented in Figure 1. This situation has to be repro-
duced on a bar-like sample in which a narrow uniformly
heated zone is formed in the middle due to the thermal
balance between the electric heating and the heat flow
towards cold copper jaws, producing real temperature
gradients, Figure 2. During the weld thermal cycle, a
part of the HAZ material at first expands on heating and
during this is compressed being restrained crosswise by
the cold portions of the parent material, while in the
second portion of this cycle the faster cooling down
portions of thermal gradient zones once again compress
it. In this second part of the thermal cycle also tensile
strains appear, in particular in the main direction of the
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
106 Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119
Figure 3: Schematic drawing of Gleeble sample for laser dilatometer
studies of phase transformations in steels with cooling rates up to 200
°C/s (a) and this sample mounted in Gleeble’s "cold" Cu jaws (b)
Slika 3: Shemati~na oblika Gleeblejevega preizku{anca za laserki
dilatometrski {tudij faznih transformacij v jeklu s hitrostjo do 200 °C/s
(a) in preizku{anec pritrjen v Gleeblejevih hladnih bakrenih ~eljustih
(b)
Figure 1: Schematic situation of the heat-affected zone in an arc weld
with electric current flow and heat flow.
Slika 1: Shema polo`aja zone toplotnega vpliva pri elektrooblo~nem
varjenju s smerjo toka toplote
Figure 2: Weld HAZ simulation in Gleeble on 10mm diameter
round-bar sample mounted in "cold" copper jaws.
Slika 2: Simulacija v Gleeblejevi napravi za toplotno zono zvara 10
mm na okrogli palici, pritrjeni v hladnih bakrenih ~eljustih
a)
b)
heat flow, and these may assist embrittlement due to
generation of dislocations and interaction of these with
interstitials 4, or even initiate intergranular cracking.
The magnitude of crosswise deformation due to
thermal gradient, which in the case of a stiff real com-
ponent may result in substantial residual stresses, can be
illustrated in the following test. On the sample like in
Figure 3, tested for phase transformations by contact-
less laser dilatometer, at the cooling rate of 160 °C/s
obtained in the hot gauge zone by heat transfer to cooler
mounting portions of the sample fixed in cold copper
jaws of the Gleeble, a permanent shrinkage of diameter
occurs after the thermal cycle, Figure 3. At lower cool-
ing rates of 30 °C/s and less, which produce weaker tem-
perature gradients, such permanent change of sample’s
diameter may be negligible, Figure 4.
1.2 Stress-relief embrittlement
An important demand for the successful physical
simulation is its practical accuracy. Certain physical
phenomena cannot be detected or studied on samples of
wrong size or improper geometry, and then the sensi-
tivity of the testing system must be adequate. Moreover,
the physical background of the studied phenomenon
must be well known, to properly design the experiment.
Here an example is given from a study on stress-
relief embrittlement, occurring in some micro-alloyed
HSLA steels and weld metals after post-weld heat-treat-
ment. The stress relieving PWHT carried out at 600 °C
converts prior fully acicular ferrite microstructure of the
HSLA weld to partly recrystallised one, Figure 5, which
is brittle and fractures in an intercrystalline mode along
columnar crystals of the weld metal, Figure 6. This
stress-relief embrittlement was associated with initial
configuration of crisscrossing a/2<111> screw disloca-
tions, Figure 7, dominating in ferrite grains along the
boundaries of prior columnar grains, and interaction of
these dislocations with interstitials to form meta-stable
a<001> edge dislocations 5. The saturated by interstitials
a<001> edge dislocations are mobile and able to enter
the columnar grain boundaries to deposit there the inter-
stitials, thus sensitizing these boundaries. An example of
such boundary is given in Figure 8. Non-embrittling
HSLA weld metals of very similar chemical compo-
sitions had another initial dislocation substructure and
during the stress-relief annealing mainly generated large
amount of tangled a/2<111> edge type accommodation
dislocations.
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119 107
Figure 4: Laser dilatometer curves of HSLA steel heated up in
Gleeble to 1020 °C and cooled by heat flow to Cu jaws with rate of
160 °C/s (a) and with rate of 30 °C/s (b).
Slika 4: Laserke dilatometrske krivulje jekla HSLA, ki je bilo v
Gleeblejevi napravi segreto na 1020 °C in ohlajeno s pretokom toplote
v ~eljusti s hitrostjo 160 °C/s (a) in s 30 °C/s (b)
Figure 6: Brittle intergranular and transgranular fracture surface of
the stress-relief annealed HSLA weld metal
Slika 6: Krhka intergranularna in transgranularna prelomna povr{ina v
napetostno `arjenem zvaru HSLA-jekla
Figure 5: Microstructure of transformed acicular ferrite of HSLA
weld metal, showing grain boundary of prior columnar crystals
Slika 5: Mikrostruktura iz transformiranega acikularnega ferita v
zvaru jekla HSLA, ki prikazuje meje primarnih stebrastih kristalnih
zrn
Testing on Gleeble for susceptibility to stress-relief
embrittlement comprises 3-step thermal cycling applied
on bulk samples of 10 mm to 12 mm diameter, while
recording diameter by crosswise contactless laser dila-
tometer; the setup of this test is schematically shown in
Figure 9. The laser dilatometer readouts from austenitic
stainless steel sample show no changes of its dimensions
during such thermal cycling, Figure 10.
Samples of a non-embrittling HSLA weld metals
during the hold periods of thermal cycles show increase
of diameter, Figure 11, due to mainly generation of
accommodation dislocations, while the embrittling
HSLA weld metals in the first cycle show tempering of
acicular ferrite (bainite) and in further cycles a steady
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
108 Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119
Figure 12: Laser measurement during thermal cycling of embrittling
HSLA weld metal
Slika 12: Laserske meritve med termi~nim cikliranjem zvara
HSLA-jekla z pojavom krhkosti
Figure 8: Sensitized grain boundary containing high density of dislo-
cations and fine precipitates in embrittled HSLA weld metal
Slika 8: Sensitivirana krstalna meja z veliko gostoto dislokacij in
majhnih izlo~kov v krhkem zvaru HSLA-jekla
Figure 7: Configuration of criss-crossing screw dislocations domi-
nating in ferrite of the stress-relief embrittling weld metal
Slika 7: Konfiguracija kri`ajo~ih se vija~nih dislokacij, ki prevladu-
jejo v feritu v zvaru, ob~utljivem za popustno krhkost
Figure 9: Schematic of laser dilatometer measurement on sample
mounted in Gleeble
Slika 9: Shema meritve z laserskim dilatometrom za vzorec v Gleeb-
lejevi napravi
Figure 10: Result of laser measurement during 3-step thermal cycling
of austenitic stainless steel
Slika 10: Rezultati meritev z laserskim dilatometrom pri termi~nem
cikliranju avstenitnega nerjavnega jekla v treh korakih
Figure 11: Laser measurement during thermal cycling of non-em-
brittling HSLA weld metal
Slika 11: Laserske meritve med termi~nim cikliranjem zvara HSLA-
jekla brez pojava krhkosti
decrease of diameter, Figure12, associated with reaction
of a/2<111> screw dislocations into the meta-stable
a<001> edge dislocations (decrease of density by the
factor of 
2.4) followed by annihilation of the last,
mostly within the grain boundaries.
1.3 Hot ductility and hot cracking
One of the main problems of weldability is the
susceptibility of some welds to hot cracking, as welding
is the technology of joining materials by bringing them
locally to the molten state and then solidifying this
region. At high temperatures just below solidus most of
metal alloys suffer loss of strength and ductility. This
occurs on heating due to liquation before reaching the
fully molten state and then on cooling from the melt it
persists within certain range of temperatures after
solidification, called brittle temperature range – BTR. As
the solidified alloy becomes stiff while not yet ductile,
hot cracks may form on cooling, when tensile strains
caused by shrinkage and assisted by restraint cannot be
compensated by ductility of this alloy. Accordingly, the
method of studying susceptibility of an alloy to hot
cracking involves tensile testing at the conditions
simulating these of the real welding (or casting) process.
The use of thermal-mechanical simulator like Gleeble,
able to reproduce on a specimen the welding thermal
cycle and impose a strain in a controlled manner, allows
achieving this goal.
Hot tensile test
Testing on Gleeble for susceptibility to weld
liquation cracking / HAZ hot cracking, is the hot tensile
testing of a number of specimens on-heating and then
on-cooling, and determining their hot ductility measured
as a reduction in area at the specimen’s neck portion
after the test. This procedure is schematically presented
in Figure 13 6.
The hot ductility of a welded alloy increases
gradually with increase of testing temperature from
ambient towards melting point, however, before reaching
this point it drops abruptly from certain maximum to nil.
Just above this nil ductility temperature (NDT) is the nil
strength temperature (NST) at which the alloy looses its
strength due to liquation i.e. formation of weak or liquid
phases along grain boundaries. The real physically
measurable melting temperature of such alloy – TL, is
higher than NST. On-cooling from the melt or from the
NST, the ductility does not recover exactly at NDT only
below it at so-called ductility recovery temperature –
DRT. The temperature span from NST to DRT is
considered to be the brittle temperature range – BTR, the
extent of which can be used as a rough criterion of the
susceptibility to hot cracking.
More exact criterion of this susceptibility is the
measure how fast does the ductility recover on-cooling
as compared with its decrease on-heating. As reference
point for this measurement the maximum of ductility
from the on-heating ductility curve is taken and the
representative areas below the on-heating and on-cooling
curves are compared, Figure 14 7. Arbitrarily consider-
ing 5 % of reduction-in-area on-cooling as the ductility
recovery point and comparing the hot ductility curves
on-heating and on-cooling, the nil ductility range (NDR),
ductility recovery rate (DRR) and ratio of ductility reco-
very (RDR) can be determined to exactly characterize
the susceptibility of an alloy to hot cracking.
The reference point for the hot ductility measure-
ments determining the susceptibility to liquation
cracking is the nil strength temperature. This temperature
has to be used as a peak of the welding thermal cycle,
on-cooling after which the hot tensile tests should be run.
To measure the NST on Gleeble an attachment is used,
schematically presented in Figure 15. This NST attach-
ment keeps the specimen under a constant tensile load of
about 50 N while allowing its heating-up with an initial
heating rate the same as of the welding thermal cycle to
be simulated up to a temperature 50 °C below the solidus
temperature of the material, then change to a heating rate
of 2–5 °C/s until the NST is reached, as shown in Figure
16.
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119 109
Figure 14: Evaluation of hot ductility curves for the hot cracking
susceptibility 7
Slika 14: Ocena krivulj vro~e duktilnosti za ob~utljivost za vro~o
pokljivost 7
Figure13: Schematic of Gleeble’s procedure for hot cracking,
including hot tensile testing on-heating up to NST and then on-cooling
after weld thermal cycle with NST as the peak temperature 6
Slika 13: Shema Gleeblejeve procedure za vro~o pokljivost, ki obsega
vro~ raztr`ni preizkus s segrevanjem do NST in ohlajanje po ter-
mi~nem ciklu varjenja z NST kot najvi{jo temperaturo 6
To avoid mixing-up of phenomena related to
liquation cracking with these of solidification cracking,
in order to measure only the hot ductility region
responsible for the liquation cracking, the NST must not
be exceeded in the simulated weld thermal cycles.
To test for the susceptibility to solidification
cracking, controlled melting and solidification of a
rod-like specimen can be carried out on Gleeble, and
after the solidification the hot ductility is determined in
the manner like described above. Here the ability of
Gleeble to melt electrically conductive specimens is
used. In this test the central portion of the specimen
protected by a crucible / quartz sleeve is brought to a
temperature above solidus and then this crucible contains
the molten / semi-liquid metal, as shown in Figure 17.
Thermal gradients between the molten portion and
mounting jaws prevent the metal from flowing out of the
crucible while controlled thermal cycle allows con-
ducting the solidification in a manner similar to that of
real casting or welding. Schematic of the assembly used
for this purpose in Gleeble is given in Figure 18.
More details of the Gleeble testing for liquation and
solidification cracking can be found in IIW document
No. II-C-042A/95 8. The Gleeble testing procedures are
also mentioned in the technical report CEN ISO/TR
17641-3:2003, under the chapter "Hot Tensile Test" 9.
The SICO test
An alternative procedure to study susceptibility to
solidification cracking on Gleeble is the strain-induced
crack opening test – SICO, developed at Dynamic Sys-
tems Inc., in USA, for studying hot deformability of
alloys. In this test the central hot portion of the Gleeble
specimen is compressed to form a bulge, Figure 19, on
outer perimeter of which cracks appear at the critical
secondary tensile strain, Figure 20. The critical strain to
fracture in SICO test is defined as the hoop strain at
onset of cracking in the bulge zone:

c = ln (Df /Do)
where Do is the initial diameter of the specimen, while
Df is the final maximum diameter in the bulge zone.
As during the controlled melting and solidification in
Gleeble the dendritic crystals mostly grow from the
direction of the main heat flow i.e. in the axial direction,
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
110 Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119
Figure 18: Schematic of Figure 17, with end nuts for hot tensile test-
ing
Slika 18: Shema slike 17 s kon~nimi vijaki za vro~ natezni preizkus
Figure 16: Time – temperature graph of Gleeble test for nil strength
temperature
Slika 16: Graf ~as – temperatura Gleeblejevega preizkusa za tempera-
turo ni~elne trdnosti
Figure 15: Nil strength attachment of Gleeble using trapezoidal
spring grip maintaining constant small tensile load on specimen
during heating
Slika 15: Shema monta`e brez sile v Gleeblejevo napravo s prijemom
z uporabo trapezoidalne vzmeti, ki ohranja konstantno majhno
natezno napetost med ogrevanjem preizku{anca
Figure 17: Gleeble’s setup for controlled melting and solidification
study
Slika 17: Gleeble postavitev za raziskave kontroliranega taljenja in
strjevanja
the mid-span segregation may occur in the specimen
causing deep and partly hidden cracking of the SICO
specimen along the central plane perpendicular to the
compression axis, Figure 21. In such situation it is
advised to check for the true critical diameter on the
cross-section of SICO sample, like it is shown in Figure
22.
As the hot tensile testing on Gleeble gives the ade-
quate characteristics of an alloy regarding its hot
cracking susceptibility, the SICO test appears to be more
accurate for measuring of critical strains to fracture and
related critical strain rates.
2 SIMULATION OF CASTING
The physical simulation actually plays an important
role in the design and application of the most efficient
industrial manufacturing process, which is the conti-
nuous casting and the hot rolling following it. As shown
in Figure 23, in different stages of process may occur
flaws such as shrinkage cavity, centre-line porosity,
facial and corner cracks 10. Physical simulation can be
used to determine flawless processing parameters, with-
out interrupting of production.
2.1 Controlled solidification and semi-solid processing
In the first phase of process like schematically shown
above, the liquid metal is poured into a crystallizer
chamber in which its outer shell has to solidify to the
extent securing the liquid core inside. When a vertically
cast slab or billet has to be bent into horizontal position,
the ductility of outer shell must allow this. Gaining
physical data for such operation includes melting and
solidifying of numerous samples in the manner like
presented in Figures 17 & 18, and hot tensile testing
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119 111
Figure 23: Schematic of continuous casting and hot rolling processes 10
Slika 23: Shema neprekinjenega litja in vro~ega valjanja 10
Figure 21: Central plane crack in SICO often appearing after melting
and solidification
Slika 21: Razpoka v centralni ravnini na SICO-preizku{ancu, ki po-
gosto nastane po taljenju in strjevanju
Figure 20: Schematic of SICO test showing crack formed on bulge
portion of sample
Slika 20: Shema SICO preizkusa z napako na izbo~enem delu
preizku{anca
Figure 19: SICO sample tested in Gleeble
Slika 19: SICO-preizku{anec med preizkusom na Gleeblejevi napravi
Figure 22: Method of measuring critical diameter Dcr when central
plane crack appears
Slika 22: Metoda za merjenje kriti~nega premera Dcr v primeru raz-
poke v centralni ravnini
them at various temperatures after solidification, to
create hot ductility maps 10.
An alternative test comprises melting of a 25 mm
diameter bar and deforming it by compression after a
partial solidification. The heating-cooling balance of
Gleeble causes in the sample thermal gradient to occur
keeping molten core inside when outer shell solidifies.
The test called 25 mm SICO gives critical strain to
fracture, strain rate and temperature, and in the case of
incomplete solidification reveals thickness of outer shell
on crashed sample, Figure 24.
Distribution of strains in the SICO sample can be
determined by numerical simulation 11. Having known
the strain distribution and temperature gradient as well,
valuable information can be gained from cross section of
the sample like given in Figure 25: cracks at maximum
perimeter occurred at temp of 1385 °C and 0.35 strain
with 0.5/s strain rate, then no-crack zone at 
0.5 of
radius appeared representative to temperature 
1400 °C
and strain of 0.2, and once again cracks in the middle of
the sample were formed representative to strain 
0.05 at
temperature of 
1425 °C, the last is the nil ductility
temperature of the tested steel.
2.2 Direct casting and rolling
The recent development in continuous casting
consists of thin strip casting followed directly by hot
rolling while retaining the heat of the melt. For simu-
lation of such process the dedicated HDS-V40 physical
simulator has been built 4,12, allowing on one specimen
after melting and solidification with controlled dendrite
size and growth direction to perform multi-step high rate
deformations representative to multi-stand rolling. The
process comprises resistance heating of a bulk 10 mm
thick, 50 mm wide and 165 mm long specimen in a
specifically designed crucible, to melt the central portion
of this specimen and then to solidify it in a controlled
manner, and then to deform its central portion by heated
plane strain anvils perpendicularly to the lengthwise
axis. Upon melting and subsequent cooling a thin shell
forms around the molten material and as soon as it
appears the melting crucible is pulled away from the
specimen to allow for deformation. Schematic of this
process is given in Figure 26. Multiple hit deformations
by plane strain compression, Figure 27, can then be per-
formed on the material while it is in the solid or semi-
solid state. Resulting from the simulation is the flat bar,
Figure 28, with central deformed zone (5), transient
zone (4) and side zones of reheated in solid state coarse-
grained material (1), mushy zone (2) and fully melted (3)
material; in this picture marked by hatching are sections
used in subsequent metallographic examination.
The reliability of the melting and solidification
procedure can be discussed in terms of non-metallic
inclusions before and after the experiment. The specimen
made of 0.2 % C plain carbon steel, had in the initial
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
112 Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119
Figure 26: Schematic of the melting and solidification test in
HDS-V40 simulator, showing the site and shape of the molten pool
and then formed columnar crystals as visible on longitudinal section
and on fracture in the mid-plane
Slika 26: Shema talilnega in strjevalnega preizkusa v HDS-V40 simu-
latorju, ki ka`e mesto in obliko staljene kopeli in nato nastalih ste-
brastih zrn, kot se vidijo na vzdol`nem prerezu in na prelomu v srednji
ravnini
Figure 25: Cross-section of SICO sample showing deep cracks along
perimeter and cracks in the middle, with flawless zone in between
Slika 25: Pre~ni prerez SICO-preizku{anca z globokimi razpokami
vzdol` oboda in razpokami v sredini z podro~jem brez napak med
obema
Figure 24: Sizes of solidified outer shell and molten pool in 25 mm
SICO bar melted at 1465 °C and then compressed at 1420 °C surface
temperature, while the middle temperature being of 
1450 °C
Slika 24: Velikosti strjene zunanje lupine in talilne kopeli v 25 mm
SICO-palici, ki je bila staljena pri 1465 °C in kr~ena pri temperaturi
povr{ine 1420 °C, ko je bila temperatura v sredini 
1450 °C
state a ferritic-pearlitic microstructure with bands elon-
gated in rolling direction and so extended non-metallic
inclusions. After the test, in the entirely melted and soli-
dified "as-cast" zone-3, chains of spheroidal inclusions
dominated along boundaries of dendrites, Figure 29. On
fracture through mid-plane of the as-solidified specimen,
which coincides with the mid-plane of the former molten
pool, individual dendritic crystals are seen, Figure 30,
with their lengths of up to 2.0 mm. This size of dendrites
well coincides with the length and width of dendritic
crystals appearing in the thin continuously cast slabs
manufactured in the industrial processes 13.
3 SIMULATION OF HOT DEFORMATION
Two major hot deformation procedures: forging and
rolling, have to be physically simulated in a different
manner, due to differences in main force direction, main
strain direction and heat flow, as schematically given in
Figure 31. Thus to simulate forging an axial com-
pression / flow stress test is used, while for simulation of
rolling the plane strain compression has to be applied.
3.1 Multi-step deformation
An exact simulation of multi-step hot rolling by plane
strain compression test requires constant strain rates to
be maintained in each step with an instantaneous stop at
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119 113
Figure 31: Schematic presentation of a roll stand and simulation of
deformation by: (a) flow stress test, and (b) plane strain test. The
letters denote: F – compression force; H – heat flow direction;  –
strain direction: X – rolling direction 14
Slika 31: Shema valjalni{kega ogrodja in simulacija s: (a) preizkusom
meje te~enja in (b) z ravninskim deformacijskim preizkusom. ^rke
pomenijo: F – tla~na sila, H – smer toka toplote, - smer deformacije,
X – smer valjanja 14
Figure 30: Cleavage fracture along mid-plane of specimen, showing
sizes of dendritic crystals
Slika 30: Cepilni prelom vzdol` centralne ravnine preizku{anca, na
katerem vidimo velikost dendritnih zrn
Figure 28: Specimen after simulation of melting and direct hot rolling
Slika 28: Preizku{anec po simulaciji taljenja in direktnega vro~ega
valjanja
Figure 27: Hot central zone of specimen and plane strain anvils in the
working chamber
Slika 27: Vro~a centralna zona preizku{anca in ~eljusti za napenjanje
v delovni komori
Figure 29: Fine spheroidal inclusions in grain boundary ferrite of
dendritic crystals
Slika 29: Fini sferi~ni vklju~ki v feritu po kristalnih mejah dendritnih
zrn
the end of deformation. To achieve such performance of
the Gleeble’s servo-hydraulic system, a dedicated defor-
mation-assisting device called Hydrawedge was de-
signed and implemented 14. The Hydrawedge, synchro-
nized with the main / primary deforming system, is
acting as a flexible mechanical stop that allows the
primary hydraulic ram to be stopped by running into an
immovable object. In order to perform exact multiple
compressions sequentially, for which the specimen must
be moved since the main hydraulic ram will stop at the
same point in space each time, the Hydrawedge is used
to program the displacements. This allows to exactly
control the amount of strain, while simultaneously and
separately controlling the strain rate at which the sample
is being deformed. Without such device, all fast servo-
hydraulic machines or give substantial over-travel or
must slow down before stopping at the right sample’s
height. In the first instance other than programmed
strains are generated while in the last case final micro-
structures are generated characteristic of slower than
programmed strain rates.
In Figure 32 the schematic of Hydrawedge is given
and in Figure 33 its operation explained. In the system
consisting of main power piston (M), punch with stop
(P), yoke (Y), sample (S) and Hydrawedge piston (H),
before deformation the system is in position like drawn
in Figure33(a). Preparation to deformation includes
pulling off the main piston (M) to form a gap "g"
between it and punch (P). Simultaneously the Hydra-
wedge piston (H) pushes the sample (S) and punch (P) to
set-up the amount of required deformation between the
punch’s stop ring and the yoke (Y), Figure 33(b). Finally
in the deformation step, Figure 33(c), the main piston
accelerates through the gap "g" reaching the required top
speed, then during the deformation decelerates in a
programmed manner to maintain constant strain rate, and
when the stop of the punch (P) hits the yoke (Y), the
deformation is finished with high accuracy. Such steps
can be then repeated several times.
Maintaining constant strain rates and instantaneous
termination of each deformation step with an aid of
Hydrawedge reveals true behaviour of the material under
processing. Theoretical deformation behaviour assumes
continuous strain hardening and increase of flow stress
with decrease of deformation temperature, Figure 34. In
reality, due to microstructure transformation processes
like dynamic and static recrystallisations and precipita-
tions overlapping with these, the true multi-step flow
stresses are often much different, Figure 35.
An example, which follows, shows how the physical
simulation can be used to generate various as-hot-rolled
microstructures. A HSLA steel, used for manufacturing
of hot rolled plates in a 7-step rolling process, as in
Figure 36, with continuous drop of temperature, had an
acicular ferrite / upper bainite microstructure with very
uniform grain size across the plate thickness, given in
Figure 37. On this steel an attempt was made by
physical simulation on Gleeble, to refine grains, separate
the phases, achieve a dual-phase microstructure, as well
as to reduce amount of deformation steps in the rolling
mill.
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
114 Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119
Figure 34: Theoretical flow stress curves of multi-step hot deforma-
tion
Slika 34: Teoreti~ne krivulje te~enja pri vro~i deformaciji v ve~ stop-
njah
Figure 32: Schematic drawing of the Hydrawedge device 14
Slika 32: Shema vle~enja na napravi Hydrawedge 14
Figure 33: Hydrawedge operation
Slika 33: Preizkus Hydrawedege
One of the methods to obtain the dual-phase micro-
structure is an intercritical annealing followed by accele-
rated cooling. An alternative is to stimulate by defor-
mation the phase separation in the two-phase intercritical
austenite-ferrite range. The new requested continuous
5-step rolling procedure needed a proof that for the
selected steel this would be achieved. The appropriate
test on Gleeble comprised applying small strains after
which the relaxation of sample was monitored. Two- and
three-hit experiments were carried out with different
time intervals after the hits. An example of a 2-hit
experiment is displayed in Figure 38. A sample
pre-loaded to 175 MPa compression stress at room
temperature was heated with constant rate of 10 °C/s to
above A3 temperature in fixed Gleeble jaws, showing
relaxation temperature of about 480 °C. After reaching
920 °C it was deformed by the first hit and then held to
relax for 100 s while the temperature was constantly
dropping. The second hit was executed at 770 °C and
during the hold for 150 s after it the austenite-to-ferrite
isothermal transformation effects were observed.
Further a free cooling was applied, during which at
first the effects were observed of austenite-to-martensite
transformation and then "stiffening" of the ferritic-mar-
tensitic microstructure.
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119 115
Figure 39: Strain vs. time and temperature profile of 5-step simulation
to achieve fine-grained microstructure
Slika 39: Odvisnost napetosti od ~asa in temperaturni profil pri
simulaciji v 5 korakih za dosego finozrnate mikrostrukture
Figure 36: Strain vs. time of the 7-step rolling schedule simulated in
Gleeble
Slika 36: Odvisnost deformacije od ~asa pri simulaciji valjanja v 7
vtikih v Gleeblejevi napravi
Figure 35: True stress-strain graph of the 7-step hot rolling simulated
in Gleeble
Slika 35: Pravi graf napetost – deformacija simulacije valjanja v 7
vtikih v Gleeblejevi napravi
Figure 37: Microstructure of acicular ferrite / bainite with uniform
distribution of carbides
Slika 37: Mikrostruktura iz acikularnega ferita/bainita z enakomerno
porazdelitvijo karbidov
Figure 38: The 2-step deformation and relaxation test in the
intercritical range followed by cooling, as simulated in Gleeble
Slika 38: Simulacija v napravi Gleeble za deformacijo v dveh korakih;
preizkus relaksacije v interkriti~nem podro~ju ter hlajenje
The final 5-step rolling procedure, Figure 39, com-
prised three initial steps in austenite phase, maintaining
strains, strain rates, as well as interpass times and tem-
peratures adequate to preserve and/or to additionally
stimulate precipitation of fine carbides. The last two
steps were executed in the austenite-ferrite two-phase
region according to the results from tests like this
presented in Figure 38. The resulting fine-grained
microstructure, which on plane strain samples reduced to
only 
1.0 total strain formed mainly in shear-bands, is
shown in Figure 40. By an increase of strains in all five
steps of this new rolling procedure the uniform fine-
grained microstructure was obtained across the whole
thickness of the plate.
3.2 Solving problems in hot rolling
Flawless rolling process requires that characteristic
of the material strains to fracture got never exceeded. In
hot deformation tests steeper temperature gradients cause
apparent strengthening of the material and decrease of its
ductility. In industrial manufacturing this phenomenon is
often responsible for appearance of corner cracking of
hot rolled billets. The lower temperatures of the corner
region always appear, Figure 41a, and corner or near-
corner cracks may occur when the flow of material to the
corner of billet is hampered by the thermal gradients.
This situation, schematically given in Figure 41b, shows
how normal force and friction forces exerted by rolls
cause the material flow in the rolling direction as well as
towards the corners, while the plasticity of the material
along the cooler corners may be limited. Solving of this
problem requires hot strength and ductility data that
might be provided by numerous hot tensile tests 14.
As an alternative to physically reproduce this
situation, a two-step SICO test may be used (Figure 42).
After forming a bulge in the first compression defor-
mation, an air (or inert gas) blow is locally applied to
produce the measured by two thermocouples Tc1 & Tc2
thermal gradient like in the real situation, and when this
gradient is achieved the second compression is applied,
with adequate second strain and proper strain rate, till the
cracks appear. This procedure gives the amount of strain
to fracture at the real strain rate, at controlled tempe-
rature and at real temperature gradient – it is not easy to
gain all these data simultaneously in any other simple
and continuous test.
4 THERMAL-MECHANICAL FATIGUE
The thermal-mechanical fatigue is a complex process
often responsible for premature failure of components in
power generation and chemical processing. In recent
four decades microstructural features were identified,
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
116 Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119
Figure 42: A two-step SICO test for susceptibility of hot rolled billets
to corner cracking, including: the first compression to form a bulge,
holding till the gradient appears and the final second compression
Slika 42: Dostopenjski SICO- preizkus ob~utljivosti vro~e valjanih
slabov za vogalne razpoke, ki obsega: najprej kr~enje za nastanek
izbokline, zadr`anje do nastanka gradienta in nato drugo, kon~no
kr~enje
Figure 41: The real situation of hot rolled steel slab (a) with cooler
corner portions, and its schematic explanation indicating the directions
of main metal flow
Slika 41: Pravo stanje vro~e valjanega slaba s (a) hladnej{imi vogali
in shematsko ozna~bo smeri glavnega toka metala
Figure 40: Fine-grained microstructure if the deformation bands after
the 5-step rolling simulation in Gleeble
Slika 40: Finozrnata mikrostruktura v deformiranem pasu po simu-
laciji valjanja v 5 vtikih v Gleeblejevi napravi
which in novel creep resisting martensitic / ferritic steels
accelerate precipitation of carbides and speed-up reco-
very and recrystallisation of matrix. Based on these
evidences a low-cycle thermal-mechanical fatigue proce-
dure called accelerated creep test (ACT) was developed
on Gleeble. The ACT speeds-up microstructural changes
by elasto-plastic tensile and compressive strains applied
on the sample during thermal cycling in the temperatures
characteristic of creep. It also uses the advantages of
direct electric resistance heating, which is the heating
mode implemented on Gleeble and also complies with
the recent knowledge on development of dislocation
substructures and their effect on precipitation processes
in ferritic and multi-phase steels 15.
4.1 The accelerated creep test – ACT
The up to date ACT procedure complies with the
following principles:
• The basic temperature and applied strains prevent
odd transformations like e.g. secondary dissolution of
carbides or intensive formation of non-equilibrium
phases.
• The final deformation at fracture is like at real creep
– just a few pct in total.
• The depletion of weld metal or steel matrix in
alloying elements is achieved similar to that of crept
steels and the carbide phases at onset of cracks are
not different.
Size and mounting of specimens in Gleeble allows a
uniformly heated zone formed in the middle-span of the
sample and the portion of material in this zone undergoes
transformation. To better define this zone, gauge portion
of a reduced diameter is made on the sample, Figure43.
The ACT was primarily invented for welded joints, in
particular for repair welding. Its first larger application
appeared in the 5thFP EU R&D project "SmartWeld"
(2001–2004) and more recently in COST-536 and
COST-538 EU actions.
The ACT samples mounted in the Gleeble’s "pocket
jaws" assembly, like in Figure 44, were subjected to
programmed cycles of the low-cycle thermal-mechanical
fatigue, run till failure or till pre-determined stress or
strain. In homogeneous materials, often before the crack
appearance on the surface, internal cracks were formed
extending perpendicularly to the sample’s axis. In micro-
structurally inhomogeneous samples the cracks usually
followed the "weakest links" like the HAZs of the welds.
4.2 Results of the ACT
In the implemented low-cycle thermal-mechanical
fatigue procedure on Gleeble data of stress, strain, strain
rate and temperature as well as dilatometric information
are recorded, out of which strain-time and stress-time
graphs can be produced like these given below in Figure
45. Usually the tests are run till failure of specimen,
however the test can be interrupted anytime and
specimens for e.g. fine fractographic and microanalytical
investigations taken before fracturing. As the tests for
different materials are run at different temperatures and
the response of material gives various stress equivalent to
the YS at elevated temperature of the test, to compare the
ACT results the duration of the test and its temperature
can be included in the following parameter:
PACT = (7 + log t) · T/100
where: t/ks = time of test, and T/K = temperature.
Then, the creep strength factor in ACT can be calcu-
lated as:
FACT = PACT · RS /100
where RS is the average stress of all ACT cycles measu-
red during relaxation on tension.
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119 117
Figure 44: ACT sample located in the Gleeble "pocket jaw" assembly,
then shape of the sample after the test and the internal crack formed
during the test in the neck portion of this sample
Slika 44: ACT-preizku{anec v `epnih ~eljustih Gleeblejeve naprave;
oblika preizku{anca po preizkusu in notranja razpoka, nastala v vratu
preizku{anca
Figure 43: Schematic drawing of cross-weld samples used for ACT
on all-weld-metal and on weld’s HAZ, and mounting of the sample in
Gleeble’s jaws set
Slika 43: Shema kri`nih varilnih preizku{ancev za ACT na metalu
zvara in coni toplotnega vpliva zvara ter pritrditev preizku{anca v
~eljusti Gleeblejeve naprave
From the graph of Figure 45a for the zero-stress
condition the progress of permanent extension can be
read, while from graph of Figure 45b the decrease of
tensile yield strength with progress of the creep can be
observed as well as the change of elasticity modulus
calculated. Micro-hardness measurements after the ACT
show results comparable with these of multi-year creep
exposed power generation components of the same grade
material; some examples are given in Table 4.1.
4.3 Verification of results
The loss of steel’s strength during creep is associated
with transformation of its microstructure, which in the
case of the Cr-Mo-V grade materials is due to precipi-
tation of carbides and their coagulation and also to
formation of intermetallic phases, so metallographic and
microanalytical investigations are needed to confirm the
reliability of the testing procedure.
In the case of the P91 steels and weld metals dis-
cussed here, their initial microstructures in the initial
tempered state contain numerous fine and medium size
carbide precipitates, relatively uniformly distributed in
the ferritic matrix retaining its post-martensitic character,
Figure 46. Such matrix, when seen in thin foil speci-
mens in TEM, mostly consists of fine subgrains aligned
in arrays of crystallographic orientations inherited from
the former martensite microstructure, Figure 47. During
exposure to creep conditions, in this initial microstruc-
ture further precipitation processes occur as well as
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
118 Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119
Figure 45: Strain-time (a) and stress-time (b) graphs, from ACT at
600 °C on a post-weld heat-treated P91 weld metal
Slika 45: ACT-grafa deformacija – ~as (a) in napetost – ~as (b) pri 600
°C za po varjenju toplotno obdelan P 91-zvar
Table 4.1: Hardness results of P91/P92 materials after creep and after
ACT
Tabela 4.1: Trdote jekel P91/P92 po preizkusu lezenja in po ACT
Material / sample / exposure
Micro-hardness HV 100G
initial after exposureor testing
P91 component 1 – pipe / 3
years at 568 °C n. a. 231
P91 component 1 – weld / 3
years at 568 °C n. a. 258
P91 component 1 – bottle / 3
years at 568 °C n. a. 243
P91 component 2 – antler / 9
years at 600 °C n. a. 229
P91 component 2 – weld / 9
years at 600 °C n. a. 208
P91 component 2 – stub / 9
years at 600 °C n. a. 177
ACT – P91 standard weld
metal with PWHT 285 223
ACT – P91 lean/soft weld
metal low PWHT 298 193
ACT – P92 standard weld
metal 294 240
ACT – P92 standard pipe 268 228
Figure 47: Arrays of subgrains with carbides in tempered martensite
of P91 steel; thin foil specimen, TEM image
Slika 47: Podro~je podzrn s karbidi v popu{~enem martenzitu v jeklu
P 91, tanka folija; TEM-posnetek
Figure 46: Tempered martensite microstructure of P91 steel; FeCl3
etched, SEM image
Slika 46: Mikrostruktura iz popu{~enega martenzita v jeklu P 91. Jed-
kano s FeCl3; SEM-posnetek
transformation of the existing carbides and their coagu-
lation, in combination with recovery and recrystallisation
of the matrix.
Accordingly, any reliable creep test should not
generate microstructures much different from these taken
from true services. The microstructures generated during
the ACT appear to be a bit finer than after exposure to
true creep while the precipitation of carbides especially
at the surface of the internal crack seems to be somewhat
more intensive, Figure 48. Nevertheless, on the fracture
surfaces of the ACT samples the precipitated phases are
identical with these after the long-term creep exposure
and their content and chemical compositions do agree
with the Thermocalc prediction of phases, which should
be present at thermodynamic equilibrium 16. On these
fracture surfaces also traces of slip lines can be observed
confirming that up to the onset of cracking the process is
mainly controlled by dislocations glide and annihilation.
Finally the microstructure after the ACT consists of well
recrystallised grains with uniform distribution of sphe-
roidal precipitates, mainly carbides, Figure 49.
5 CONCLUSIONS
1. Up-to-date physical simulation procedures allow
studying materials behaviour at conditions very close
to real industrial processing or applications.
2. The process parameters such like temperature, sense
and amount of strain, strain rate as well as thermal
gradients can be adequately reproduced and their
values accurately recorded.
3. By means of physical simulation a large variety of
microstructures and associated mechanical properties
can be obtained and studied in a short time and for a
tiny fraction of full-scale industrial experiments.
4. The materials behaviour data gained from physical
simulation experiments can be further used in
computer modeling and control of the industrial
manufacturing processes.
6 REFERENCES
1 E. F. Nippes & W. F. Savage; The Weldability of Ship Steel – A
Study of the Effect of Travel Speed, Preheat Temperature and Arc
Power Level on the Notch Toughness of the Weld Metal and the
Heat-Affected Zone, Welding Journal 25 (1946), Res Suppl, p. 776s
2 E. F. Nippes & W. F. Savage; Tests of Specimens Simulating Weld
Heat-Affected Zones, Welding Journal 28 (1949), Res Suppl, p. 599s
3 W. F. Savage et al; An Investigation of the Hot Ductility of High
Temperature Alloys, Welding Journal 34 (1955), Res Suppl, p.183s
4 Information on http://www.gleeble.com/
5 S. T. Mandziej; Physical Simulation of Welding, in Proc ASM Intl
Conf on Welding & Joining, Madrid, Spain, March 1997, p.253
6 W. Lin, J. C. Lippold, W. A. Baeslack III; An Evaluation of
Heat-Affected Zone Liquation Cracking Susceptibility, Part I:
Development of a Method for Quantification’, Welding Journal 72
(1993), Res Suppl, p.135
7 C. D. Lundin et al; Hot ductility and hot cracking behavior of
Modified 316 Stainless Steels designed for high temperature service,
Welding Journal 72 (1993), Res Suppl, p.189
8 A Standard Procedure for Hot Cracking Test, Dynamic Systems Inc,
Poestenkill, NY, USA, 1995, IIW Doc. II-C-042A/95
9 CEN ISO/TR 17641-3:2003, Destructive tests on welds in metallic
materials – Hot cracking tests for weldments – Arc welding
processes – Part 3: Externally loaded tests
10 H. G. Suzuki, S. Nishimura, S. Yamaguchi; Physical Simulation of
the Continuous Casting of Steels, in Proc Physical Simulation of
Welding, Hot Forming and Continuous Casting, MTL 92-43(TR),
CANMET, Canada, 1988, p. II-1
11 L. Trebacz, R. Kuziak, M. Pietrzyk; Numerical Model of the SICO
Test, Proc STEELSIM 2005, Brno CR, 2005, p.487
12 S. T. Mandziej, J. D. Vosburgh, R. Kawalla, H.-G. Schoss; Physical
Simulation of Thin Slab Continuous Casting, Materials Science
Forum, Vols 539-547, 2007, p.4149
13 B. Engel, M. Albedyhl, M. Brühl, Ch. Klinkenberg, H. Langner, H.
Pircher, K. Wünnenberg; Stahl und Eisen 118 (1998) 5, 41
14 H. S. Ferguson; Fundamentals of Physical Simulation, in Proc Intl
Symposium on Physical Simulation, TU Delft 1992, p. 1
15 S. T. Mandziej; Low-Energy Dislocations and Ductility of Ferritic
Steels, in Fundamental Aspects of Dislocation Interactions, Materials
Science & Engineering A, 164 (1993) 275
16 S. T. Mandziej, A. Výrostková: Evolution of Cr-Mo-V weld metal
microstructure during creep testing – Part 1: P91 material, Welding
in the World, 52 (2008)1/2, 3
Materiali in tehnologije / Materials and technology 44 (2010) 3, 105–119 119
S. T. MANDZIEJ: PHYSICAL SIMULATION OF METALLURGICAL PROCESSES
Figure 49: Recrystallised fine ferrite grains retaining oriented arrays
of carbides in P91 sample after ACT at 600 °C; TEM, thin foil
Slika 49: Rekristalizirana fina feritna zrna, ki so obdr`ala orientirano
porazdelitev karbidov v preizku{ancu iz jekla P91 po ACT pri 600 °C,
TEM, tanka folija
Figure 48: Fracture surface of ACT sample of P91 steel showing high
density of carbides, FeCl3 etched, SEM image
Slika 48: Prelomna povr{ina ACT-preizku{anca iz jekla P 91, ki ka`e
veliko gostoto karbidov; jedkano s FeCl3; SEM-posnetek

M. PELLIZZARI: THERMODYNAMIC MODELING FOR THE ALLOY DESIGN OF HIGH SPEED STEELS ...
THERMODYNAMIC MODELING FOR THE ALLOY
DESIGN OF HIGH SPEED STEELS AND HIGH
CHROMIUM CAST IRONS
TERMODINAMI^NO MODELIRANJE NA^RTOVANJA SESTAVE
HITROREZNIH JEKEL IN LITINE Z VELIKO VSEBNOSTJO
KROMA
Massimo Pellizzari
Department of Materials Engineering and Industrial Technologies, via Mesiano 77, 38050 Trento – Italy
Massimo.Pellizzari@ing.unitn.it
Prejem rokopisa – received: 2009-11-16; sprejem za objavo – accepted for publication: 2003-03-10
In recent years, Thermo-Calc was successfully used by the author for the development of as-cast High Speed Steels and High
Chromium Irons. The correlation between solidification process and microstructure was studied in view of influence of alloying
elements introduced to promote carbide precipitation. The solidification process occurs under non-equilibrium conditions,
because of microsegregation phenomena connected with the solidification structure: the liquid between dendrites becomes
progressively enriched in solute and its composition increasingly differs from that predicted by the equilibrium diagram. This
behavior can be studied with good approximation by using the model proposed by Scheil and Gulliver. In this way it was
possible to refine the phase constitution of High Speed Steels, even if kinetic-related phenomena still limit its correct prediction.
Present results also show that microstructural tailoring is possible looking at the correlation existing between the fraction of
liquid phase at eutectic MC carbides precipitation. The morphology of V-rich particles changes from a continuous interdendritic
network to a globular dissociated eutectic, showing higher toughness. The composition of HiCrI should result as near as
possible to the eutectic, to maximize eutectic carbide amount. Calculations allowed to define the parameter TL-TEi (TL = liquidus
temperature; TEi = eutectic start temperature) as representative of the material hypoeutecticity, and, on the basis of experimental
results, a value of 20 °C was safely established in the development of the new composition.
Key words: thermodynamic modelling, solidification, high chromium alloy, high speed steel, phase composition, MC carbides
Zadnja leta je avtor uspe{no uporabil Thermo-Calc za razvoj hitroreznih jekel (HSS) in veliko vsebnostjo litine kroma (HiCrI).
Raziskana je bila korelacija med procesom strjevanja in mikrostrukturo z vidika legirnih elementov, ki so bili dodani za
pove~anje precipitacije karbidov. Proces strjevanja je neravnote`en zaradi povezanosti med strjevalno strukturo in
mikrosegregacijo: talina med dendriti postopno postane obogatena s topljencem, njena sestava pa se nara{~ajo~e razlikuje od
tiste, ki jo napoveduje ravnote`ni diagram. Tako vedenje je mogo~e raziskati z zadostnim pribli`kom, ~e uporabimo model, ki
sta ga predlagala Scheil in Gulliver. Tako lahko izbolj{amo fazno sestavo hitroreznih jekel, ~eprav kineti~ni pojavi omejujejo
natan~no napoved. Prikazani rezultati dokazujejo, da je mogo~e krojiti mikrostrukture z opazovanjem korelacije med dele`em
taline pri precipitacijo MC-karbidov. Morfologija zrn, bogatih z V, se spremeni od povezane interdritske mre`e v globularni
disociirani evtektik in dobi ve~jo `ilavost. Sestava litine z veliko vsebnostjo kroma (HiCrI) naj bo ~im bli`ja evtekti~ni, da se
tako pove~a dele` evtekti~nih karbidov. Z izra~uni je bil dolo~en parameter TL – TEi (TL = likvidis temperatura; TEi = temperatura
za~etka evtektika), ki je merilo hipoevtekti~nosti. Na podlagi eksperimtalnih rezultatov je bila opredeljena vrednost parametra
20 °C za razvoj novih sestav.
Klju~ne besede: termodinami~no modeliranje, strjevanje, sestava faz, hitrorezno jeklo, litina z veliko vsebnostjo kroma,
MC-karbidi
1 INTRODUCTION
The solidification and precipitation sequence play an
important role on the microstructure and the final proper-
ties of high speed steels (HSS) and high chromium cast
irons (HiCrI). The optimization of properties is related to
the possibility to govern the phase precipitation,
microsegreation and transformation temperature during
solidification. This becomes crucial for spincast hot roll
materials, that are not submitted to thermomechanical
treatment after solidification. Hence, hardness, wear
resistance and toughness (…) are largely dependent on
the amount and type of primary and eutectic carbides,
but also on the composition of the metallic matrix giving
rise to secondary hardening during final treatment 1–3.
Different models have been applied to simulate the
solidification behaviour of multicomponent systems like
high speed steels 4–7 and cast irons 7,8 using thermo-
dynamic methods. In general, the solidification does not
follow the equilibrium and limitations regarding the
diffusivity in the solid and, partially in the liquid phase,
have to be taken into account 9,10. The Scheil-Gulliver
model 11,12, considering the total lack of diffusion within
the solid and the complete mixing in the liquid
(produced by efficient stirring and diffusion), defines the
upper limit for the solute segregation in the liquid and
the minimum attainable solidus temperature. In the
present work Thermo-Calc 13 software was used to model
the solidification process under both, thermodynamic
equilibrium and under the conditions stated by the Scheil
model.
Materiali in tehnologije / Materials and technology 44 (2010) 3, 121–127 121
UDK 669.14.018.252:669.13:669.26 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 44(3)121(2010)
2 THERMODYNAMIC MODELS AND
DATABASE
The phases in the studied multicomponent systems
are described by the sub-lattice model 14,15. The molar
Gibbs free energy of a phase m in the system can be
expressed as
G G TS G Gm m m m m= − + +
ref id ex mg
where G m
ref is the reference free energy, −TS m
id is the
ideal mixing entropy, G m
ex is the excess energy term and
G m
mg is the change in energy caused by magnetic order-
ing. The thermodynamic data used for the calculations
are contained in the TCFE2000 database. Phases
considered are reported in Table 1.
3 ALLOY DESIGN OF HIGH CHROMIUM IRON
The nominal chemical composition of standard HiCrI
is reported in Table 2. The base-alloy contains mainly
carbon and chromium, with only minor addition of
secondary hardening alloying elements (Mo, V and W).
The hypo-eutectic nature of this alloy can be observed in
the microstructure in Figure 1, displaying the material
after quenching from 950 °C and double tempering. The
result is also confirmed by the calculated isopleth in
M. PELLIZZARI: THERMODYNAMIC MODELING FOR THE ALLOY DESIGN OF HIGH SPEED STEELS ...
122 Materiali in tehnologije / Materials and technology 44 (2010) 3, 121–127
Table 1: Model of the phases considered in the calculations
Tabela 1: Modeli faz, upo{tevanih pri izra~unih
Phase N. sub. N of sitesPer sublatticee
Sublattice Species
(Va = vacancies)
Liquid 1 1 Fe ,Si, Mn, Cr, Ni, Mo, V, Nb: C, Va
FCC (austenite) 2 1 : 1 Fe, Si, Mn, Cr, Ni, Mo, V, Nb: C, Va
BCC (ferrite) 2 1 : 3 Fe, Si, Mn, Cr, Ni, Mo, V, Nb: C, Va
MC 2 1 : 1 Fe, Si, Mn, Cr, Ni, Mo, V, Nb: C, Va
M2C 2 2 : 1 Fe, Si, Mn, Cr, Ni, Mo, V, Nb: C, Va
M3C (cementite) 2 3 : 1 Fe, Si, Mn, Cr, Ni, Mo, V, Nb: C
M6C 4 2 : 2 : 2 : 1 Fe : Mo, W: Fe, Si, Mn, Cr, Ni, Mo, V, Nb: C
M7C3 2 7 : 3 Fe, Si, Mn, Cr, Ni, Mo, V, Nb: C
M23C6 3 20 : 3 : 6 Fe, Cr: Fe, Cr, Si, Mn, Cr, Ni, Mo, V, Nb: C
Table 2: Chemical composition of standard HiCrI
Tabela 2: Kemi~na sestav standardne HiCrI
C Si Mn Cr Ni Mo V W
2.4–2.7 0.3–0.8 0.4–0.8 16–20 0.3–2.0 0.2–2.5 0.2–0.4 0.2–0.6
Figure 2: a) Isopleth of standard HiCrI and b) mole fraction of phases
vs temperature microstructure
Slika 2: a) Isopleth standardne HiCrI; b) molski dele` faz v odvisnosti
od tempereture
Figure 1: Microstructure of standard HiCrI
Slika 1: a) Mikrostruktura standardne HiCrI
Figure 2a. Microstructure after heat treatment comprises
tempered martensite dendrites and a eutectic consisting
of an interpenetrating network of M7C3 carbides and
tempered martensite.
The volume percentage of eutectic carbides assessed
with Image Analysis is 28.0 ± 4 %. This agrees well with
that predicted by Thermo-Calc, using TCFE2000 data-
base: the carbide volume percentage, as calculated by the
lever rule at 950 °C, is of 32 % (Figure 2b). The discre-
pancy can be explained in view of the lower kinetics of
carbides precipitation in the solid state than in the liquid
phase. This is confirmed by Figure 2b showing that the
fraction of M7C3 at solidus is much lower (25 %) than at
950 °C.
The solidification of the melt starts with the preci-
pitation of primary austenite at 1287 °C (TL = liquidus),
while, the eutectic reaction starts at 1260 °C (TEi) and
ends at 1228 °C (TEf = TS = solidus). Looking to a novel
cast iron with higher wear resistance, the composition of
the modified HCrI should results as near as possible to
the eutectic to maximize eutectic carbide amount. How-
ever, in designing the new composition, containing
higher percentages of alloying elements than in the stan-
dard grade, attention was paid on avoiding hypereutectic
solidification which may occur in internal sections of the
shell 16, due to segregation. Large pro-eutectic carbides
are known to worsen wear resistance and toughness, as
well.
Therefore, the new composition was tailored to be
slightly hypoeutectic. Since alloying elements influence
eutectic carbon, Thermo-Calc was used to simulate
solidification of alloys with different compositions, using
standard products to validate the theoretical predictions.
The content in Cr and Mo was modified. The width of
the solidification range of the base alloy by varying
content in Cr and Mo is reported in Figures 3a and 3b,
respectively. While the increasing Cr causes a general
reduction of solidification range, Mo plays the opposite
effect 17. Table 3 resumes the influence of an increasing
content of both alloying on the solidification of standard
HiCr. The parameter TL – TEi (TL = liquidus temperature;
TEi = eutectic start temperature) was taken as represen-
tative of the material hypoeutecticity and, on the basis of
the experimental results, a value of 20 °C was safely
established in the development of the new composition.
Data in Figure 3 show that the parameter TL – TEi is
more sensitive to an increase in Cr than in Mo. Thus, in a
material with 20 % Cr proeutectic carbides may easily
M. PELLIZZARI: THERMODYNAMIC MODELING FOR THE ALLOY DESIGN OF HIGH SPEED STEELS ...
Materiali in tehnologije / Materials and technology 44 (2010) 3, 121–127 123
Figure 3: Influence of a) Cr and b) Mo, on the solidification range of
standard HiCrI
Slika 3: Vpliv a) Cr in b) Mo na strjevalni interval standardne HiCrI
Figure 4: a) Mole fraction of solid phases vs. temperature b) pro-
eutectic (Nb,V)C carbide within the metallic matrix
Slika 4: a) Molski dele` faz v odvisnosti od temperature, b) proevtek-
ti~ni (Nb,V)C-karbid v kovinski matici
form at solidification, while no proeutectic carbides
could be find up to high Mo content.
Table 3: Influence of Mo and Cr on the solidification range of
standard HiCrI
Tabela 3: Vpliv Mo in Cr na strjevalni interval standardne HiCrI
Standard
HiCrI
Standard
HiCrI 3% Mo
Standard
HiCrI 20% Cr
TL 1295 °C  
TEi 1260 °C  
TEf 1230 °C 
 
TE = TEi – TEf 30 °C  

TL – TEi 35 °C  

An improved high Mo HCrI grade was thus designed
on the base of the above results to obtain a value of TL –
TEi greater than 20, after proper modification of the C
content of the standard composition. A further develop-
ment was the introduction of a stronger carbide former
than Mo, like V and Nb. Depending on the content of
these two elements, pro-eutectic precipitation of MC
carbides is possible (Figure 4a) directly from the liquid
at much higher temperature than eutectic M7C3 carbides.
Under this conditions, MC carbides are located in the
interior of dendrites (Figure 4b), causing a substantial
strengthening of the metallic matrix, without any
embrittlement produced by interdendritic precipitation.
4 ALLOY DESIGN OF HIGH SPEED STEEL
The nominal composition of the base high speed steel
in reported in Table 4.
Table 4: Nominal composition of the base HSS
Tabela 4: Nominalna sestava osnovnega HSS
C Si Mn Cr Ni Mo V
1.65-1.70 0.8-0.9 0.85-0.90 5.400 0.500 2.900 4.600
The related pseudo binary phase diagram calculated
with Thermo-Calc is displayed in Figure 5a and the
mole fraction of phases provided by equilibrium are
shown in Figure 5b.
The solidification starts with the formation of
primary dendrites of austenite at 1348 °C. The fraction
increases up to 0.42 at 1273 °C, as the eutectic reaction,
i.e., L  Aus+MC carbides starts. The amount of eutectic
phases then increases by decreasing the temperature
down to the calculated solidus temperature (1243 °C). At
this point, the provided molar fraction of MC carbides is
8.8 %. In equilibrium, no other eutectic reactions are
predicted.
Thermal analysis highlights that solidification starts
at 1355 °C, in good agreement with the calculated value
(Figure 6). The first eutectic reaction, i. e., L  Aus +
MC, starts at about 1320 °C and shows the fastest kinetic
at about 1280 °C in correspondence of the DTA peak. A
second eutectic reaction, i. e., L  Aus + M7C3, starts at
about 1180 °C and shows the maximum rate at 1160 °C.
Solidification ends at about 1090 °C, i. e., at a much
lower temperature than that provided by equilibrium
M. PELLIZZARI: THERMODYNAMIC MODELING FOR THE ALLOY DESIGN OF HIGH SPEED STEELS ...
124 Materiali in tehnologije / Materials and technology 44 (2010) 3, 121–127
Figure 6: DTA showing the reactions occurring on solidification of
HSS. In the frame the microstructure showing the presence of MC and
M7C3 eutectic carbides
Slika 6: DTA, ki prikazuje reakcije med strjevanjem HSS. V okvirju
je mikrostruktura, kjer so prikazani evtekti~ni karbidi MC in M7C3
Figure 5: a) Isopleth of the studied HSS (dotted line represent the
composition of the studied HSS). b) Diagram reporting the mole
fraction of phases vs temperature.
Slika 5: a) Izoplete raziskanega HSS (~rtkane ~rte pomenijo raziskani
HSS); b) diagram, ki ponazarja molski dele` faz v odvisnosti od
temperature
calculations (1243 °C). Discrepancies between calcu-
lations and DTA can be explained in view of the micro-
segregation to which liquid is exposed during solidifi-
cation, caused mainly by the limited diffusion of solute
atoms in solid phase. The result of simulation using the
Scheil-Gulliver model in Thermo-Calc better reflects the
experimental one. In Figure 7, the comparison between
calculated molar fractions and experimental data indicate
a quite good agreement. Now, beneath the eutectic
reaction allowing the precipitation of MC carbides, a
second reaction involving the precipitation of M7C3 is
provided, as well. The fractions of carbides assessed
with image analysis and calculations are very close, such
us the solidus temperatures. As expected, in view of the
maximum possible segregation of the liquid provided by
the Scheil model, the calculated solidus underestimates
the experimental value.
Hence, it can be stated that solidification proceeds
under non equilibrium conditions, is caused principally
by the limited diffusion in the solid phase.
This allows the author to use the Scheil-Gulliver
model in the alloy design of such steel grade. Interden-
dritic eutectic carbides reduce toughness of HSS, since
they constitute an almost continuous brittle network in
M. PELLIZZARI: THERMODYNAMIC MODELING FOR THE ALLOY DESIGN OF HIGH SPEED STEELS ...
Materiali in tehnologije / Materials and technology 44 (2010) 3, 121–127 125
Figure 9: Increasing fraction of dissociated eutectic carbides by
increasing V/Cr content (from up to down)
Slika 9: Nara{~ajo~i dele` disociacije evtekti~nih karbidov pri rasti
razmerja vsebnosti V/Cr (z zgoraj na dol)
Figure 7: Molar fraction of solid phases provided by Thermo-Calc
simulation using the Scheil Gulliver model
Slika 7: Molski dele` faz, dolo~en na podlagi Thermo-Calc simulacije
po Scheil Gulliver modelu
Figure 8: Morphology of MC carbides as a function of the content in
V and Nb
Slika 8: Morfologija karbidov MC kot funkcija vsebnosti V in Nb
Table 5: Computer simulation of the solidification sequence in HSS
(Thermo-Calc), using the Scheil Gulliver model
Tabela 5: Ra~unalni{ka simulacija sekvenc strjevanja HSS (Thermo-
Calc) po modelu Scheil Gulliver
Temperatures, T/°C Volume Pct of EutecticCarbides, /%
TL TMC TM7 TS MC M7C3 M2C TOT
1348 1273 1175 1087 8.0 2.0 trace 10.0
Table 6: Computer simulation of the solidification sequence in HSS
(Thermo-Calc), using the Scheil Gulliver model
Tabela 6: Ra~unalni{ka simulacija sekvenc strjevanja HSS (Thermo-
Calc) po modelu Scheil-Gulliver
DTA Temperatures, T/°C Volume Pct of EutecticCarbides, /%
TL TMC TM7 TS MC M7C3 M2C TOT
1355 1320 1180 1110 7.0 2.8 trace 10.0 (a) (b) (c)
the microstructure. Moreover, crack usually nucleates
either at the carbide-matrix interface or inside the
carbide particle, because the different elastic properties
of the two constituents rise local stresses. Depending on
the chemical composition, MC carbides tend to
precipitate with different morphology.
By high C and V concentration VC homogeneously
precipitates in eutectic cells to form a dissociated
eutectic (Figure 8a). A relative high amount of fine and
globular particles precipitate together with eutectic
austenite, so that they are expected to be less detrimental
for toughness than interdendritic carbides. The criteria to
obtain this kind of microstructure is that the eutectic
reaction is started with a high fraction of residual liquid
phase. This allows eutectic cells to develop fully. Figure
8a shows the fraction of solid phases in the base HSS
added with 6 % V. Two effects have to be underlined.
The MC start temperature is increased with respect to the
base material (1281 °C vs 1273 °C) and the corres-
ponding fraction of liquid phase is higher (0.69 vs 0.58),
as well. Indeed, Figure 8b shows the opposite result for
the steel with V content decreased to 1.5 %. The MC
start temperature drops to 1210 °C and the corresponding
fraction of liquid phase to 0.38. Experimental results
(Figure 9) confirm that the increasing content in V
facilitates the formation of a microstructure with a high
fraction of dissociated eutectic.
It is interesting to point out that Thermo-Calc allows
the easy calculation of the MC start temperature (Figure
10a) and of the fraction of liquid phase (Figure 10b) by
varying the content in V. Continuous lines in Figure 10b
highlight the increasing fraction of liquid phase at MC
precipitation start as the content in V increases. Corres-
pondingly, the fraction of solid, i.e. austenite up to 7.2 %
V and ferrite above 6.5 %, decreases.
With the aid of data in Figure 10 a critical value for
the fraction of austenite, for the formation of dissociated
MC carbides could be determined. Increasing the V
content above this value, the amount of eutectic cells
progressively increases. A limiting factor is represented
by the segregation of VC carbides during centrifugal
casting of bimetallic rolls. Because of the lower density
of these particles the liquid phase, MC carbides tend to
segregate at the shell-core interface. This phenomenon is
responsible for poor bonding and must be controlled to
preserve roll quality. Hence, as proposed previously, Nb
can be added, allowing the formation of composite Nb-V
MC carbides with higher density. Nb is an MC former
stronger than V. The addition of a small amount of this
element dramatically increases the precipitation tempe-
rature of MC carbides (Figure 10). In Figure 10a and
10b dashed lines are referred to the standard alloy with
0.5 % Nb. As MC precipitate in the 1 % V – 0.5 % Nb
alloyed high speed steel the fraction of liquid is about
0.8, i.e. much higher than that observed in the base
material (0.2). As the content in V is close to the mass
fraction 10 %, the fraction of liquid at MC precipitation
start becomes almost 1, indicating the ultimate condition
before the occurrence of proeutectic precipitation of MC
carbides. This circumstance is verified by the higher
amount of Nb, as confirmed by Figure 8c, showing the
fraction of solid phases in the base HSS with addition of
1.5 % Nb. NbC particles now precipitate directly from
the liquid at a much higher temperature than Nb-free or
low-Nb grades promoting the heterogeneous nucleation
of austenite dendrites. From this point of view, a refine-
ment of the solidification structure can be obtained with
properly selecting/handling Niobium. A basic form of
inoculation can be realized.
Finally, in order to drive solidification transfor-
mations towards dissociate eutectic and to minimize the
amount of interdendritic carbides, the chemical com-
position must be designed with a prevailing amount of V
and a limited amount of Cr, Mo and W. As an example,
if the amount in Cr is increased keeping V constant, the
temperature of Cr-rich M7C3 eutectic carbides is
increased, as their total fraction with respect to VC. The
possibility of eutectic cells development is hindered by
the concurrent precipitation of solid Aus + M7C3 (Figure
9). Moreover, carbon content has to be designed to both
sustain the precipitation of carbides and harden tempered
martensite.
M. PELLIZZARI: THERMODYNAMIC MODELING FOR THE ALLOY DESIGN OF HIGH SPEED STEELS ...
126 Materiali in tehnologije / Materials and technology 44 (2010) 3, 121–127
Figure 10: a) Eutectic MC start temperature and b) Mole fraction of
phases as a function of V content at eutectic MC start
Slika 10: a) za~etna temperatura MC-evtektika in b) molski dele` faz
v odvinsosti od vsebnosti V pri za~etku MC-evtektika
5 CONCLUSIONS
The present paper briefly resumes the authors
knowledge in the field of alloy design of high Speed
Steels and High Chromium Irons by means of
thermodynamic modelling. The basic criteria for the
development of new alloys were illustrated in view of the
eutectic reactions occurring during solidification. The
influence of type and amount of different alloying
elements on the solidification sequence was analysed and
correlated to the microstructure. The results confirm the
possibility to successfully employ computer modelling in
the alloy design of high alloyed steels and cast irons.
6 REFERENCES
1 Molinari A., Tremea A., Pellizzari M., Biggi A., Corbo G.: Mater.
Sci. and Technol. 18 (2002), 1574–1580
2 K. C. Hwang, S. Lee, H. C. Lee: Mater. Sci. Eng. A 254 (1998), 282
3 S. Karagöz, H. F. Fischmeister: Met. Trans. A 29 (1998), 205–216
4 J. A. Golczewsky, H. F. Fischmeister: Z. Metallkd. 84 (1993) 12,
860–866
5 J. A. Golczewsky, H. F. Fischmeister: Steel Research. 63 (1992) 8,
354–360
6 P. Gustafson: Computer assisted development of High Speed Steels.
In The SGTE Casebook, Thermodynamic at work. Chapter 6. 70–76
7 G. Coelho, J. A. Golczewsky, H. F. Fischmeister: Met. Trans. A 34
(2003), 1749–1758
8 C. G. Schon, A. Sinatora, CALPHAD 22 (1998), 437
9 H. F. Fischmeister, R. Riedl, S. Karagöz: Met. Trans. A 20 (1989),
2133
10 M. Boccalini, H. Goldenstein: Intl. Mater. Reviews, 46 (2001) 2,
92–114
11 G. M. Gulliver: J. Inst. Metals 9 (1913), 120–157
12 E. Scheil: Z. Metallkd. 34 (1942), 70–72
13 B. Sundman et al., CALPHAD 9 (1985), 153
14 H. Ohtani, T. Tnaka, M. Hasebe, T. Nishizawa: Calphad 12 (1988),
225
15 M. Hillert, LI Staffansson: Acta Chem. Scand. 24 (1970), 3618
16 Molinari A., Pellizzari M., Biggi A., Corbo G., Tremea A.:
Development of spincast hot rolls through microstructural tailoring.
Proc. 44th Mechanical Working and Steel Processing, Orlando
(Florida), 8-11 september, 2002, Iron & Steel Society:Vol. XL, pp.
1233–1244
17 J. A. Tenorio et.al., Int. J. Cast Metals Res. 13 (2000), 99
M. PELLIZZARI: THERMODYNAMIC MODELING FOR THE ALLOY DESIGN OF HIGH SPEED STEELS ...
Materiali in tehnologije / Materials and technology 44 (2010) 3, 121–127 127

U. ESME: USE OF GREY BASED TAGUCHI METHOD IN BALL BURNISHING PROCESS ...
USE OF GREY BASED TAGUCHI METHOD IN BALL
BURNISHING PROCESS FOR THE OPTIMIZATION OF
SURFACE ROUGHNESS AND MICROHARDNESS OF AA
7075 ALUMINUM ALLOY
UPORABA GREY-TAGUCHIJEVE METODE PRI PROCESU
GLAJENJA ZA OPTIMIZACIJO POVR[INSKE HRAPAVOSTI IN
MIKROTRDOTE ALUMINIJEVE ZLITINE AA 7075
Ugur Esme
Mersin University Tarsus Technical Education Faculty, Department of Mechanical Education, 33140, Tarsus-Mersin/Turkey
uguresme@gmail.com
Prejem rokopisa – received: 2009-11-09; sprejem za objavo – accepted for publication: 2010-03-20
This study investigated the multi-response optimization of burnishing process for an optimal parametric combination to yield
favorable surface roughness and microhardness using the Grey relational analysis and Taguchi method. Sixteen experimental
runs based on an orthogonal array of Taguchi method were performed to derive objective functions to be optimized within
experimental domain. The objective functions have been selected in relation of burnishing parameters; burnishing force, number
of passes, feed rate and burnishing speed. The Taguchi approach followed by Grey relational analysis was applyed to solve the
multi-response optimization problem. The significance of the factors on overall quality characteristics of the burnishing process
has also been evaluated quantitatively with the variance method (ANOVA). Optimal results were verified through confirmation
experiments. This shows application feasibility of the Grey relation analysis in combination with Taguchi technique for
continuous improvement in product quality in manufacturing industry.
Keywords: ball burnishing process, Grey relation analysis, Taguchi method
V tej {tudiji je raziskana ve~odgovorna optimizacija procesa glajenja z dosego optimalnih kombinacij parametrov za ugodno
povr{insko hrapavost in mikrotrdoto z uporabo Greyjeve analize odvisnosti in Taguchijeve metode. [estnajst eksperimentov v
ortogonalni porazdelitvi po metodi Taguchi je bilo uporabljenih za razvoj objektivnih funkcij za optimizacijo v
eksperimentalnem polju. Objektivne funkcije so bile izbrane v odvisnosti od parametrov glajenja; sila glajenja, {tevilo prehodov,
hitrost podajanja in hitrost glajenja. Taguchijev pribli`ek in Greyjeva analiza odvisnosti sta bila uporabljena za re{itev
ve~odgovornega problema. Kvantitativno je bil ocenjen tudi pomen dejavnikov kakovosti procesa glajenja z metodo variance
(ANOVA). Optimalni rezultati so bili potrjeni s preizkusi. Delo dokazuje uporabnost Greyjeve analize odvisnosti in Taguchijeve
tehnike za stalno izbolj{anje kakovosti proizvodov v predelovalni industriji.
Klju~ne besede: krogelno glajenje, Greyjeva analiza odvisnosti, Taguchijeva metoda
1 INTRODUCTION
The function performance of a machined component
such as fatigue strength, load bearing capacity, friction,
etc. depends to a large extent on the surface as
topography, hardness, nature of stress and strain induced
on the surface region. Nowadays, about 50% of the
energy supplied is lost in the friction of elements in
relative motion1,2. Roughness values less than 0.1 mm
are required for good aesthetic appearance, easy mould
release, good corrosion resistance, and high fatigue
strength. During recent years, however, considerable
attention has been paid to the post-machining metal
finishing operations such as burnishing which improves
the surface characteristics by plastic deformation of the
surface layers2,3.
Burnishing is considered as a cold-working finishing
process, differing from other cold-working, surface
treatment processes such as shot peening and sand
blasting, etc. in that it produces a good surface finish and
also induces residual compressive stresses at the metallic
surface layers4. Accordingly, burnishing distinguishes
itself from chip-forming finishing processes such as
grinding, honing, lapping and super-finishing which
induce residual tensile stresses at the machined surface
layers5,6. Also, burnishing is economically desirable,
because it is a simple and cheap process, requiring less
time and skill to obtain a high-quality surface finish4,5.
Beside producing a good surface finish, the burnish-
ing process has additional advantages over other machi-
ning processes, such as securing increased hardness,
corrosion resistance and fatigue life as a result of
producing compressive residual stress. Residual stresses
are probably the most important aspect in assessing
integrity because of their direct influence on perfor-
mance in service. Thus, control of the burnishing process
(burnishing conditions) in such a way as to produce
compressive residual stresses in the surface region could
lead to considerable improvement in component life. A
comprehensive classification of burnishing tools and
their application has been given by Shneider7. A lite-
rature survey shows that work on the burnishing process
Materiali in tehnologije / Materials and technology 44 (2010) 3, 129–135 129
UDK 669.715:620.178 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 44(3)129(2010)
has been conducted by many researchers and the process
improves also the properties of the parts, e.g. higher
wear resistance2,8,9 increased hardness10-12, surface
quality2,3,14 and increased maximum residual stress in
compression11. The parameters affecting the surface
finish are: burnishing force, feed rate, ball material,
number of passes, workpiece material, and lubrication2,3.
It is necessary to find an optimal process condition
capable of producing desired surface quality and hard-
ness. However, this optimization should be performed in
such a way that all the objectives should fulfill
simultaneously. Such an optimization technique is called
multi-response optimization15.
The majority of the research existing in literature on
the effect of burnishing parameters on the burnished
surface is of experimental nature and very few analytical
models are available in the literature.
The Taguchi method is very popular for solving
optimization problems in the field of production
engineering.16,17 The method utilizes a well-balanced
experimental design (allows a limited number of
experimental runs) called orthogonal array design, and
signal-to-noise ratio (S/N ratio), which serve as objective
function to be optimized (maximized) within the experi-
mental domain. However, traditional Taguchi method
cannot solve multi-objective optimization problem. To
overcome this, the Taguchi method coupled with Grey
relational analysis has a wide area of application in
manufacturing processes. This approach can solve
multi-response optimization problem simultaneously15,18.
Planning the experiments through the Taguchi ortho-
gonal array has been used quite successfully in process
optimization19–24. Therefore, in this study the Taguchi
L16(44) orthogonal array was applied to plan the
experiments on burnishing process.
Four controlling factors including burnishing force
(F), number of passes (N), feed rate (f) and burnishing
speed (V) on the surface roughness (Ra) and micro-
hardness (HV) with four levels for each factor were
selected. The Grey relational analysis was then applied
to examine how the burnishing parameters influenced the
surface roughness and microhardness and an optimal
parameter combination was then obtained. Through
analyzing the Grey relational grade matrix, the most
influential factors for individual quality targets of
burnishing process can be identified. Additionally, the
analysis of variance (ANOVA) was also utilized to
examine the most significant factors for the surface
roughness and microhardness in burnishing process.
2 GREY RELATIONAL ANALYSIS
In Grey relational analysis, experimental data i.e.,
measured features of quality characteristics are first
normalized ranging from zero to one. This process is
known as Grey relational generation. Next, based on
normalized experimental data, Grey relational coefficient
is calculated to represent the correlation between the
desired and actual experimental data15. Then overall
Grey relational grade is determined by averaging the
Grey relational coefficient corresponding to selected
responses. The overall performance characteristic of the
multiple response process depends on the calculated
Grey relational grade. This approach converts a multiple
response process optimization problem into a single
response optimization situation with the objective
function which is the overall Grey relational grade. The
optimal parametric combination is then evaluated which
would result in the highest Grey relational grade. The
optimal factor setting for maximizing overall Grey
relational grade can be obtained by Taguchi method15.
In Grey relational generation, the normalized Ra
values corresponding to the smaller-the-better (SB)
criterion which can be expressed as:
x k
y k y k
y k y ki
i i
i i
( )
max ( ) ( )
max ( ) min ( )
=
−
−
(1)
HV100 should follow the larger-the-better (LB)
criterion, which can be expressed as:
x k
y k y k
y k y ki
i i
i i
( )
( ) min ( )
max ( ) min ( )
=
−
−
(2)
where xi(k) is the value after the Grey relational
generation, min yi(k) is the smallest value of yi(k) for the
kth response, and max yi(k) is the largest value of yi(k)
for the kth response15. An ideal sequence is [x0(k) (k=1,
2, 3......, 16)] for the responses. The definition of Grey
relational grade in the course of Grey relational analysis
is to reveal the degree of relation between the 16
sequences [x0(k) and xi(k), i=1, 2, 3,.......,16]. The Grey
relational coefficient i(k) can be calculated as:


i i
k
k
( )
( )
min max
max
=
−
+
∆ ∆
∆ ∆0
(3)
where ∆ 0 0i ix k x k= −( ) ( ) the absolute value of the dif-
ference of x0(k) and xi(k);  is the distinguishing coeffi-
cient 0 	  	 1; ∆ min
min min ( ) ( )= ∀ ∈ ∀ −j i k x k x kj0 =
the smallest value of 0i; and
∆ max
max max ( ) ( )= ∀ ∈ ∀ −j i k x k x kj0 is the largest value
of 0i. After averaging the Grey relational coefficients,
the Grey relational grade 
i can be computed as:

 i i
k
n
n
k=
=
∑1
1
( ) (4)
where n is the number of process responses. The higher
value of Grey relational grade corresponds to intense
relational degree between the reference sequence x0(k)
and the given sequence xi(k). The reference sequence
x0(k) represents the best process sequence. Therefore,
higher Grey relational grade means that the correspon-
ding parameter combination is closer to the optimal15.
The mean response for the Grey relational grade with its
U. ESME: USE OF GREY BASED TAGUCHI METHOD IN BALL BURNISHING PROCESS ...
130 Materiali in tehnologije / Materials and technology 44 (2010) 3, 129–135
grand mean and the main effect plot of Grey relational
grade are very important because optimal process
condition can be evaluated from this plot15.
3 EXPERIMENTAL DETAILS AND TEST
RESULTS
3.1 Workpiece Material
In this study, high strength precipitation hardening
7XXX series wrought aluminum alloy AA 7075 was
used. The strength and good mechanical properties make
the AA 7075 aluminum alloy appropriate for the use in
aerospace industry. The chemical composition and
mechanical properties of the workpiece material is given
in Table 1.
Table 1: Chemical composition and mechanical properties of AA7075
aluminum alloy
Tabela 1: Kemi~na sestava in mehanske lastnosti aluminijeve zlitine
AA 7075
Chemical
composi-
tion (%)
Al Cu Mg Cr Zn
90.0 1.60 2.50 0.23 5.60
Mecha-
nical pro-
perties
Tensile
strength
(MPa)
Yield
strength
(MPa)
Shear
strength
(MPa)
Fatique
strength
(MPa)
Hardness
(HV100)
220 95 150 160 150
The workpiece material, as shown in Figure 1, was
prepared with the diameter of 30 mm and 70 mm in
length as a three part each having 20 mm length.
3.2 Machines and Equipments
An 18 mm diameter ball was used for burnishing.
The detailed and drawing is shown in Figure 2. When
the ball or roller is pressed against the surface of the
metallic specimen, a pre-calibrated spring was com-
pressed. This spring is being used mainly to reduce the
possible sticking of the tool onto the surface25.
The experiments were performed on a FANUC
GT-250B CNC lathe. The burnishing tool was mounted
on the CNC turret. Dry turning and burnishing were used
in all the experimental work, but alcohol was used to
clean the specimens before burnishing. Cleaning of the
ball was carried out continuously in order to prevent hard
particles from entering on the contact surface between
the tool and the specimen, such hard particles usually
leaving deep scratches, which may damage the burnished
surface of the specimen25.
Phynix TR-100 model surface roughness tester was
used to measure the surface roughness of the burnished
samples. Cut off length was chosen as 0.3 for each
roughness measurement. Vickers microhardness tester
with 100 g load (HV100) was used for microhardness
measurements25. Six measurements of surface roughness
and microhardness were taken from the samples and
average of the values were used in the multi-criteria
optimization.
3.3 Process Parameters and Test Results
In full factorial design, the number of experimental
runs exponentially increases as the number of factors as
well as their level increases. This requires huge
experimentation cost and considerable time. So, in order
to compromise these two adverse factors and to search
the optimal process condition through a limited number
of experimental runs Taguchi’s L16(44) orthogonal array
consisting of 16 sets of data was selected to optimize the
multiple performance characteristics of surface rough-
ness. The burnishing parameters used in this study are
shown in Table 2.
Table 2: Process parameters and their limits
Tabela 2: Parametri in omejitve procesa
Parameters Nota-tion Unit
Levels of factors
1 2 3 4
Burnishing
force F N 58.86
* 117.72 176.58 235.44
Number of
passes N – 1
* 2 3 4
Feed rate f mm/min 0.1
* 0.2 0.3 0.4
Burnishing
speed V rpm 200
* 300 500 700
*Initial factor settings
Table 3 shows the selected design matrix based on
Taguchi L16(44) orthogonal array consisting of 16 sets of
coded conditions and the experimental results for the
U. ESME: USE OF GREY BASED TAGUCHI METHOD IN BALL BURNISHING PROCESS ...
Materiali in tehnologije / Materials and technology 44 (2010) 3, 129–135 131
Figure 2: Detailed drawing of the ball burnishing tool: (1) casing; (2)
adapter cover; (3) spring25
Slika 2: Na~rt gladilnega orodja: (1) ohi{je; (2) prilagoditveni pokrov,
(3) vzmet25
Figure 1: Dimensions of workpiece material25
Slika 1: Mere preizku{anca
responses of Ra and HV100. All these data were utilized
for the analysis and evaluation of optimal parameter
combination required to achieve desired surface quality
within the experimental domain.
Table 3: Orthogonal array L16(44) of the experimental runs and results
Tabela 3: Ortogonalna porazdelitev L16(44) preizkusov in rezultati
Run no
Process parameters Experimentalresults
F N f V Ra (µm) HV100
1 1 1 1 1 0.61 160
2 1 2 2 2 0.57 167
3 1 3 3 3 0.53 173
4 1 4 4 4 0.59 187
5 2 1 2 3 0.36 165
6 2 2 1 4 0.18 178
7 2 3 4 1 0.25 181
8 2 4 3 2 0.20 193
9 3 1 3 4 0.20 172
10 3 2 4 3 0.22 187
11 3 3 1 2 0.08 196
12 3 4 2 1 0.23 210
13 4 1 4 2 0.30 191
14 4 2 3 1 0.15 199
15 4 3 2 4 0.14 205
16 4 4 1 3 0.19 212
4 PARAMETRIC OPTIMIZATION OF
BURNISHING PROCESS
4.1 Evaluation of Optimal Process Condition
First, by using Eqs. (1) and (2), experimental data
were normalized to obtain Grey relational generation15.
The normalized data and 0i for each of the responses
are given in Table 4 and Table 5 respectively. For Ra
smaller-the-better (SB) and for HV larger-the-better (LB)
criterion has been selected.
Table 4: Grey relational generation of each performance characteri-
stics
Tabela 4: Generacija Greyjeve odvisnosti za karakteristike vsake
performance
Run no
Ra HV
Smaller-the-better Larger-the-better
Ideal sequence 1 1
1 0.000 0.000
2 0.043 0.075
3 0.250 0.151
4 0.519 0.038
5 0.096 0.472
6 0.346 0.811
7 0.404 0.679
8 0.635 0.774
9 0.231 0.774
10 0.519 0.736
11 0.692 1.000
12 0.365 0.717
13 0.596 0.585
14 0.750 0.868
15 0.865 0.887
16 1.000 0.792
Table 5: Evaluation of 0i for each of the responses
Tabela 5: Ocena 0i za vsak odgovor
Run no Ra HV
Ideal sequence 1 1
1 1.000 1.000
2 0.957 0.925
3 0.750 0.849
4 0.481 0.962
5 0.904 0.528
6 0.654 0.189
7 0.596 0.321
8 0.365 0.226
9 0.769 0.226
10 0.481 0.264
11 0.308 0.000
12 0.635 0.283
13 0.404 0.415
14 0.250 0.132
15 0.135 0.113
16 0.000 0.208
Table 6 shows the calculated Grey relational
coefficients (with Ra = 0.67, HV = 0.33) of each perfor-
mance characteristic using Eq. (3).
Table 6: Grey relational coefficient of each performance characte-
ristics (Ra = 0.67, HV = 0.33)
Tabela 6: Greyjev odvisnostni koeficient za karakteristike vsake
performance (Ra = 0.67, HV = 0.33)
Run no Ra HV
Ideal sequence 1 1
1 0.333 0.333
2 0.343 0.351
3 0.400 0.371
4 0.510 0.342
5 0.356 0.486
6 0.433 0.726
7 0.456 0.609
8 0.578 0.688
9 0.394 0.688
10 0.510 0.654
11 0.619 1.000
12 0.441 0.639
13 0.553 0.546
14 0.667 0.791
15 0.788 0.815
16 1.000 0.707
The Grey relational coefficients, given in Table 7, for
each response was accumulated by using Eq. (4) to
evaluate Grey relational grade, which is the overall
representative of all the features of burnishing quality.
Thus, the multi-criteria optimization problem was
transformed into a single equivalent objective function
optimization problem using the combination of Taguchi
U. ESME: USE OF GREY BASED TAGUCHI METHOD IN BALL BURNISHING PROCESS ...
132 Materiali in tehnologije / Materials and technology 44 (2010) 3, 129–135
approach and Grey relational analyses. Higher is the
value of Grey relational grade, the corresponding factor
combination is said to be close to the optimal.
Table 7: Grey relational grade
Tabela 7: Greyjeva stopnja odvisnosti
Run no Grey relationalgrade Rank
1 0.333 16
2 0.348 15
3 0.380 14
4 0.397 13
5 0.443 12
6 0.629 6
7 0.558 10
8 0.651 5
9 0.591 8
10 0.606 7
11 0.874 1
12 0.573 9
13 0.548 11
14 0.750 4
15 0.806 2
16 0.803 3
When the Taguchi experimental design is carried out,
a function would be offered (efficiency evaluation) about
one design factor as the standard to evaluate the
efficiency and to understand the experiment efficiency.
While Taguchi experimental design took the quality loss
as the base, it would design one statistic to evaluate
efficiency, which is called the S/N ratio. In this section
we had one quality characteristic that is grey relational
grade for Taguchi analysis26. Table 8 shows the S/N ratio
based on the larger-the-better criterion for overall Grey
relational grade calculated by using Eq. (5).
S N
n y ii
n
/ lg= −
⎡
⎣⎢
⎤
⎦⎥=
∑10 1 12
1
(5)
where n is the number of measurements, and yi is the
measured characteristic value.
Table 8: S/N ratio for overall Grey relational grade
Tabela 8: Razmerje S/N za splo{no Greyjevo odvisnostno stopnjo
Run no S/N
1 -9.54
2 -9.16
3 -8.40
4 -8.02
5 -7.07
6 -4.02
7 -5.06
8 -3.72
9 -4.57
10 -4.34
11 -1.17
12 -4.83
13 -5.21
14 -2.50
15 -1.87
16 -1.90
Graphical representation of S/N ratio for overall Grey
relational grade is shown in Figure 3. The dashed line is
the value of the total mean of the S/N ratio.
As indicated in Figure 3, the optimal condition for
the burnishing of AA7075 aluminum alloy becomes
F4N3f1V4. Table 9 shows the mean Grey relational grade
ratio for each level of the process parameters.
Table 9: Response table for the mean Grey relational grade
Tabela 9: Tabela odgovorov za povpre~no Greyjevo odvisnostno
stopnjo
Factors
Grey relational grade
Level 1 Level 2 Level 3 Level 4 Max-Min
F 0.36 0.57 0.66 0.73 0.37
N 0.48 0.58 0.65 0.61 0.17
f 0.66 0.54 0.59 0.53 0.13
V 0.55 0.61 0.56 0.61 0.06
Total mean Grey relational grade= 0.58
4.2 Analysis of Variance (ANOVA)
The purpose of the analysis of variance (ANOVA) is
to investigate which burnishing parameters significantly
affect the performance characteristices. This is
accomplished by separating the total variability of the
grey relational grades, which is measured by the sum of
the squared deviations from the total mean of the grey
relational grade, into contributions by each burnishing
parameters and the error. Thus;
SS SS SST F e= + (6)
where
SS j m
j
p
T = −
=
∑ ( )
 
 2
1
(7)
U. ESME: USE OF GREY BASED TAGUCHI METHOD IN BALL BURNISHING PROCESS ...
Materiali in tehnologije / Materials and technology 44 (2010) 3, 129–135 133
Figure 3: S/N ratio plot for the overall Grey relational grade
Slika 3: S/N razmerje za splo{no Greyjevo stopnjo
and
SST Total sum of squared deviations about the mean

j Mean response for jth experiment

m Grand mean of the response
p Number of experiments in the orthogonal array
SSF Sum of squared deviations due to each factor
SSe Sum of squared deviations due to error
In addition, the F test was used to determine which
burnishing parameters have a significant effect on the
performance characteristic. Usually, the change of the
burnishing parameter has a significant effect on the
performance characteristics when the F value is large.
ANOVA for overall Grey relational grade is shown in
Table 10.
Table 10: ANOVA results
Tabela 10: ANOVA rezultati
Parameter Degree ofFreedom
Sum of
Square
Mean
Square
F Contribu-tion (%)
F 3 0.300 0.100 19.00 71.59
N 3 0.066 0.022 4.20 15.75
f 3 0.043 0.014 2.72 10.26
V 3 0.009 0.003 0.63 2.14
Error 3 0.001 0.005 0.23
Total 15 0.419 100
According to this analysis, the most effective
parameters with respect to surface roughness and
microhardness are burnishing force, number of passes,
burnishing feed and burnishing speed. Percent
contribution indicates the relative power of a factor to
reduce variation. For a factor with a high percent
contribution, has a great influence on the performance.
The percent contributions of the burnishing parameters
on the surface roughness and microhardness are shown
in Table 10 and Figure 4. Burnishing force (71.59%)
was found to be the major factor affecting surface
roughness and microhardness, whereas number of passes
(15.75%), burnishing feed (10.26%) and burnishing
speed (2.14%) were found to be the second, third and
fourth ranking factor respectively.
4.3 Confirmation Test
After evaluating the optimal parameter settings, the
next step is to predict and verify the enhancement of
quality characteristics using the optimal parametric
combination. The estimated Grey relational grade 

using the optimal level of the design parameters can be
calculated as:

 ( )
 
 
= + −
=
∑m j m
i
o
1
(8)
where 
m is the total mean Grey relational grade, 
i is
the mean Grey relational grade at the optimal level, and
o is the number of the main design parameters that
affect the quality characteristics. Table 11 indicates the
comparison of the predicted surface roughness and
microhardness with that of actual by using the optimal
burnishing conditions; good agreement between the
actual and predicted results was obtained. Also, impro-
vement in overall Grey relational grade was found to be
as 0.65.
Table 11: Results of confirmation test
Tabela 11: Rezultati potrditvenih preizkusov
Initial
factor
settings
Optimal process
condition
Prediction Experiment
Factor levels F1N1f1V1 F4N3f1V4 F4N3f1V4
Ra 0.61 0.12
HV100 160 200
S/N ratio of overall Grey
relational grade -9.54 -0.51 -0.80
Overall Grey relational
grade 0.33 0.91 0.98
Improvement in Grey relational grade=0.65
In Taguchi method, the only performance feature is
the overall Grey relational grade and the aim should be
to search a parameter setting that can achieve highest
overall Grey relational grade. The Grey relational grade
is the representative of all individual performance cha-
racteristics. In this study, objective functions have been
selected in relation to parameters of surface roughness
and microhardness. The weight calculations were done
by using Analytic Hierarchy Process (AHP) and the
weights were found to be as 0.67 and 0.33 for the res-
ponses of surface roughness and microhardness respecti-
vely. The results showed that using optimal parameter
setting (F4N3f1V4) caused lower microhardness together
with higher microhardness.
5 CONCLUSIONS
Taguchi method is a very effective tool for process
optimization under limited number of experimental runs.
Essential requirements for all types burnishing processes
are smoother surface with higher surface hardness. This
study has concentrated on the application of Taguchi
method coupled with Grey relation analysis for solving
U. ESME: USE OF GREY BASED TAGUCHI METHOD IN BALL BURNISHING PROCESS ...
134 Materiali in tehnologije / Materials and technology 44 (2010) 3, 129–135
Figure 4: Contribution percentage of the burnishing parameters
Slika 4: Odstotna porazdelitev parametrov glajenja
multi criteria optimization problem in the field of
burnishing process. Experimental results have shown that
surface roughness and microhardness of burnished
aluminum alloy are greatly improved by using Grey
based Taguchi method.
6 REFERENCES
1 L. Kukielka, Journal of Mechanical Technology, 19 (1989), 319–356
2 M. H. El-Axir, International Journal of Machine Tools &
Manufacture, 40 (2000), 1603–1617
3 N. H. Loh, S. C. Tam, S. Miyazawa, Journal of Mechanical Working
Technology, 18 (1989), 53–61
4 A. M. Hassan, Journal of Materials Processing Technology, 72
(1997), 385–391
5 C. Wick, R. F. Veilleux, Soc. Manuf. Eng., 3 (1985), 16–38
6 T. Siva Prasad, B. Kotiveerachari, J. Inst. Eng., 69 (1988), 55–58
7 G. Y. Shneider, Machines and Tooling, 1 (1989), 34–35
8 A. N. Niberg, Soviet Engineering Research, 7 (1987), 67
9 P. C. Michael, N. Saka, E. Rabinowicz, Wear, 132 (1989), 265
10 N. H. Loh, S. C. Tam, S. Miyazawa, Wear, 129 (1989), 235
11 I. Yashcheritsyn, E. I. Pyatosin, V. V. Votchuga, Soviet Journal of
Friction and Wear, 8 (1987), 87
12 M. Fattouh, M. H. El-Axir, S. M. Serage, Wear, 100 (1988), 127
13 A. S. H. Al-Bsharat, Investigation into the effect of burnishing pro-
cess on some properties of cylindrical metallic materials, Jordan
University of Science and Technology, MSc Thesis, (1994)
14 O. M. Abd Al-Wahhab, The effects of roller-burnishing on some pro-
perties of non-ferrous metals, Jordan University of Science and
Technology, MSc Thesis, (1996)
15 S. Datta, A. Bandyopadhyay, P. K. Pal, International Journal of
Advanced Manufacturing Technology, 39 (2008), 1136–1143
16 W. H., Yang, Y. S. Tarng, Journal of Materials Processing Techno-
logy, 84 (1998), 122–129
17 H. Rowlands, J. Antony, The TQM Magazine, 12 (2000), 78–83
18 S. H. Lim, C. M. Lee, W. J. Chung, International Journal of Precision
Engineering and Manufacturing, 7 (2006), 18–23
19 D. C. Chen, C. F. Chen, Journal of Materials Processing Technology,
190 (2007), 130–137
20 C. P. Fung, P. C. Kang, Journal of Materials Processing Technology,
170 (2005), 602–610
21 S. H. Tang, V. J. Tan, S. M. Sapuan, S. Sulaiman, N. Ismail, R.
Samin, Journal of Materials Processing Technology, 182 (2007),
418–426
22 P. Vijian, V. P. Arunachalam, Journal of Materials Processing Tech-
nology, 180 (2006), 161–166
23 L. J. Yang, Journal of Materials Processing Technology, 185 (2007),
113–119
24 J. Z. Zhang, J. C. Chen, E. D. Kirby, Journal of Materials Processing
Technology, 184 (2007), 233–239
25 U. Esme, A. Sagbas, F. Kahraman, M. K. Kulekci, Materials and
Technology 42 (2008), 215–219
26 F. C. Kung, Y. J. Weng, Polymer-Plastics Technology and Engi-
neering, 47 (2008), 1154–1161
U. ESME: USE OF GREY BASED TAGUCHI METHOD IN BALL BURNISHING PROCESS ...
Materiali in tehnologije / Materials and technology 44 (2010) 3, 129–135 135

S. KORES ET AL.: EFFECT OF CERIUM ADDITIONS ON THE AlSi17 CASTING ALLOY
EFFECT OF CERIUM ADDITIONS ON THE AlSi17
CASTING ALLOY
DODATEK CERIJA LIVNI ZLITINI AlSi17
Stanislav Kores, Maja Von~ina, Borut Kosec, Primo` Mrvar, Jo`ef Medved
University of Ljubljana, Faculty of Natural Sciences and Engineering, Department of Materials and Metallurgy,
A{ker~eva 12, 1000 Ljubljana, Slovenia
stanislav.kores@ntf.uni-lj.si
Prejem rokopisa – received: 2010-01-08; sprejem za objavo – accepted for publication: 2010-02-01
The influence of the primary silicon phase and of eutectic silicon on the solidification process of hypereutectic Al-Si alloys with
the addition of cerium is presented. The solidification was analyzed by a simple thermal analysis and a simultaneous thermal
analysis, and the microstructures were examined with conventional light microscopy and scanning electron microscopy. The
simultaneous refinement of both the primary and the eutectic silicon particles was achieved with cerium additions. The results
showed that the addition of cerium had a very positive influence on the alloy’s tensile strength.
Keywords: hypereutectic Al-Si alloys, cerium, silicon refinement, modification
^lanek opisuje vpliv dodatka cerija na primarno fazo silicija in na evtektski silicij pri strjevanju nadevtektskih Al-Si-zlitinah.
Proces strjevanja je bil analiziran z enostavno in s simultano termi~no analizo. S konvencionalnim svetlobnim mikroskopom ter
z elektronskim mikroskopom je bila opravljena mikrostrukturna analiza. Z dodatkom cerija je bilo dose`eno udrobnjevanje tako
primarno izlo~ene faze silicija kot tudi evtektskega silicija. Preiskave so pokazale, da ima dodatek cerija pozitiven u~inek na
natezno trdnost zlitin.
Klju~ne besede: nadevtektske Al-Si zlitine, cerij, udrobnjevanje silicija, modifikacija
1 INTRODUCTION
The application of aluminium alloys for automotive
parts, such as pistons, cylinder liners, engine blocks and
other parts is important because of the alloys’ wear
resistance, heat resistance and low thermal expansion.1,2
The dominant group of Al-Si foundry alloys contains the
mass fraction of Si between 5 and 25 %, and additions of
Mg, Ni and Cu. Silicon, as the major alloying addition,
represents one of the most effective ways to obtain
high-quality aluminium casting alloys, mainly because
of their high fluidity due to the relatively large volume of
the Al-Si eutectic.3, 4 Al-Si alloys with more than 12 % Si
have a hypereutectic microstructure, usually consisting
of a primary silicon phase in a eutectic matrix. These
alloys are used for engine blocks or cylinder liners, and
are the ideal solution for the substitution of cast-iron
cylinder liners in the manufacturing of a monolithic
engine block.5, 3
The primary silicon phase has a harmful effect on the
extrudability, machinability, strength, and ductility of the
alloy, since coarse silicon phases lead to premature crack
initiation and fracture when in tension.4 With the aim to
avoid these problems, the structural modifications are
achieved with the addition of elements that refine the
silicon phases to a better shape.1
The use of silicon modifiers and high cooling rates
refines the primary silicon particles.6 Sodium and
strontium are both effective modifiers of the Al-Si
eutectic, but it is difficult to control and maintain the
modification effect of sodium because of the volatili-
zation and oxidation losses, especially when longer
holding times and higher temperatures are required.
Strontium does not cause such problems; however, it is
reported that it may cause micro- and macroporosity in
Al-Si alloy castings.7 The refinement of primary silicon
grains is usually achieved with the addition of
phosphorus to the melt. The diagram in Figure 1 shows
the influence of phosphorus additions on the size of the
primary silicon grains8. It is reported4,7–12 that rare-earth
elements are capable of modifying the Al-Si eutectic too,
and it has also been proved that rare-earth elements
Materiali in tehnologije / Materials and technology 44 (2010) 3, 137–140 137
UDK 669.131:669.85/.86 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 44(3)137(2010)
Figure 1: Influence of phosphorus additions on the refinement of
primary silicon particles7
Slika 1: Vpliv dodatka fosforja na udrobnitev primarnega silicija7
represent one way to achieve longer effects for refining
agents.7 It is, for example, mentioned in reference13 that
cerium did not refine the primary silicon grains, but it
had a moderate effect on modifying the silicon eutectic.
A small cerium addition could provide ternary high-
temperature stable 1 and 2 phases in equilibrium with
the Al-Si-rich melt. These phases might act as nucleation
sites for (aluminium) or (silicon) particles in both hypo-
and hypereutectic Al-Si alloys.14 Kowata et al.15 reported
that the primary Si in hypereutectic Al-20 % Si alloys
was refined with rare-earth metals.
The goal of the present investigation was to check the
possibility of whether cerium acted as a refining agent
for the primary silicon phase and the silicon eutectic in
the pure AlSi17 alloy and in the commercial
AlSi17CuMg alloy. The influence of various cerium
additions on the tensile strength was also examined.
2 EXPERIMENTAL
Pure Al-Si alloys with the mass fraction of Si 17 %
and the commercial AlSi17CuMg alloy and various
additions of cerium were used for the experiments. The
alloys were melted in an electric resistance furnace in a
3-kg crucible and cerium was added as a 99.9 % pure
metal. The melt with a temperature of 750 °C was
poured into the measuring cast-iron cell for the thermal
analysis. The cooling curve was recorded with a K-type
thermocouple at a cooling rate of approximately 20 K/s.
With the Thermo-Calc program we recorded all the
thermodynamically possible equilibrium phases existing
under the defined conditions. The equilibrium binary
phase diagram was determined and the composition of
the microstructure could be predicted. The samples for
light microscopy were prepared with a standard metallo-
graphic procedure and examined with an Olympus BX61
light microscope equipped with a DP 70 video camera.
The simultaneous thermal analysis (STA) of specimens
of the starting materials was performed with the STA
449 Netzsch apparatus. Two equal corundum cups were
placed on the platinum sensor. One contained the
examined material, the other one, which was empty,
represented the reference material. The measurements
were carried out in a protective atmosphere of 99.999 %
pure argon. The specimens were heated up to 720 °C at a
heating rate of 10 K/min and cooled with the same
cooling rate to room temperature.
3 RESULTS AND DISCUSSION
To act as a nucleus, the compound must precipitate
prior to the primary silicon phase. A thermodynamic
simulation of the AlSi17 alloy with 80 µg/g (ppm)
phosphorus was made with the Thermo-Calc program,
and Figure 2 a shows that the AlP compound precipi-
tates before the primary silicon phase.
The evidence that cerium also forms nuclei, a
simulation with the Thermo-Calc program, was carried
out for various additions of cerium. Figure 2 b presents
the isopleths of the phase diagram for the AlSi17 alloy
with 4.5 % Ce. It is evident that the AlCe compound
precipitated before the primary silicon. The several
proposed mechanisms of refinement for primary silicon
grains by the addition of phosphorus can be summarized
as follows. The widely accepted and popular theory is
based on the nucleation of silicon grains on the AlP
compound when phosphorus is added to the alloy. It has
been proposed that the AlP compound with the melting
point above 982 °C and a cubic structure similar to that
of silicon, acted as a nucleation agent and with
heterogeneous nucleation produced the refinement of the
primary grains 4.
The modification effect of the cerium addition on the
solidification of the AlSi17 alloy is shown on the cooling
curve in Figure 3. The addition of 0.5 % Ce affected the
primary silicon size distribution, and this could be seen
on the cooling curve without undercooling. The
solidification time was shorter than with the pure AlSi17
alloy. The diagram demonstrates that the alloy with 4.5
% Ce had a lower liquidus temperature, below 600 °C.
At the same time, the solidification time was longer than
for the pure alloy. The idea that added cerium formed
nuclei for the primary silicon precipitation was refuted,
S. KORES ET AL.: EFFECT OF CERIUM ADDITIONS ON THE AlSi17 CASTING ALLOY
138 Materiali in tehnologije / Materials and technology 44 (2010) 3, 137–140
Figure 2: Isopleth diagram of AlSi17 alloy with 80 µg/g (ppm) P (a),
and the mass fraction of Ce w(Ce) = 4.5 % (b)
Slika 2: Izopletni fazni diagram zlitine AlSi17 z 80 µg/g (ppm) P (a)
in masni dele` Ce w(Ce) = 4,5 % (b)
since the liquidus temperature was lower than with the
pure alloy.
Figure 4 shows the microstructures of the un-
modified (a) and modified AlSi17 alloy with 1 % Ce (b)
in the light microscope. It demonstrates that this addition
modified the size of the primary silicon grains, from
coarse shapes to fine polyhedral shapes. Also, the
microstructure of the eutectic was finer.
The microstructure of the AlSi17 alloy with 4.5 % Ce
is presented in Figure 5. It demonstrates that the higher
cerium content does not cause the refinement of the
primary silicon grains. A phase with a highly polyhedral
shape that contained cerium was found in the micro-
structure.
STA was performed with specimens from the thermal
analysis. Figure 6 indicated that the liquidus temperature
of the alloy decreased with an increasing cerium content.
The greatest reduction of the liquidus temperature was
with the alloy containing 1 % Ce, and the drop was from
686.6 °C (pure AlSi17 alloy) to 591.9 °C.
With the aim to prove that pure cerium possibly
forms nuclei, the sample containing 1 % cerium was
S. KORES ET AL.: EFFECT OF CERIUM ADDITIONS ON THE AlSi17 CASTING ALLOY
Materiali in tehnologije / Materials and technology 44 (2010) 3, 137–140 139
Figure 5: AlSi17 alloy with w = 4.5 % Ce
Slika 5: Zlitina AlSi17 z w = 4,5 % Ce
Figure 3: Effect of cerium addition on the AlSi17 alloy
Slika 3: U~inek dodatka cerija na zlitino AlSi17
Figure 4: Microstructures of the AlSi17 alloy (a) and the alloy with w
= 1 % addition of cerium
Slika 4: Mikrostruktura zlitine AlSi17 (a) in zlitine z dodatkom w = 1 %
cerija
Figure 7: Scanning electron micrograph of the AlSi17 with w = 1 %
Ce
Slika 7: Posnetek z vrsti~nim elektronskim mikroskopom zlitine
AlSi17 z w = 1 % Ce
Figure 6: DSC-curves for the AlSi17 alloy with various additions of
cerium
Slika 6: DSC-krivulje zlitine AlSi17 z razli~nimi dodatki cerija
examined with a scanning electron microscope. Figure 7
shows a micrograph of the examined alloy. Cerium did
not form any nuclei for the precipitation of the primary
silicon phase and precipitated on the surface of the
primary silicon phase itself.
The examination of various cerium additions to the
commercial AlSi17CuMg alloy with the mass fractions
1.6 % Cu, and 0.3 % Mg was performed too. Figures 8 a
to 8 d show the microstructures of the base AlSi17CuMg
alloy (a), of alloys with 0.5 mass % Ce (b), with 1 mass
% Ce (c) and with 1.5 % Ce. The addition of 1 % Ce had
the greatest effect with regards to refining the primary
silicon phase and also of the Al-Si eutectic.
The influence of cerium additions on the mechanical
properties of the AlSi17CuMg alloy was determined
with tensile tests. The results of the tensile strength tests
are presented in Table 1. It was found that the cerium
additions had a positive influence on the tensile strength.
Table 1: Tensile strengths of the examined test samples
Tabela 1: Natezne trdnosti preiskovanih vzorcev
Alloy Pure alloy with w =0.5 % Ce
with w =
1 % Ce
with w =
1.5 % Ce
AlSi17CuMg 170 MPa 192 MPa 195 MPa 186 MPa
4 CONCLUSIONS
The refinement of the primary silicon and of the
eutectic silicon morphology was achieved with additions
of cerium (up to 1 % Ce) to hypereutectic Al-Si alloys.
The best results for the microstructural and strength
properties were obtained with the addition of 1 % Ce.
With an increase in the cerium additions, the
precipitation temperature of the primary silicon phase
decreased. The addition of 1 % Ce produced the greatest
reduction in the liquidus temperature, from 686.6 °C to
591.9 °C.
5 REFERENCES
1 Y. Hong-Kun, Z. Di, Transactions of Nonferrous Metals Society,
China, April (2003) 358
2 O. Zak, B. Tonn, Giesserei-Praxis, Germany, September (2009) 281
3 F. C. Robles Hernandez, J. H. Sokolowski, Journal of Alloys and
Compounds, 419 (2006), 180
4 J. Chang, I. Moon, C. Choi, Journal of Materials Science 33 (1998)
5015
5 J. E. Gruzleski and B. M. Closset, American Foundrymen’s Society,
Inc., Des Plaines, Illinois, (1990) 19
6 L. Bäckerud, G. Chai, J. Tamminem, Solidification Characteristics of
Aluminium Alloys, American Foundry Society/Skanaluminium, U.
S. A Vol. 3. (1990)
7 B. J. Ye, C. R. Jr. Loper, D.Y. Lu, et al., AFS Transaction, 30 (1985)
533
8 W. Vogel, V. Schneider, Giesserei 78, Nr. 23, (1991), 848–852
9 H. Sano, Metal Powder Report (UK), 49 (1) (1994) 26
10 J. Y. Chang, G. H. Kim, I. G. Moon at al., Scripta Materialia, 39
(1998) 307
11 R. Sharan, N. P. Saksen, International Cast Metals Journal 3 (1978)
29
12 M. Ravi, U. T. S. Pillai, G. C. Pai, Metallurgical and Materials
Transactions 27A (1996) 1283
13 J. C. Wiess, C. R. Loper Jr., AFS Transaction 32 (1987) 51
14 J. Gröbner, D. Mirkovi}, R. Schmid-Fetzer, Metallurgical and Mate-
rials Transactions A, 35A (2004) 3349
15 T. Kowata, H. Horie, S. Hiratsuka, A. Chida, Imono 66 (1994) 803
S. KORES ET AL.: EFFECT OF CERIUM ADDITIONS ON THE AlSi17 CASTING ALLOY
140 Materiali in tehnologije / Materials and technology 44 (2010) 3, 137–140
Figure 8: Influence of cerium additions on the microstructure of the
AlSi17CuMg alloy; a) pure alloy, b) with mass fractions w = 0.5 %
Ce, c) with w = 1 % Ce, d) with w = 1.5 % Ce
Slika 8: Vpliv dodatka cerija na mikrostrukturo zlitine AlSi17CuMg,
a) brez dodatka, b) z masnimi dele`i w = 0,5 % Ce, c) z w = 1 % Ce,
d) z w = 1,5 % Ce
V. CHYGYRYNS’KYY ET AL.: A GENERALIZED THEORY OF PLASTICITY
A GENERALIZED THEORY OF PLASTICITY
POSPLO[ENA TEORIJA PLASTI^NOSTI
Chygyryns’kyy Victorovich Valeryy1, Kachan Aleksey Yakovlevich1,
Ilija Mamuzi}2, Ben’ Anna Nikolaevna1
1Zaporozhskyy National Technical University, Str. Zukovsky 64, 69063 Zaporozhskye, Ukraine
2University of Zagreb, Faculty for Metallurgy, Aleja narodnih heroja 3, 44000 Sisak, Croatia
valerij@zntu.edu.ua
Prejem rokopisa – received: 2008-11-21; sprejem za objavo – accepted for publication: 2009-11-23
A closed solution of the plane problem of the generalized theory of plasticity and a model of the complex plastic medium were
theoretically developed. Solutions with the use of the deformation theory and the theory of plastic yielding were developed. The
solution for a simple strengthening medium was deduced.
Key words: metal plasticity, analytical solution, mathematical model, plastic medium, process parameters
Razvita sta bila zaprta re{itev splo{ne teorije plasti~nosti in teoreti~en model kompleksnega plasti~nega medija. Opredeljene so
bile re{itve z uporabo teorije deformacije in teorije plasti~nega te~enja. Razvita je bila re{itev za preprost utrditveni medij.
Klju~ne besede: plasti~nost kovin, analiti~na re{itev, matermati~ni model, plasti~ni medij, parametri procesa
1 INTRODUCTION
A characteristic of the new method based on a closed
solution of the plane problem of theory of plasticity is a
simplified analysis of the deformation mode of the
medium and the theoretical connection to the medium
mechanical characteristics through the process para-
meters. The analytical solution of the plane problem of
the theory of plasticity for a strengthening medium is
known.1 The developed complex model for the
strengthening of the plastic medium is based on the shear
resistance to the plastic deformation and is a function of
the coordinates of the nucleus of deformation. This
approach offers a new possibility to evolve a new
solution for a problem, including the generalized theory
of plasticity. The approach includes equations and
criteria: an equilibrium equation, and the criteria of
yielding, the equation of incompressibility, of the
deformation rate and the deformation as well as
equations of continuity of the deformation rate and the
deformation:
– the equilibrium equations
∂
∂
∂
∂
 x xy
x y
+ = 0;
∂
∂
∂
∂
 xy y
x y
+ = 0
– the criterion of yielding
( )  x y xy k− + ⋅ = ⋅
2 2 24 4
– the constraint equations for the rates of deformation
and deformation
 

 


x y
xy
x y
xy
F
−
⋅
=
−
=
2 1'
;
 

 


x y
xy
x y
xy
F
−
⋅
=
−
=
2 2
(1)
– the equations of incompressibility for the rates of
deformation and the deformation
 x y+ = 0;  x y+ = 0
– the equation of continuity for the deformation rates
and the deformation
∂
∂
∂
∂
∂
∂ ∂
2
2
2
2
2
  
x y xy
y x y x
+ =
'
;
∂
∂
∂
∂
∂
∂ ∂
2
2
2
2
2
  
x y xy
y x y x
+ =
– the equation of heat conductivity
a
T
x
T
y
T
t
2
2
2
2
2
∂
∂
∂
∂
∂
∂
+
⎛
⎝
⎜
⎞
⎠
⎟ =
The model of the complex plastic medium is defined
with
T H Tm m mi i i= ⋅ ⋅ ⋅ ( ) ( ) ( )1 2 3G (2)
The system of equations (1) includes the equations of
the deformation theory of plasticity and the theory of
plastic yielding with the addition of the equation of heat
conductivity.2 The model (2) is a real strengthening
medium with the boundary conditions for stresses3
[ ] n i= ⋅ −T Asin F 2 , Ti = k
or 
 
  n =
−
⋅ ⋅
x y
xy2
2 2sin cos (3)
The additional conditions are given by the specific
contact forces (3) of the change of friction according to
the sinusoidal law of deformational and high-speed strain
hardening. All the intensities and the temperature depend
on the coordinates of the deformation nucleus.
Materiali in tehnologije / Materials and technology 44 (2010) 3, 141–145 141
UDK 539.374.001.8.621.7-111 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 44(3)141(2010)
2 THEORETICAL DEVELOPMENT
With the aim to obtain the model (2), let us consider
three second-order equations in form of non-uniform
hyperbolic partial derivations:
∂
∂
∂
∂
∂
∂ ∂
2
2
2
2
2
22
 
 	
xy xy
xyx y y x
k− = ⋅ − ,
∂
∂
∂
∂
∂
∂ ∂
2
2
2
2
2
1
2
1 

x x
xy x y x F
− = ⋅ ⋅ (4)
∂
∂
∂
∂
∂
∂ ∂
2
2
2
2
2
2
2
1 

x x
xy x y x F
− = ⋅ ⋅
The boundary conditions (3) correspond to the
substitution  xy k A= ⋅ sin F. A complex dependence of
the coordinates is assumed with k f H T x yi i= ( , , , , )G . In
this case, k C= ⋅ 
exp ', with 
' ( , , , , )= f H T x yi iG , with
Gi, Hi, T standing for the intensity of the deformation, the
rates of deformation and the temperature.
The derivatives are taken as for the complex
function,4 and after substitution in the first equation (3)
we obtain:
{ [( )' ' ' ' '
 
 
 
 
 H x x t x x H x xH T H⋅ + ⋅ + ⋅ + ⋅ + ⋅ +G G
]+ ⋅ + − ⋅ + ⋅ + ⋅ −
 
 
 
t x y H y y t y yT A H T' ' ' ') ( )F G2
[ ] }− ⋅ + ⋅ + ⋅ − + ⋅( )' ' '
 
 
H y y t y x xyH T A AG F F2 2
[ ]{⋅ + ⋅ ⋅ + ⋅ + ⋅ + ⋅sin( ) ( )' ' 'A H T AH x x t x yF G F2 
 
 

[ ]⋅ − ⋅ + ⋅ + ⋅ + −A H T Ax H y y t y xxF G F( )' ' '
 
 

− − ⋅ ⋅ ⋅ + ⋅ +A H H Hyy HH x y H xyF 2 (
' '
 

}− ⋅ ⋅ + ⋅ + ⋅ ⋅ + ⋅
 
 
 
 ' ' ' ' )G G Gx y xy tt x y t xyT T T ·
·cos (AF) = 0 (5)
Equation (5) is equal to zero if the parts in the square
brackets are equal to zero. Then,

 
 
H x x t x yH T A
' ' '⋅ + ⋅ + ⋅ =G F

 
 
H y y t y xH T A
' ' '⋅ + ⋅ + ⋅ =G F
(
 
 
H x x t x x yxH T A
' ' ' )⋅ + ⋅ + ⋅ = −G F
(
 
 
H y y t y y xyH T A
' ' ' )⋅ + ⋅ + ⋅ = −G F
A H H Hyy HH x y H xy x yF G G

 
 
 
' ' '⋅ ⋅ + ⋅ ⋅ ⋅ +

 
 

' ' ' )⋅ + ⋅ ⋅ + ⋅G xy tt x y t xyT T T
A H H Hxx HH x y H xy x yF G G
 
 
 
 
' ' '⋅ ⋅ + ⋅ ⋅ ⋅ +

 
 

' ' ' )⋅ + ⋅ ⋅ + ⋅G xy tt x y t xyT T T
The operations with the complex function allow us to
determine the exponent index as the sum of three
functions accounting for the effect of the deformation
degree and the rate, and of the temperature:

 
 
 
 
 
 
 
 	   	 
' ' ' ' ' ' '(= − = + + = −A A A A
The shear resistance and the components of the
tensor of the stresses are:
k C A A A A= ⋅ − ⋅ − ⋅ − ⋅  	 
 
 
 exp( exp( exp( sin(
' ' ' F (6)
 
 
 
   	 x C A A A A= ⋅ − ⋅ − ⋅ − ⋅exp( exp( exp( cos(
' ' ' F
 0 + +f y C( )
 
 
 
   	 y C A A A A= − ⋅ − ⋅ − ⋅ − ⋅exp( exp( exp( cos(
' ' ' F
 0 + +f x C( )
with 
 
 
 
 	 x x x x yA
' ' ' '( ) ( ) ( )= + + = − F

 
 
 
 	 y y y y xA
' ' ' '( ) ( ) ( )= + + = F
By substituting the stress values into the equation of
constraint we obtain:
 

x y
xy
A
−
=
2
ctg F;
 


x y
xy
B
−
=' ctg 1F
ctg ctgA B FF F= =1 1
 


x y
xy
B
−
=' ctg 2F
ctg ctgA B FF F= =2 2
It is possible to establish the relation between the
shears and the linear figures of the deformation rates and
the deformations. Taking into account the equations of
non-compressibility we obtain:

  xy x xF
B' = ⋅ ⋅ ⋅ ⋅2
1
1
1= 2 tg F

  xy x xF
B= ⋅ ⋅ ⋅ ⋅2
1
2
2= 2 tg F
In order to simplify, we define:
 
 x C B= ⋅ ⋅exp cos
''
1F
 
 	x C B= ⋅ ⋅exp cos
''
2F
By substituting these relations into the equations of
continuity of the deformation rate and the deformation
(1) or (4), we obtain:
[ ]− − + + + − ⋅ +
 
 
 
   xx x y yy y xB B B' ' ' ' ' ' ' '( ) ( ) sin1 2 1 1F F F
[ ]+ − ⋅ + + − ⋅2 1 1 1 1( ) ( ) ( )'' ' 'B B B Bx y x y xx yyF F F F
 
 
⋅ = ⋅ ⋅ + ⋅ ⋅cos sin cos''B B B Bxy xy1 1 1 12 2F F F F
 (7)
as well as
[ ]− − + + + − ⋅ +
 
 
 
	 	 	 	xx x y yy y xB B B' ' ' ' ' ' ' '( ) ( ) sin2 2 2 2F F F
[ ]+ − ⋅ + + − ⋅2 2 2 2 2( ) ( ) ( )'' ' 'B B B Bx y x y xx yyF F F F
 
	 	
⋅ = ⋅ ⋅ + ⋅ ⋅cos sin cos''B B B Bxy xy2 2 2 22 2F F F F
 	 (8)
Brackets identical to (5) appear in equations (7) and
(8). For


' ' ) x yB= − 1F 

' ' ) y xB= 1F

 	
' ' ) x yB= − 2F 
 	
' ' ) y xB= 2F
the equations are transformed into identities, with,

 

' ' = −B1 , 
 
	
' ' = −B2 as the indices of the exponents of
the functions determining the fields of the deformation
rate and the deformation, B1F and B2F are the trigo-
nometric functions determining the fields of the defor-
mation rate and the deformation.
The expressions for the deformation rate and the
deformation are:
142 Materiali in tehnologije / Materials and technology 44 (2010) 3, 141–145
V. CHYGYRYNS’KYY ET AL.: A GENERALIZED THEORY OF PLASTICITY
  
 x y C B= − = ⋅ ⋅ =exp cos
''
1F
= ⋅ − ⋅C B B 
exp ( cos1 1F (9)

 
 xy C B
' ' ' sin= ⋅ ⋅exp 1F = ⋅ − ⋅C B B 
exp( 1 1sin F
H C C Bi = ⋅ = ⋅ −2 2 1  
 
exp exp (
''
  
 	x y C B= − = ⋅ ⋅ =exp
'' cos 2F
= ⋅ ⋅C B B 
exp(– 2 2cos F (10)

 
 	xy C B= ⋅ ⋅exp
'' sin 2F = ⋅ − ⋅C B B 
exp( 2 2sin F
G i C C B= ⋅ = ⋅ −2 2 2 	 
 
exp exp (
''
with 

' ' ) y xB= 1F 

' ' ) x yB= − 1F

 	
' ' ) y xB= 2F 
 	
' ' ) x yB= − 2F
With a comparison of expressions (9), (10) and (7)
we can confirm that all the expressions have functional
dependencies on the coordinates 
 and F (the indices of
the exponents and the examples of the trigonometric
functions).
It is of interest to obtain similar dependencies for the
solution for the temperature fields that could allow us to
solve this task theoretically. Let us consider a differential
equation for the stationary temperature field
∂
∂
∂
∂
2
2
2
2 0
T
x
T
y
+ =
For this case we look for the solution in the form of
T C B BT= ⋅ ⋅ +exp (
 
' ' ) (sin cos )3 3F F (11)
with: (
 
' ' ) x yB= − 3F , (
 
' ' ) y xB= − 3F
We will demonstrate that expression (11) is a solution
of the Laplace equation. By substituting the derivatives
(11) into the equation of the heat conductivity and a
simplification we obtain:
{ [ ] [ ]( + ( ( (
 
 
 
   ' ' ' ' ' ' ' ') ) ) )xx x y x y yyB B+ ⋅ − + +3 3F F
[ ] [ ]}+ ( (
 
 ' ' ' ') ) (siny x y xB B B+ ⋅ − ⋅ +3 3 3F F F
[+ B B Bx x xx ycos ) ) )
'' ' '
3 3 32 2F F F+ ⋅ ⋅ + + ⋅ ⋅( (
 
 
]⋅ ⋅ − =B +B B By yy3 3 3 3 0F F F F(cos sin ) (12)
In the case of equality to zero, the brackets,
[ ](
 ' ' ) x yB+ 3F , [ ](
 ' ' ) y xB− 3F in equation (12) esta-
blish a relation in the form of:
(
 
' ' ) xx yxB= − 3F , (
 
' ' ) yy xyB= − 3F
B xx yx3F = (
 
' ' ) , B yy xy3F = (

' ' )
The last correlations correspond to Cauchy-Rieman
condition and are functions determined by the Laplace
equation corresponding to relation (11).
From the comparison of the solutions (7) to (11)
(conditions superimposed on functions) it was concluded
that 
 

' ' = −B3 for the stressed and deformed conditions
and the temperature fields can be used to determine a
common parametric function, which is included into the
fields of the stresses, the deformations, the rates of
deformation and the temperatures, allowing us to express
them mathematically, one with another. Thus,
exp (−
⋅
⎛
⎝
⎜
⎞
⎠
⎟ =
⋅
⎛
⎝
⎜
⎞
⎠
⎟ =


 
H
C C
i
B
i
B
2 2
1 1
1 2G


T
C B BT
B
⋅ +
⎛
⎝
⎜
⎞
⎠
⎟
(sin cos )3 3
1
3
F F
With substitution into an expression for resistance to
deformation, we obtain
( ) ( ) ( )T H Ti i
A
B
i
A
B
A
B= ⋅ ⋅
1
1
2
2
3
3
' ' '
'G (13)
The form of expression (13) corresponds to a depen-
dence of the yield stress from the rate, the degree of
deformation and the temperature proposed in1.
3 ANALYSIS OF THE RESULTS
In the analysis the expressions (6) are used to study
the stressed condition of the plastic medium in the case
of a flat upsetting of rough plates. If the problem is
reduced to a more simple mathematical model (A2
' = A3
' =
0), expression (6) will correspond to the solutions in5.
k C A= ⋅ − 
exp ( 1
' )
 
xy C A A= ⋅ − ⋅exp ( 1
' ) sin( )F
 
 x C A A f y C= ⋅ − ⋅ + + +exp ( 1 0
' ) cos( ) ( )F
 
 y C A A f x C= − ⋅ − ⋅ + + +exp ( 1 0
' ) cos( ) ( )F (14)
Applying the condition for plasticity
 0 2= − ⋅k Acos F, C k=  , the functions AF and 

become harmonic. Starting from the Laplace equation
and the Cauchy-Rieman conditions we obtain the
expressions for determining the functions in the form of
the coordinate polynomial.
A AA x y AA x y x yF = ⋅ ⋅ − ⋅ ⋅ ⋅ −6 13
2 2( )
[{
 ' . ( ) . (= − ⋅ ⋅ − − ⋅ ⋅ +05 0 256 2 2 13 4AA x y AA x
]}+ − ⋅ ⋅y x y4 2 215) .
The constants in the expressions were determined as
proceeding from the real boundary conditions
AA
l h6
0
4= ⋅
⋅
Y
, AA
l h
l h l h13 1 3
16
2
= ⋅ ⋅
− ⋅
⋅ ⋅ +
Y
( )
[ ]Y0 2 1= ⋅ ⋅ −arctg f f( )
[ ]Y1 17 1= ⋅ ⋅ −arctg . ( )f f
and the coefficient
C
k
A 

= ⋅ −0
0cos
)'
F
exp (
A AA
l h
AA
l h l h
F0 6 13
2 2
4 4 4 4
= ⋅
⋅
− ⋅
⋅
⋅ −
⎛
⎝
⎜
⎞
⎠
⎟

 

' .= − = − ⋅ ⋅ −
⎛
⎝
⎜
⎞
⎠
⎟ −
⎧
⎨
⎩
A AA
l h
0 6
2 2
05
4 4
− ⋅ +
⎛
⎝
⎜
⎞
⎠
⎟ − ⋅
⋅⎡
⎣⎢
⎤
⎦⎥
⎫
⎬
⎭
AA
l h l h
13
4 4 2 2
0 25
16 16
15
16
. .
Materiali in tehnologije / Materials and technology 44 (2010) 3, 141–145 143
V. CHYGYRYNS’KYY ET AL.: A GENERALIZED THEORY OF PLASTICITY
By substituting the components of the tensor of
stresses into (14), we obtain


  

x k
A
A k= − ⋅ ⋅ +
exp ( ' ' )
cos
cos
0
0
0F
F


  

y k
A
A k= − ⋅ ⋅ +3
0
0
0
exp ( ' ' )
cos
cos
F
F


  

xy k
A
A= ⋅ ⋅
exp ( ' ' )
cos
sin
0
0F
F (15)
The results of the calculation according to equations
(15) in Figures 1 to 4 show that the distribution of the
contact stresses is related to the factor of the shape of the
stress nucleus l/h and the friction coefficient f, the
relative normal stresses y/2k0 and the relative tangent
stresses xy/k0. The results of the calculation correspond
to the real distribution diagrams of the contact stresses.6
It should be emphasized that expressions (15) are uni-
form for the entire nucleus of deformation and there is
no need to break it into separate zones of contact fric-
tion.7 Figures 5 and 6 show the distribution of stresses
V. CHYGYRYNS’KYY ET AL.: A GENERALIZED THEORY OF PLASTICITY
144 Materiali in tehnologije / Materials and technology 44 (2010) 3, 141–145
Figure 4: Distribution of the normal stresses along the strip height
during the upsetting with rough strikers f = 0,3, l/h = 2...15
Slika 4: Porazdelitev normalnih napetosti po vi{ini plo{~e pri kr~enju
s te`kimi kladivi f = 0,3, l/h = 2...15
Figure 2: Distribution of the normal and tangent stresses at the
contact during upsetting with rough strikers f = 0,3, l/h = 1...15
Slika 2: Porazdelitev kontaktnih normalnih in tangetnih napetosti pri
kr~enju s te`kimi kladivi f = 0,3, l/h = 1...15
Figure 3: Distribution of the normal stresses along the plate height
during the upsetting with rough strikers l/h = 8, f = 0,1...0,5
Slika 3: Porazdelitev normalnih napetosti po vi{ini plo{~e pri kr~enju
s te`kimi kladivi: l/h = 8, f = 0.1…0.5
Figure 1: Distribution of the normal and tangent stresses at the
contact during upsetting with rough strikers l/h = 8, f = 0,1...0,5
Slika 1: Porazdelitev kontaktnih normalnih in tangetnih napetosti pri
kr~enju s te`kimi kladivi l/h = 8, f = 0,1...0,5
in the nucleus of deformation, which also include the
shape factor and the friction coefficient.
The obtained results display qualitatively and quanti-
tatively the general patterns of the distribution fields of
the tensor of stresses over the entire nucleus of defor-
mation. The results meet fully the requirements of the
boundary conditions. In particular, the proposed proce-
dure and expressions15 can be recommended for the
calculation of various problems in applications.
The proposed complex model of a plastic medium
based on the closed solution can be considered as a
generalization of the theory of plasticity, uniting the
theories of deformation and of plastic yielding.
4 REFERENCES
1 V. V. Chygyryns’kyy, I. Mamuzi}, G. V. Bergeman: Analysis of the
State of Stress of a Medium under Conditions of Inhomogeneous
Plastic Yielding; Metalurgija. Zagreb. 43 (2004), 87–93
2 A. N. Tikhonov, A. A. Samarsky: Uravnenia matematicheskoy fisiki
(Equations of mathematical physics); Nauka, (1977), 735
3 N. N. Malinin: Prikladnaya teoria plastichnosti i polsychesti
(Applied theory of plasticity and yielding); Mashinostroyeniye,
(1975), 399
4 L. M. Kachanov: Osnovy teorii plastichnosti (Fundamentals of the
theory of plasticity); Nauka, (1969), 419
5 G. E. Arkulis, V. G. Dorogobid: Teoria plastichnosti (Theory of
plasticity); Metalurgiya, (1987), 251
6 A. P. Chekmarev, P. L. Klimenko: Experimentalnoe issledovanie
udel’nyh davleniy na kontaktnoy poverhnosti pri prokatke v kalibrah
(Experimental investigations of partial pressures on the contact
surface in rolling in calibers); Obrabotka metallov davleniem
(Pressure forming of metals); Sbornik trudov Dnepropetrovskogo
metallurgicheskogo instituta; Khar’kov, M., 1960. – Vypusk 39
7 M. V. Storozhev, E. A. Popov: Teoria obrabotki metallov davleniem
(Theory of pressure working of metals); Mashinostroyeniye, (1977),
422
LIST OF SYMBOLS
 – normal components of the stress tensor;
 – tangential components of the stress tensor;
 – linear components of the strain-rate tensor;

 – shear components of the strain-rate tensor;
 – linear deformation along the axes x and y;
 – factor of correspondence between the tangent stress and the
temperature-deformation parameter of the centre to deformation;
Gi – intensity of the shift of the deformation;
n – tangential contact stress on an arbitrary inclined area;
 – angle of inclination of the contact area;
k – shearing plastic deformation strength;
F – harmonic function depending on the coordinates of the deforma-
tion zone and the argument of a trigonometric function;
F0 – argument of trigonometric function for x = ±l/2 and y = ±h/2;
A – constant characterizing the trigonometric function for the state of
stress of the plastic medium;
A6, A13 – constant factor, characterizing the shearing tangent stress in
the zone of reduction
Y0, Y1 – values taking into account the influence of the factor of fric-
tion;
B – constant value characterizing the trigonometric function for the
state of strain of the plastic medium;

0 – factor exhibitors for x = ±l/2 and y = ±h/2;

 ' – harmonic function, exponential index, characterizing the shearing
stress distribution in the zone of reduction;

 ' ' – a harmonic function, exponential index, characterizing the distri-
bution of the rate of shearing in the zone of reduction;
C – constant value determining the state of stress of the plastic me-
dium;
C – constant value characterizing the state of strain of the plastic me-
dium;
T – temperature in the k-th point;
Ti – intensity of tangential stress;
Hi – intensity of shearing rates;
CT – constant value characterizing the temperature field;
a – coefficient of temperature conductivity;
l and h – length and height of the deformation nucleus during strip up-
setting;
f – friction coefficient;
k0 – contact shear resistance at the beginning of the deformation nu-
cleus
Materiali in tehnologije / Materials and technology 44 (2010) 3, 141–145 145
V. CHYGYRYNS’KYY ET AL.: A GENERALIZED THEORY OF PLASTICITY

O. GUVEN ET AL.: COMPARATIVE MODELING OF WIRE ELECTRICAL DISCHARGE MACHINING ...
COMPARATIVE MODELING OF WIRE ELECTRICAL
DISCHARGE MACHINING (WEDM) PROCESS USING
BACK PROPAGATION (BPN) AND GENERAL
REGRESSION NEURAL NETWORKS (GRNN)
PRIMERJALNO MODELIRANJE ELEKTROEROZIJSKE @I^NE
OBDELAVE (WEDM) Z UPORABO POVRATNOSTI (BPN) IN
SPO[NE NEVRONSKE REGRESIJSKE MRE@E (GRNN)
Onur Guven1, Ugur Esme2*, Iren Ersoz Kaya3, Yigit Kazancoglu4, M. Kemal
Kulekci5, Cem Boga6
1Mersin University, Engineering Faculty, Department of Mechanical Engineering, 33400, Mersin/Turkey
2,3,5Mersin University, Tarsus Technical Education Faculty Department of Mechanical Education, 33480, Tarsus-Mersin/Turkey
4Izmir University of Economics, Department of Business Administration, 35330, Balcova-Izmir/Turkey
6Cukurova University, Vocational School of Karaisali, 01770, Karaisali-Adana/Turkey
uguresme@gmail.com
Prejem rokopisa – received: 2010-01-22; sprejem za objavo – accepted for publication: 2010-03-02
The use of two neural networks techniques to model wire electrical discharge machining process (WEDM) is explored in this
paper. Both the back-propagation (BPN) and General Regression Neural Networks (GRNN) are used to determine and compare
the WEDM parameters with the features of the surface roughness. A comparison between the back-propagation and general
regression neural networks in the modeling of the WEDM process is given. It is shown that both the back-propagation and
general regression neural networks can model the WEDM process with reasonable accuracy. However, back propagation neural
network has better learning ability for the wire electrical discharge machining process than the general regression neural
network. Also, the back-propagation network has better generalization ability for the wire electrical discharge machining
process than does the general regression neural network.
Keywords: WEDM, neural network, modeling, BPN, GRNN
Raziskana je uporaba dveh nevronskih mre` za modeliranje elekroerozijske `i~ne obdelave (WEDM). Obe metodi: povratnostna
(BPN) in splo{na regresijska nevronska mre`a (GRNN), sta uporabljeni za dolo~itev in primerjavo VEDM-procesa. Dokazano
je, da sta obe metodi primerni za modeliranje WEDM s sprejemljivo natan~nostjo. Vendar pa ima povratnostna nevronska mre`a
bolj{o sposobnost u~enja in bolj{o sposobnost posplo{enja procesa kot splo{na regresijska nevronska mre`a.
Klju~ne besede: WEDM, nevronska mre`a, modeliranje, BPN, GRNN
1 INTRODUCTION
Manufacturing industry is becoming ever more
time-conscious with regard to the global economy, and
the need for rapid prototyping and small production
batches is increasing. These trends have placed a pre-
mium on the use of new and advanced technologies for
quickly turning raw materials into usable goods; with no
time being required for tooling.1 Wire electrical
discharge machining (WEDM) technology has been
found to be one of the most recent developed advanced
non-traditional methods used in industry for material
processing with the distinct advantages of no thermal
distortion, high machining versatility, high flexibility,
rapid machining and high accuracy of complex parts.2
The degree of accuracy of workpiece dimensions obtain-
able and the fine surface finishes make WEDM particu-
larly valuable for applications involving manufacture of
stamping dies, extrusion dies and prototype parts.
Without WEDM the fabrication of precision workpieces
requires many hours of manual grinding and polishing.3–6
The most important performance measures in
WEDM are cutting speed, workpiece surface roughness
and cutting width. Discharge current, discharge capaci-
tance, pulse duration, pulse frequency, wire speed, wire
tension, average working voltage and dielectric flushing
conditions are the machining parameters which affect the
performance measures.3–7
Tosun et al.3 determined the effect of machining
parameters on the cutting width and material removal
rate based on the Taguchi method. Tosun and Cogun4
investigated experimentally the effect of cutting
parameters on wire electrode wear. Tosun et al.5
investigated the effect of the cutting parameters on size
of erosion craters (diameter and depth) on wire electrode
experimentally and theoretically. Cogun and Savsar6
investigated the random behaviour of the time-lag
durations of discharge pulses using a statistical model for
different pulse durations, pulse pause durations, and
discharge currents in EDM.
Esme et al.7 modeled the surface roughness in
WEDM process using design of experiments and neural
networks. Scott et al.8 have developed formulas for the
solution of a multi-objective optimization problem to
select the best parameter settings on a WEDM machine.
Materiali in tehnologije / Materials and technology 44 (2010) 3, 147–152 147
UDK 669.14.018:519.68 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 44(3)147(2010)
They used a factorial design model to predict the
measures of performances as a function of a variety of
machining parameters. Wang and Rajurkar9 have
developed a WEDM frequency monitoring system to
detect on-line the thermal load on the wire to prevent the
wire from rupture. Spur and Schoenbeck10 have investi-
gated a finite element model and they have explained the
impact of a discharge on the anode as a heat source on a
semi-infinite solid whose size and intensity are time-
dependent in WEDM. Tarng et al.11 developed a neural
network system to determine settings of pulse duration,
pulse interval, peak current, open circuit voltage, servo
reference voltage, electric capacitance and wire speed for
the estimation of cutting speed and surface finish.
Spedding and Wang12 presented parametric combination
by using artificial neural networks and they also
characterized the roughness and waviness of workpiece
surface and the cutting speed. Liao et al.13 performed an
experimental study to determine the variation of the
machining parameters on the MRR, gap width and
surface roughness. They have determined the level of
importance of the machining parameters on the metal
removal rate (MRR). Lok and Lee14 compared the
machining performance in terms of MRR and surface
finish by the processing of two advanced ceramics under
different cutting conditions using WEDM. Rama-
krishnan and Karunamoorthy15 developed an artificial
neural network with Taguchi parameter design. Tsai et
al.16 relationships between the heterogeneous second
phase and the machinability evaluation of the ferritic SG
cast irons in the WEDM process. Sarkar et al.17 studied
on the features of trim cutting operation of wire elec-
trical discharge machining of 
-titanium aluminide.
Caydas et al.18 developed an adaptive neuro-fuzzy
inference system (ANFIS) for modeling the surface
roughness in WEDM process.
As indicated in the previous studies, most of the
research works are focused on the effect of machining
parameters, discharge energy, theory and experimental
verification crater formation on the wire electrode. The
present study focused on the comparative modeling and
prediction of surface roughness to compare the techni-
ques of back propagation network (BPN) and general
regression neural network (GRNN).
2 EXPERIMENTAL DETAILS
As shown in Figure 1, the experimental studies were
performed on an Acutex WEDM machine tool. Different
settings of pulse duration (t), open circuit voltage (V),
wire speed (S) and dielectric flushing pressure (p) were
used in the experiments. Table feed rate (8.2 mm/min),
pulse interval time (18 µs), and wire tension (1800 g) are
kept constant during the experiments7.
AISI 4340 steel plate was used as a workpiece
material with (150 × 150 × 10) mm dimensions. CuZn37
Suncut brass wire with 0.25 mm diameter and tensile
strength of 900 N/mm2 was used in the experiments.
Workpiece average surface roughness (Ra) measurements
were made by using Phynix TR–100 portable surface
roughness tester. Cut-off length () and traversing length
(l) were set as 0.3 mm and 5 mm, respectively. Pulse
duration, open circuit voltage, wire speed and dielectric
flushing pressure were selected as input parameters and
surface roughness (Ra) was selected as an output para-
meter7.
Four measurements were made and their average was
taken as Ra value for a machined work surface. After
collecting the experimental results both techniques
namely back propagation neural network (BPN) and
general regression neural network (GRNN) techniques
were carried out to predict surface roughness (Ra).
3 ARTIFICIAL NEURAL NETWORKS (ANN)
It is well known that modeling the relationships
between the input and output variables for non-linear,
coupled, multi-variable systems is very difficult. In
recent years, neural networks have demonstrated great
O. GUVEN ET AL.: COMPARATIVE MODELING OF WIRE ELECTRICAL DISCHARGE MACHINING ...
148 Materiali in tehnologije / Materials and technology 44 (2010) 3, 147–152
Figure 1: Acutex WEDM used in the experiments7
Slika 1: Acutex WEMM, uporabljen za preizkuse7
Figure 2: A non-linear, coupled, and multi-variable system19
Slika 2: Nelinearen, povezan in multivariabilni sistem19
potential in the modeling of the input–output relation-
ships of complicated systems.19,20 Consider that X = {x1,
x2…, xm} is the input vector of the system where m is the
number of input variables and Y = {y1, y2…, yn} is the
corresponding output vector of the system where n is the
number of output variables19 as shown in Figure 2. In
this section, the use of back-propagation and general
regression networks to construct the relationships
between the input vector X and output vector Y of the
system will be explored.
3.1 Back-Propagation Networks (BPN)
The back-propagation network is composed of many
interconnected neurons that are often grouped into input,
hidden and output layers. The neurons of the input layer
are used to receive the input vector X of the system and
the neurons of the output layer are used to generate the
corresponding output vector Y of the system. The
back-propagation network used in this study is shown in
Figure 3. For each neuron a summation function for all
the weighted inputs are calculated as:
net j
k k
j
i
kw o= µ
−∑ 1 (1)
where netkj is the summation function for all the inputs
of the j-th neuron in the k-th layer, wkji is the weight
from the i-th neuron to the j-th neuron and oik–1 is the
output of the i-th neuron in the (k–1)-th layer.
Setting 5-hidden layers resulted in lowest error
between predicted and experimental results. Therefore,
in the present work, 4-inputs, 5-hidden layer, 1 output
layer (4 : 5 : 1 model) back propagation neural network
has been used. The used BPN algorithm is shown in
Figure 3.
As shown in Eq. (1), the neuron evaluates the inputs
and determines the strength of each one through its
weighting factor, i.e. the larger the weight between two
neurons, the stronger is the influence of the connection.19
The result of the summation function can be treated as an
input to an activation function from which the output of
the neuron is determined. The output of the neuron is
then transmitted along the weighted outgoing connec-
tions to serve as an input to subsequent neurons. To
modify the connection weights properly, a supervised
learning algorithm involving two phases is employed.21
The first is the forward phase which occurs when an
input vector X is presented and propagated forward
through the network to compute an output for each
neuron.19,20 Hence, an error between the desired output yj
and actual output oj of the neural network is computed.19
The summation of the square of the error E can be
expressed as:
( )E y oj j
j
n
= −
−
∑1
2
2
1
(2)
The second is the backward phase which is an
iterative error reduction performed in a backward direc-
tion. To minimize the error between the desired and
actual outputs of the neural network as rapidly as possi-
ble, the gradient descent method, adding a momentum
term,21 is used. The new incremental change of weight
wkji(n + 1) can be expressed as:
∆ ∆w n
E
w
w nji
k
ji
k ji
k( ) ( )+ = − +1  
∂
∂
(3)
where  is the learning rate,  is the momentum coeffi-
cient and n is the index of iteration. Through this
learning process, the network memorizes the relation-
ships between input vector X and output vector Y of the
system through the connection weights.19–21
4 GENERAL REGRESSION NEURAL
NETWORKS (GRNN)
The General Regression Neural Networks (GRNN)
introduced by Donald Specht in 1990 is a memory-based
feed forward neural network based on the approximate
estimation of the probability density function from
observed samples using Parzen-window estimation.22 It
approximates any arbitrary function between input and
output vectors. This approach removes the necessity to
specify a functional form of estimation.
The method utilizes a probabilistic model between an
independent random vector X (input) and a dependent
scalar random variable Y (output). Let x and y be the
particular measured values of X and Y, respectively, and

( , )g x y is the joint continuous probability density function
of X and Y. A good choice for a non-parametric estimate
of the probability density function g is the Parzen
window estimator as proposed by Parzen and performed
for multidimensional cases by Cacoullos.22–26 Given a
sample of n real D dimensional xi vectors and corres-
ponding scalar yi values, the estimate of joint probability
density in GRNN is given by;

( , )
( )
exp
( ) ( )
( ) / ( )
g x y
n
x x x x
d d
i
T
i= − −
⎛
⎝
⎜
+ +
1
2
1
21 2 1 2π  ⎜
⎞
⎠
⎟⎟
−⎛
⎝
⎜⎜
⎞
⎠
⎟⎟
⎡
⎣
⎢
⎤
⎦
⎥
=
∑ exp ( )y yi
i
n 2
2
1 2
(4)
O. GUVEN ET AL.: COMPARATIVE MODELING OF WIRE ELECTRICAL DISCHARGE MACHINING ...
Materiali in tehnologije / Materials and technology 44 (2010) 3, 147–152 149
Figure 3: BPN network used for modeling
Slika 3: BPN mre`a, uporabljena za modeliranje
where 
 is the window width of a sample probability,
called the smoothing factor of the kernel24. The
expected value of Y given x (the regression of Y on x) is
given by;
[ ]E Y/x
Y g x Y Y
g x Y Y
=
⋅
−∞
+∞
−∞
+∞
∫
∫
( , )
( , )
d
d
(5)
Using Eq. (4), Eq. (5) becomes;
[ ]
[ ]

( )
exp( )
exp( )
y x E Y/x
y d
d
i i
i
n
i
i
n= =
=
=
∑
∑
1
1
(6)
where di is the distance between the input vector and the
ith training vector, and is given by;
d
x x x x
i
i
T
i2
22
= −
− −( ) ( )

(7)
The estimate 
y(x) is thus a weighted average of all the
observed yi values where each weight is exponentially
proportional to its Euclidean distance from x.
As shown in Figure 4, the structure of the GRNN
consists of 4 layers; the input layer, the hidden (pattern)
layer, the summation layer and the output layer. As a
preprocessing step, all input variables of the training data
are scaled. Then, they are copied as the weights into the
pattern units.
As a preprocessing step, all input variables of the
training data are scaled. Then, they are copied as the
weights into the pattern units. The summation layer has
two units that can be denoted as the numerator and the
denominator of Eq. (6). The output layer gives the
estimate of the expected value of 
y(x). If y and 
y are the
vector variables, the results above are generalized by
adding with one summation unit for each component of y
in the output layer.
The only adjustable parameter of the network is 
,
the smoothing factor for the kernel function. It is critical
to decide an optimum value for 
. The larger values of
this factor cause the density to be smooth, and 
y(x) then
converges to the sample mean of the observed yi. On the
other hand, when 
 is chosen very small, the density is
forced to have non-gaussian shapes. Then, the oscillatory
points have a bad effect on the estimate. All values of yi
are taken into account where the points closer to x are
given heavier weights, if the optimum value of 
 is
selected.24–26 Therefore, in this study, 
 was chosen as
0.57 due to the optimum value of success rate that was
found after iterative calculation of 
 values between 0.1
and 0.9.
The main advantage of GRNN according to other
techniques is fast learning. It is a one-pass training
algorithm, and does not require an iterative process. The
training time is just the loading time of the training
matrix. Also, it can handle both linear and non-linear
data because the regression surface is instantly described
everywhere, even just one sample is enough for this.
Thus, other existing pattern nodes tolerate faulty
samples. Another advantage is the fact that adding new
samples to the training set does not require re-calibration
of the model. As the sample size increases, the estimate
surface converges to the optimal regression surface.
Thus, it requires many training samples to span the
variation in the data and all these to be stored for the
future use. Solely, this causes a trouble of an increase in
the amount of the computation to evaluate new points. In
the course of time, highly improvements in the speed of
the computer’s processing power prevent this being a
major problem. Furthermore, this also can be overcome
by applying the various clustering techniques for
grouping samples that each center is represented by this
group of samples.22–26 However, there is only one dis-
advantage that there is no intuitive method for choosing
the optimal smoothing factor.
5 RESULTS AND DISCUSSIONS
In this study, twenty-eight set of data under different
process condition was used for training and testing of the
BPN and GRNN. Sixteen of them were used as a training
purpose and the rest were used as testing purposes.
Table 1 shows the design matrix and training set used for
BPN and GRNN analysis.
Testing the validation of BPN and GRNN results was
made using the input parameters according to the design
matrix given in Table 2.
These comparisons have been depicted in terms of
percentage error in Figure 5 for validation set of experi-
ments. From Table 2 it is evident that for our set of data
the BPN result predicts the surface roughness nearer to
the experimental values than the GRNN results. But,
O. GUVEN ET AL.: COMPARATIVE MODELING OF WIRE ELECTRICAL DISCHARGE MACHINING ...
150 Materiali in tehnologije / Materials and technology 44 (2010) 3, 147–152
Figure 4: Constructed GRNN network
Slika 4: Zgrajena GRNN-mre`a
O. GUVEN ET AL.: COMPARATIVE MODELING OF WIRE ELECTRICAL DISCHARGE MACHINING ...
Materiali in tehnologije / Materials and technology 44 (2010) 3, 147–152 151
Figure 6: Comparison of predicted and experimental results
Slika 6: Primerjava napovedanih in eksperimentalnih rezultatov
Figure 5: BPN and GRNN errors in prediction of the surface rough-
ness
Slika 5: BPN- in GRNN-napake pri napovedi hrapavosti povr{ine
Table 1: Neural network training set7
Tabela 1: Podatki za trening nevronske mre`e7
WEDM input Parameters Output
Exp. no t/ns V/V S/(mm/min) p/(kg/cm2) Ra/µm
1 200 300 12 16 2.12
2 200 60 4 16 1.13
3 900 60 4 6 2.14
4 200 60 12 16 1.24
5 200 300 12 6 2.32
6 200 300 4 16 1.98
7 900 60 12 16 2.15
8 900 300 12 6 3.85
9 200 300 4 6 2.10
10 900 300 4 16 3.24
11 900 60 12 6 2.26
12 900 300 12 16 3.65
13 900 60 4 16 2.01
14 200 60 4 6 1.18
15 900 300 4 6 3.55
16 200 60 12 6 1.24
Table 2: Test and comparison of BPN and GRNN results
Tabela 2: Preizkusi in primerjava BPN- in GRNN-rezultatov
Modeling
Exp. No
WEDM input parameters Back PropagationNeural Network (BPN)
General Regression
Neural Network (GRNN)
t/ns V/V S/(mm/min) p/(kg/cm2) Ra/µm predicted error predicted error
1 300 80 4 6 1.30 1.26 3.08 1.38 -6.15
2 400 90 5 8 1.50 1.36 9.33 1.55 -3.33
3 500 150 6 10 2.08 2.02 2.88 1.96 5.77
4 700 250 10 14 3.18 3.48 -9.43 3.22 -1.26
5 350 60 12 16 1.29 1.24 3.88 1.43 -10.85
6 450 70 5 16 1.58 1.53 3.16 1.50 5.06
7 550 100 8 11 2.08 2.02 2.88 1.88 9.62
8 750 180 4 6 2.92 3.11 -6.51 2.72 6.85
9 850 200 10 8 3.27 3.47 -6.12 3.14 3.98
10 200 300 12 8 2.23 2.37 -6.28 2.31 -3.59
11 250 300 4 10 1.96 2.00 -2.04 2.14 -9.18
12 300 250 6 20 1.89 1.81 4.23 2.02 -6.88
Average error: 4.99%
CPU time = 2.3 min
Average error: 6.04%
CPU time = 0.074 sec
GRNN is much faster than BPN with the CPU times of
0.074 s and 2.3 min respectively. In the prediction of
surface roughness values the average errors for BPN and
GRNN are calculated as 4.99 % and 6.04 % respectively.
The value of the multiple coefficient of R2 is obtained
as 0.99 for BPN and 0.96 for GRNN which means that
the fitted line is very close to the experimental results.
Figure 6 represents the comparison of predicted (both
BPN and GRNN) and actual results. Both BPN and
GRNN results showed that the predicted values have
been very close to experimental values.
6 CONCLUSIONS
The prediction of optimal machining conditions for
the required surface finish and dimensional accuracy
plays a very important role in the process planning of the
wire erosion discharge machining process. This paper
has described a neural network approach and comparison
of Back Propagation Networks (BPN) and General
Regression Neural Networks (GRNN) networks for the
modeling of wire electrical discharge machining process
using small set of data. Both the BPN and GRNN
networks were used to construct the complicated
relationships between the process parameters and the
surface roughness. The experimental results has showed
that the BPN network has better learning ability (with
average error of 4.99 % and multiple coefficient of R2 of
99 %) for the wire electrical discharge machining
process than the GRNN (with average error of 6.04 %
and multiple coefficient of R2 of 0.96) network. Training
of BPN network consumed more CPU time (elapsed
time 2.3 min) than the GRNN (elapsed time 0.074 s). In
addition to this, the back propagation network has better
generalization ability for the wire electrical discharge
machining process than the general regression neural
network modeling.
7 REFERENCES
1 J. Wang, , 19 (2003), 114
2 U. Caydas, A. Hascalik, Journal of Materials Processing Technology,
202 (2008), 574–582
3 N. Tosun, C. Cogun, G. Tosun, Journal of Materials Processing
Technology, 152 (2004), 316–322
4 N. Tosun, C. Cogun, H. Pihtili, Int. J. Adv. Manuf. Technol., 21
(2003), 857–865
5 N. Tosun, C. Cogun, A. Inan, Machining Science and Technology, 7
(2003), 209–219
6 C. Cogun, M. Savsar, International Journal of Machine Tools and
Manufacture, 3 (1990), 467–474
7 U. Esme, A. Sagbas, F. Kahraman, Iranian Journal of Science &
Technology, Transaction B, Engineering, 33 (2009), 231–240
8 D. Scott, S. Boyina, K. P. Rajurkar, Int. J. Prod. Res., 11 (1991),
2189–2207
9 W. M. Wang, K. P. Rajurkar, Monitoring sparking frequency and
predicting wire breakage in WEDM, Sensors and Signal Processing
for Manufacturing, ASME, Production Engineering Division (PED),
New York, 55 (1992), 49–64
10 G. Spur, J. Schoenbeck, Anode erosion in wire-EDM-A theoretical
model. CIRP Ann., 1 (1993), 253–256
11 Y. S. Tarng, S. C. Ma, L. K. Chung, International Journal of Machine
Tools and Manufacture, 35 (1995), 1693–1701
12 T. A. Spedding, Z. Q. Wang, Precis. Eng., 20 (1997), 5–15
13 Y. S. Liao, J. T. Huang, H. C. Su, Journal of Materials Processing
Technology, 71 (1997), 487–493
14 Y. K. Lok, T. C. Lee, Journal of Materials Processing Technology, 63
(1997), 839–843
15 R. Ramakrishnana, L. Karunamoorthyb, Journal of Materials
Processing Technology, 207(2008), 343–349
16 T. C Tsai, J.T. Horng, N.M. Liu, C.C. Chou, K.T. Chiang, Materials
and Design, 29 (2008), 1762–1767
17 S. Sarkar, S. Mitra, B. Bhattacharyya, Journal of Materials
Processing Technology, 205(2008), 376–387
18 U. Caydas, A. Hascalik, S. Ekici, doi:10.1016/j.eswa.2008.07.019,
(2008)
19 S. C. Juang, Y. S. Tarng, H. R. Lii, Journal of Materials Processing
Technology, 75 (1998), 54–62
20 J. A. Freeman, D. M. Skapura, Neural Networks: Algorithms, Appli-
cation and Programming Techniques, Addison–Wesley, New York,
(1991)
21 J. McClelland, D. Rumelhart, Parallel Distributed Processing, MIT
Press, Cambridge, MA, 1 (1986)
22 D. Specht, A General Regression Neural Network. IEEE Trans.
Neural Networks, 2 (1991), 568–576
23 T. Cacoullos, Estimation of a multivariate density. Annals of Inst.
Stat. Math., 18 (1996), 178–189
24 D. Yeung, C. Parzen, Window Network Intrusion Detectors, Pro-
ceedings of the Sixteenth International Conference on Pattern
Recognition, Quebec City, Canada, 4 (2002), 385–388
25 I. Ersoz Kaya, T. Ibrikci, A. Cakmak, Journal of Computational
Intelligence in Bioinformatics, 1 (2008), 1–11
26 U. Esme, I. Ersoz, A. Ozbek, F. Kahraman, A. Sagbas, Cukurova
University Journal of Engineering and Architecture, 23 (2008),
57–66
O. GUVEN ET AL.: COMPARATIVE MODELING OF WIRE ELECTRICAL DISCHARGE MACHINING ...
152 Materiali in tehnologije / Materials and technology 44 (2010) 3, 147–152
K. ELER[I^ ET AL.: DEGRADATION OF BACTERIA ESCHERICHIA COLI ...
DEGRADATION OF BACTERIA ESCHERICHIA COLI BY
TREATMENT WITH Ar ION BEAM AND NEUTRAL
OXYGEN ATOMS
UNI^EVANJE BAKTERIJ ESCHERICHIA COLI S CURKOM IONOV
Ar IN NEVTRALNIH ATOMOV KISIKA
Kristina Eler{i~1, Ita Junkar1, Ale{ [pes1, Nina Hauptman2,
Marta Klanj{ek-Gunde2, Alenka Vesel1*
1Jozef Stefan Institute, Jamova cesta 39, 1000 Ljubljana, Slovenia
2National Institute of Chemistry, Hajdrihova 19, 1000 Ljubljana, Slovenia
alenka.vesel@ijs.si
Prejem rokopisa – received: 2009-06-18; sprejem za objavo – accepted for publication: 2010-03-12
Scanning electron microscopy was used to determine the difference between bacteria degradation by two types of particles
presented in gaseous plasma, i.e. positively charged ions and neutral oxygen atoms. The source of ions was an argon ion gun
with the ion energy of 1 keV and the flux of 3 × 1018 m–2 s–1. The source of neutral oxygen atoms was inductively coupled
oxygen plasma supplying the flux of oxygen atoms of about 1.5 × 1023 m–2 s–1. The ion beam treatment time was 1800 s while
the oxygen atom treatment time was 300 s. Bacteria Escherichia coli, strain ATCC 25922 were deposited onto well activated
aluminum at the concentration of about 3 × 106 cfu and exposed to both particles. SEM analysis was performed using a field
emission microscope with the energy of primary electrons of 1 keV. SEM images revealed huge difference in morphology of
bacteria treated by both methods. While ions tend to drill holes into bacterial cell wall, the atoms caused a more even disruption
of bacterial cell wall. The results were explained by kinetic, potential and charging effects.
Key words: bacteria, Escherichia coli, sterilization, degradation, oxygen plasma, atoms, ions, SEM
Z vrsti~no elektronsko mikroskopijo smo raziskovali razliko v degradaciji bakterij pri obdelavi z dvema razli~nima vrstama
delcev v plinski plazmi: s pozitivno nabitimi in z nevtralnimi kisikovimi atomi. Vir ionov argona z energijo 1 keV in tokom 3 ×
1018 m–2 s–1 je bila ionska pu{ka. Vir nevtralnih atomov kisika s tokom 1,5 × 1023 m–2 s–1 na povr{ino vzorcev pa je bila
induktivno sklopljena kisikova plazma. ^as obdelave z ioni je bil 3000 s, medtem ko je bil ~as obdelave s kisikovimi atomi 300
s. Bakterije Escherichia coli, sev ATCC 25922 smo nanesli na dobro aktivirano povr{ino aluminija in jih potem izpostavili
curkom obeh vrst delcev. Koncentracija bakterij je bila 3 × 106 cfu. Po obdelavi smo povr{ino vzorcev analizirali z vrsti~no
elektronsko mikroskopijo (SEM). SEM-slike so razkrile veliko razliko v morfologiji bakterij, obdelanih z atomi oziroma ioni.
Medtem ko ioni povzro~ijo nastanek lukenj v celi~ni steni bakterij, pa atomi bolj enakomerno degradacijo celi~ne stene.
Dobljene rezultate smo razlo`ili z vplivom kineti~nih in potencialnih efektov ter vplivom nabijanja povr{ine.
Klju~ne besede: bakterije, Escherichia coli, sterilizacija, degradacija, kisikova plazma, atomi, ioni, SEM
1 INTRODUCTION
Plasma sterilization has attracted much attention in
the past decade due to possible application for steriliza-
tion of delicate materials that cannot stand autoclaving in
humid air at 130 °C. Several different types of discharges
have been used to create plasma suitable for destruction
of vital bacteria and their spores.1–9 The discharges in-
clude low and atmospheric pressure. Among atmospheric
discharges, RF and microwave plasma torches are partic-
ularly popular, while the dielectric barrier glow dis-
charge was not found as efficient. The same applies also
for otherwise popular corona discharges. The low pres-
sure discharges suitable for destruction of bacteria at low
temperature include the DC, RF and microwave dis-
charges.10–14 Radiofrequency discharges are particularly
popular since they assure for a high density of plasma
radicals and rather low kinetic temperature of neutral
gas.
Most authors presented results on bacterial
deactivation as a function of discharge parameters. The
discharge parameters that are often varied include the
type of gas or gas mixture, the pressure in the discharge
tube and the gas flow, the discharge power, the
dimensions and the type of material used for the
discharge chamber, etc. Much less work, however, has
been done on determination of sterilization effects versus
plasma parameters. Not surprisingly, the explanations of
observed sterilization effects are often contradictory.
Many authors explain sterilization by destruction of
bacterial DNA caused by UV photons from plasma.
Other authors state that sterilization is due to chemical
etching of the bacterial cell wall with radicals such as O,
N, H, etc. Some other authors take into account also the
kinetic effects of bombardment with positive ions, and
most authors agree that synergetic effects play an
important role.
In order to understand the role of different plasma
particles it is the best to separate them and treat bacteria
only with one type plasma particles. At the experiments
presented in this paper we exposed bacteria separately to
2 types of different plasma particles: energetic non-reac-
tive ions and neutral oxygen atoms with the kinetic tem-
perature of 300 K.
Materiali in tehnologije / Materials and technology 44 (2010) 3, 153–156 153
UDK 543.428.2:579.24 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 44(3)153(2010)
2 EXPERIMENTAL
2.1 Sample preparation
Bacteria Escherichia coli (E. coli) were cultivated ac-
cording to the standard procedure. In experiment we
used bacteria E. coli strain ATCC 25922. It was grown at
37 °C, on LB plates for 24 h. Cells were then resuspend-
ed in sterile water. Number of cells was adjusted to ap-
proximately 3 × 106 cfu (colony forming unites).
Live bacteria were deposited onto commercially
available aluminum foils. Substrates were first carefully
cleaned with wet chemical treatment, and then activated
with a brief exposure to oxygen plasma in order to assure
the removal of any traces of organic contaminants and
achieve optimal hydrophilicity. A drop of water contain-
ing vital bacteria was placed onto the substrate. Due to
highly activated surface, the bacteria-containing water
drop was spread on a large surface. Such spreading al-
lowed for two dimensional distributions of bacteria with
out overlapping.
2.2 Experimental system
Samples were treated either by neutral oxygen atoms
in an afterglow chamber of oxygen plasma reactor or by
positively charged Ar ions from a commercial ion gun.
The schematic of the experimental setup for the case of
oxygen atoms is shown in Figure 1. The vacuum system
is pumped with a two stage rotary pump. The effective
pumping speed at the exit of the experimental chamber is
almost identical to the nominal pumping speed of the
pump, i.e.16 m3/h. The experimental chamber is
connected to a discharge chamber through a narrow tube
that allows for a difference in the effective pumping
speeds between the experimental and discharge
chambers and thus a pretty high drift velocity of gas
through the narrow tube. Both chambers as well as the
connection tube are made from borosilicate glass Schott
8250. This glass has a low recombination coefficient for
the reaction O + O  O2.15,16 Such a configuration
assures for experiments at constant (i.e. room) tempera-
ture and constant density of oxygen atoms in the vicinity
of substrates. The density of neutral oxygen atoms is
measured with a catalytic probe.17-19 At the experimental
pressure of 75 Pa the O density is about 1 × 1021 m–3. The
resultant flux of neutral oxygen atoms onto the surface
of the sample is then j = ¼ nv = 1.5 × 1023 m–2 s–1.
The experimental setup for treatment of bacteria with
Ar ions is shown schematically in Figure 2. The source
of Ar ions is a commercial ion gun used for sputtering of
materials during depth profiling. Ar ion beam with the
energy of 1 keV at an incidence angle of 45° and a raster
of 3 mm × 3 mm was used for treating bacteria. The ion
current is 0.15 A/m2 giving the ion flux onto the surface
of the substrate with bacteria of 3 × 1018 m–2 s–1. We used
no charge compensation during treatment of bacteria
with argon ions.
2.3 SEM imaging
Scanning electron micrographs of substrates with
bacteria were obtained using a field emission microscope
Karl Zeiss Supra 35 VP. A 1 kV accelerating voltage was
used to record images.
3 RESULTS
SEM image of untreated E. coli bacteria is shown in
Figure 3. The image does not look very sharp. This is
not an artifact of the microscope but rather the conse-
quence of the presence of the capsule on the surface of
bacteria as well as between bacteria. Namely, the capsule
is composed predominantly of chemically bonded water
as well as some sugars, proteins and lipids – material
that are a bad scatterer for electrons. That’s why the
SEM image looks rather dim.
A SEM image of a bacteria treated by Ar ions is
shown in Figure 4. The bacteria are badly damaged and
definitely not capable of revitalization.
K. ELER[I^ ET AL.: DEGRADATION OF BACTERIA ESCHERICHIA COLI ...
154 Materiali in tehnologije / Materials and technology 44 (2010) 3, 153–156
Figure 2: The experimental setup for treatment of bacteria with Ar
ions: 1 – UHV chamber, 2 – pumping system, 3 – vacuum gauge, 4 –
sample, 5 – ion gun, 6 – energetic ions.
Slika 2: Shema eksperimentalnega sistema za obdelavo bakterij z ioni
Ar: 1 – UVV komora, 2 – ~rpalni sistem, 3 – vakuummeter, 4 –
vzorec, 5 – ionska pu{ka, 6 – energijski ioni
Figure 1: The experimental setup for treatment of bacteria with
neutral oxygen atoms. 1 – vacuum pump, 2 – experimental chamber, 3
– discharge chamber, 4 – sample, 5 – vacuum gauge, 6 – catalytic
probe, 7 – inlet valve, 8 – oxygen flask
Slika 1: Shema eksperimentalnega sistema za obdelavo bakterij z
nevtralnimi atomi kisika: 1 – vakuumska ~rpalka, 2 – eksperimentalna
komora, 3 – razelektritvena komora, 4 – vzorec, 5 – vakuummeter, 6 –
kataliti~na sonda, 7 – dozirni ventil, 8 – jeklenka s kisikom
A SEM image of bacteria treated in the afterglow of
the oxygen plasma, i.e. with neutral oxygen atoms only,
is presented in Figure 5. In this case, the surface mor-
phology is very different from that observed in Figure 4.
4 DISCUSSION
Figures 3, 4 and 5 represent SEM images of bacteria
E. coli. Bacteria presented in Figure 3 are live what has
been confirmed by cultivation using the standard plate
count technique. Bacteria are covered with a thin film of
jelly of lipopolysaccharides and is called capsule. The
majority of lipopolysaccharide cover material has
chemically bonded water. This thin cover is (about 400
nm or more) capsular polysaccharide gel20 which serves
as a medium for gluing bacteria together as well as for
sticking onto surfaces. The capsule also facilitates
formation of three dimensional clusters of bacteria. Such
clustering was not observed at our experiments since we
activated the surface of the aluminum prior to bacterial
deposition. The surface of activated aluminum foil is
perfectly hydrophilic thus allowing for two- dimensional
spreading of bacteria on its surface. Such procedure for
bacteria fixation therefore allows for uniform treatment
of bacteria with plasma particles.
An exposure of bacteria to argon ions causes a strong
damage. Figure 4 represents the SEM image of bacteria
after receiving the argon ion dose of 5.4 × 1021 m–2. The
bacteria are definitely not capable of revitalization what
was proved also by control experiments using the plate
count technique. It is interesting that the damage caused
by ions is far from being uniform. Namely, a hole – like
structure of the bacterial cells is observed. Although it is
known that ion beam etching is never perfectly
homogeneous and isotropic, such rich surface
morphology cannot be due to common effects observed
at ion beam etching of organic materials. The observed
morphology may be attributed to appearance of the local
surface electrical charge during treatment with positively
charged ions. Namely, the electrical conductivity of
bacteria is poor. Since the composition of the cell wall is
far from being uniform, some spots on the surface may
keep larger charge than other. The surface charge
influence the local uniformity of the ion flux on the
surface causing local focusing and thus further
non-uniformity of the ion beam etching. Finally, the
bacteria obtain morphology as shown in Figure 4. The
ions practically cannot reach the uppermost part of
bacteria since positive charge prevents it.
The SEM image of bacteria treated with oxygen at-
oms shows a completely different picture. In this case,
the badly damaged bacteria are flattened, also. In fact,
little material remained after receiving the dose of ap-
proximately 4.5 × 1025 m–3. The remains observed on the
surface of the aluminum foil after treatment with oxygen
atoms represent only ash – mostly inorganic remains of
the bacterial material after rather complete oxidation of
organic material. This picture is in agreement with previ-
ous observations on selective etching of organic materi-
als by oxygen radicals21.
5 CONCLUSIONS
Bacteria E. coli were deposited onto aluminum foils
and exposed to positively charged argon ions or neutral
oxygen atoms in the ground state. In both cases, the sam-
ples were kept at room temperature. Since argon is inert
gas that does not interact chemically with organic mate-
rial, the interaction was almost completely kinetic. Apart
K. ELER[I^ ET AL.: DEGRADATION OF BACTERIA ESCHERICHIA COLI ...
Materiali in tehnologije / Materials and technology 44 (2010) 3, 153–156 155
Figure 5: SEM image of bacteria treated with oxygen atoms
Slika 5: SEM-slika bakterije, obdelane z atomi kisika
Figure 4: SEM image of bacteria treated with argon ions
Slika 4: SEM-slika bakterije, obdelane z ioni argona
Figure 3: SEM image of untreated bacteria
Slika 3: SEM-slika neobdelane bakterije
from radiation damage, the argon ions caused sputtering
of the bacterial material. The sputtering was extremely
inhomogeneous what was explained by local charging of
the bacteria. In the case of oxygen atoms, any kinetic ef-
fect is neglected since the O atoms are thermal at room
temperature. In this case, rather uniform degradation of
bacteria occurred and only ashes remained after the treat-
ment. The interaction of O atoms with bacteria is there-
fore purely chemical. In both cases, bacteria were badly
damaged and unable to revitalize.
ACKNOWLEDGEMENT
This research was funded by Slovenian Research
Agency, Contract No. P2 – 0082.
6 REFERENCES
1 Moisan M, Barbeau J, Moreau S, Pelletier J, Tabrizian M, Yahia L’H,
Low-temperature sterilization using gas plasmas: a review of the ex-
periments and an analysis of the inactivation mechanisms, Int. J.
Pharm. 226 (2001), 1–2
2 Soloshenko IA, Tsiolko VV, Khomich VA, Shchedrin AI, Ryabtsev
AV, Bazhenov VY, Mikhno IL, Sterilization of medical products in
low-pressure glow discharges, Plasma Phys. Reports 26 (2000),
792–800
3 Park BJ, Takatori K, Lee MH, Han DW, Woo YI, Son HJ, Kim JK,
Chung KH, Hyun SO, Park JC, Escherichia coli sterilization and
lipopolysaccharide inactivation using microwave-induced argon
plasma at atmospheric pressure, Surf. Coat. Technol. 201 (2007),
5738–5741
4 Fridman G, Brooks AD, Balasubramanian M, Fridman A, Gutsol A,
Vasilets VN, Ayan H, Friedman G, Comparison of direct and indirect
effects of non-thermal atmospheric-pressure plasma on bacteria,
Plasma Process. Polym. 4 (2007), 370–375
5 Moreau S, Moisan M, Tabrizian M, Barbeau J, Pelletier J, Ricard A,
Yahia L’H, Using the flowing afterglow of a plasma to inactivate Ba-
cillus subtilis spores: Influence of the operating conditions, J. Appl.
Phys. 88 (2000), 1166–1174
6 Kylian O, Rossi F, Sterilization and decontamination of medical in-
struments by low-pressure plasma discharges: application of
Ar/O-2/N-2 ternary mixture, J. Phys. D Appl. Phys. 42 (2009),
085207
7 Ricard A, Moisan M, Moreau S, Determination, through titration
with NO, of the concentration of oxygen atoms in the flowing after-
glow of Ar-O2 and N2-O2 plasmas used for sterilization purposes, J.
Phys. D Appl. Phys. 34 (2001), 1203–1212
8 Lerouge S, Wertheimer MR, Marchand R, Tabrizian M, Yahia LH,
Effect of gas composition on spore mortality and etching during
low-pressure plasma sterilization, J. Biomed. Mater. Res. 51 (2000),
128–135
9 Stapelmann K, Kylian O, Denis B, Rossi F, On the application of in-
ductively coupled plasma discharges sustained in Ar/O2/N2 ternary
mixture for sterilization and decontamination of medical instruments,
J. Phys. D Appl. Phys. 41 (2008), 192005
10 Cvelbar U, Vujo{evi} D, Vratnica Z, Mozeti~ M, The influence of
substrate material on bacteria sterilization in an oxygen plasma glow
discharge, J. Phys. D Appl. Phys. 39 (2006), 3487–3493
11 Vujo{evi} D, Mozeti~ M, Cvelbar U, Krstulovi} N, Milo{evi~ S, Op-
tical emission spectroscopy characterization of oxygen plasma dur-
ing degradation of Escherichia coli, J. Appl. Physic. 101 (2007),
103305-1–103305-7
12 Canal C, Gaboriau F, Villeger S, Cvelbar U, Ricard A, Studies on an-
tibacterial dressings obtained by fluorinated post-discharge plasma,
Int. J. Pharm. 367 (2009), 155–161
13 Vratnica Z, Vujo{evi} D, Cvelbar U, Mozeti~ M, Degradation of bac-
teria by weakly ionized highly dissociated redio-frequency oxygen
plasma, IEEE Trans. Plasma Sci. 36 (2008), 1300–1301
14 Vujo{evi} D, Vratnica Z, Vesel A, Cvelbar U, Mozeti~ M, Drenik A,
Mozeti~ T, Klanj{ek Gunde M, Hauptman N, Oxygen plasma steril-
ization of bacteria, Mater. Technol. 40 (2006), 227–232
15 Drenik A, Cvelbar U, Vesel A, Mozeti~ M, Weakly ionized oxygen
plasma, Inf. MIDEM 35 (2005), 85–91
16 Sorli I, Petasch W, Kegel B, Schmid H, Liebl G, Plasma processes 1.
Plasma basics, plasma generation, Inform. Midem, 26 (1996), 35–45
17 Babi~ D, Poberaj I, Mozeti~ M, Fiber optic catalytic probe for
weakly ionized oxygen plasma characterization, Rev. Sci. Instrum.
72 (2001), 4110–4114
18 Poberaj I, Mozeti~ M, Babi~ D, Comparison of fiber optics and stan-
dard nickel catalytic probes for determination of neutral oxygen at-
oms concentration, J. Vac. Sci. Technol. A 20 (2002), 189–193
19 Mozeti~ M, Characterization of reactive plasmas with catalytic
probes, Surf. Coat. Technol. 201 (2007), 4837–4842
20 Stirm S, Freund-Mölbert E, Escherichia coli Capsule Bacteriophages
II. Morphology, J Virol. 8 (3) (1971), 330–342
21 Mozeti~ M, Controlled oxidation of organic compounds in oxygen
plasma, Vacuum 71 (2003), 237–240
K. ELER[I^ ET AL.: DEGRADATION OF BACTERIA ESCHERICHIA COLI ...
156 Materiali in tehnologije / Materials and technology 44 (2010) 3, 153–156
S. KRUM ET AL.: EFFECT OF HEAT TREATMENT ON THE MICROSTRUCTURE AND PROPERTIES ...
EFFECT OF HEAT TREATMENT ON THE
MICROSTRUCTURE AND PROPERTIES OF Cr-V
LEDEBURITIC STEEL
VPLIV TOPLOTNE OBDELAVE NA MIKROSTRUKTURO IN
LASTNOSTI LEDEBURITNEGA JEKLA Cr-V
Stanislav Krum, Jana Sobotová, Petr Jur~i
Department of Material Science, Faculty of Mechanical Engineering, CTU in Prague, Karlovo nam. 13, 12135 Prague, Czech Republic
s.krum@seznam.cz
Prejem rokopisa – received: 2009-11-16; sprejem za objavo – accepted for publication: 2010-03-01
Duplex coating of high-alloyed tool steels is the newest and most promising technology that allows improvements to tools in
terms of wear resistance and service time. However, the deposition of thin films onto the nitrided surface is associated with
problems that are not yet solved. To clarify the problem of the adhesion and mechanical properties of the system
substrate/coating it is necessary to prepare the substrate with well-defined microstructural parameters. The Vanadis 6 PM
ledeburitic steel was chosen as a substrate for surface processing. It was austenitised, quenched and double tempered to a
desired hardness of HRC = (57–58). Some specimens were also sub-zero treated in order to improve the transformation of
austenite to martensite. The microstructure of the heat-processed steel was examined using optical microscopy and scanning
electron microscopy. The resistance to nucleation of fracture was assessed with the three-point bending-strength test and the
fracture surfaces were subjected to a fractographic analysis.
Keywords: Sub-zero treatment, hardness, ledeburitic tool steel
Dupleksno pokritje visokolegiranih orodnih jekel je nova in obetajo~a tehnologija za pove~anje obrabne obstojnosti in dobe
uporabnosti orodij. Vendar je nanos tanke plasti na nitrirano povr{ino povezana s problemi, ki {e niso re{eni. Za razlago
problema adhezije in mehanskih lastnosti sistema podlaga-pokritje je treba pripraviti podlago z dobro opredeljenimi parametri
mikrostrukture. Jeklo Vanadis 6 PM je bilo izbrano za podlago za nadaljevanje procesiranja. Bilo je avstenitizirano, kaljeno in
dvakrat popu{~eno na `eleno trdoto HRC = (57–58). Nekateri preizku{anci so bili tudi podhlajeni zaradi pove~anja obsega
premene avstenita v martenzit. Mikrostruktura je bila preiskana z opti~no in vrsti~no elektronsko mikroskopijo. Odpornost proti
za~etku razpoke je bila preverjena s trito~kovnim upogibom, povr{ina prelomov pa fraktografsko preiskana.
Klju~ne besede: podhladitev, trdota, ledeburitno jeklo
1 INTRODUCTION
Generally, heat treatment improves a steel’s tough-
ness and hardness, and it is absolutely necessary for the
proper functioning of tool steels. The heat treatment
usually consists of austenitizing, quenching and is
followed by multiple tempering. After this procedure the
tool steels gain properties that are suitable for industrial
applications.
In tool steels, retained austenite is an undesired
component because it lowers the hardness after the heat
treatment. Additionally, it can transform into martensite
and/or bainite at elevated working temperatures or under
high loading. This worsens the operational suitability of
tools and strong efforts are made to minimize the amount
of retained austenite.
With this aim, the sub-zero treatment was introduced
in the 1960s. This treatment hinders the retained
austenite stabilization in steels with the Mf temperature
lower than room temperature. A typical example of these
steels is the group of ledeburitic steels, with a higher
carbon content and with a large amount of alloying
elements.
Nevertheless, the opinions on the effect of the
sub-zero treatment on the working utility of tools are
diverse because the nature of the process has not yet
been entirely clarified. In 1962 Rapatz concluded that
there is no advantage in using these methods for lede-
buritic high-speed steels 1. Also, Hoyle did not
recommend this type of heat treatment for ledeburitic
steels 2. On the other hand, Berns stated different
recommendations for ledeburitic steels. He found that it
was possible to significantly increase the hardness of
chromium ledeburitic steels with an immediate sub-zero
treatment after quenching 3. He stated that the immediate
sub-zero treatment after quenching from the usual
temperatures and tempering allowed a hardness increase
of the low-carbon-content steels. For the steel
X155CrVMo 12 1, it is possible to reach HRC = 67 and
in the case of superhard ledeburitic steel X290Cr even
HRC = 69. Thus, it is evident that older studies focused
on the research of this type of heat treatment did not
produce definite results. On top of that, in the case of
ledeburitic high-speed steels the results were not even
optimistic. Nevertheless, in recent years the experiments
engaged in sub-zero treatment have become more
realistic.
The goal of this study was to investigate what
happens when Cr-V ledeburitic steels are plasma
nitrided. For this purpose, it was necessary to prepare a
Materiali in tehnologije / Materials and technology 44 (2010) 3, 157–161 157
UDK 669.14.018.252:620.18:621.785 ISSN 1580-2949
Professional article/Strokovni ~lanek MTAEC9, 44(3)157(2010)
substrate with a well-defined microstructure and
properties. In this paper, the Vanadis 6 steel processed
using various heat treatments, including a sub-zero
period, is investigated.
2 EXPERIMENTAL
2.1 Material
The samples were made from Vanadis 6 – a ledebu-
ritic cold-work tool steel with the chemical composition
of 2.1 % C, 7 % Cr, 6 % V and an initial hardness of
HV10 = 284 Fifteen specimens for three-point bend tests
were net-shape machined to dimensions of (10 × 10 ×
100) mm and then submitted to the heat treatments in
Table 1.
2.2 Methods
Several measurements and experiments for deter-
mining the material and structural properties were
performed. The hardness was measured with the
EMCOTEST M4C 075G3 hardness tester using the
Vickers HV10 and Rockwell HRC methods. These two
methods were selected because of the expected high
material surface hardness. The resistance against crack
initiation was determined with a three-point bending test.
The microstructure of the steel in the as-received state
and of heat treated steel were examined with optical
microscopy and scanning electron microscopy, applying
a Neophot 32 optical microscope and a Jeol JSM 5410
scanning electron microscope. Also, the quantitative
analysis of the carbides’ dissolution during the auste-
nitizing, which consisted of determining the particle
count and the size classes, was performed. The first out
was executed using the micrograph of the as-received
material and the second one using the micrograph after
the heat treatment A. Within the fractography study,
micrographs of two material states were taken: after
heat-treatment A (i.e. without the sub-zero treatment)
and after heat treatment C (i.e., with sub-zero treatment
–196 °C; 7 h).
3 EXPERIMENTAL RESULTS
3.1 Optical and scanning microscopy
The as-received steel Vanadis 6 has a microstructure
with fine carbide particles uniformly distributed through-
out the ferritic matrix. These carbides are of eutectic,
secondary and eutectoid origin, as shown in the micro-
graphs in Figure 1 and Figure 2.
After the heat treatment (Figure 3), the matrix con-
sists of tempered martensite and particles of the carbides
M7C3 and MC.
The material submitted to heat treatment B, i.e., with
sub-zero treatment, shows, at first glance, a similar
microstructure, MC and M7C3 carbide particles in a
martensitic matrix with particles of carbide M7C3 fewer
and smaller than after heat-treatment A, Figure 4.
The results of the quantitative analysis of the carbide
particles are presented in Figure 5 (as-received) and
S. KRUM ET AL.: EFFECT OF HEAT TREATMENT ON THE MICROSTRUCTURE AND PROPERTIES ...
158 Materiali in tehnologije / Materials and technology 44 (2010) 3, 157–161
Table 1: Heat treatment details
Tabela 1: Podatki o toplotni obdelavi
Specimen Heat-treatmentmarking
Heat-treatment process details
Austenitizing
Quenching
p(N2)/bar
Sub-zero treatment Tempering
1 − 5 A 1030 °C; 30 min 6 – 2 × 530 °C; 1 h
6 − 10 B 1030 °C; 30 min 6 −196 °C; 4 h 2 × 530 °C; 1 h
11 − 15 C 1030 °C; 30 min 6 −196 °C; 7 h 2 × 530 °C; 1 h
Figure 2: Microstructure of the as-received state (SEM)
Slika 2: Mikrostruktura prejetega jekla, SEM-posnetek
Figure 1: Microstructure of the as-received material state (optical
microscopy)
Slika 1: Mikrostruktura prejetega jekla, opti~ni posnetek
Figure 6 (after heat treatment). It is obvious that after
the heat treatment the number of carbide particles is
lower and their size is smaller in comparison to the
as-received material. The main reason is that the
as-received material contains a larger number of
ultra-fine eutectoid carbides that undergo a complete
dissolution in the austenite. In addition, also a part of the
secondary carbides is dissolved in the austenite during
the heating up to the final austenitizing temperature
3.2 Hardness
The results of the hardness measurements obtained
from the samples of all the heat treatments are presented
in Table 2. The hardness of the as-received material was
HV10 = 284 and the average hardness after heat treatment
without a sub-zero treatment was of HV10 = (748 ± 6.9)
(HRC = (60 ± 0.2)). The process with the 4-h sub-zero
treatment at –196 °C leads to an average surface hard-
ness of HV10 = (734 ± 6.9) (HRC = (58 ± 1.0)). Finally,
the process including the 7-h sub-zero treatment at –196
°C leads to the average surface hardness of HV10 = (721
± 5.6) (HRC = (58 ± 0.4)).
It is evident that after the sub-zero treatment the
hardness is slightly lower – of the order of tens of
Vickers units compared to the heat treatment without this
treatment. This hardness decrease is higher with a longer
duration of the sub-zero period. These findings do not
agree with 2,4, which stated that the hardness increased
after this procedure.
Table 2: Average hardness and standard deviation after different heat
treatments
Tabela 2: Povpre~na trdota in standardna deviacija po razli~ni toplotni
obdelavi
Measurement Method HV10 HRC
Heat Treatment A B C A B C
Average Value of
Hardness 748 734 721 60 58 58
Standard Deviation 6.9 6.9 5.6 0.2 1.0 0.4
3.3 Three-point bending strength
The three-point bending strength was used to
determine the materials’ resistance to fracture initiation.
The measured parameters were the maximum load until
fracture and the bending strength. The tests were
performed on all the samples and the average values of
the samples with the same heat treatment are presented
S. KRUM ET AL.: EFFECT OF HEAT TREATMENT ON THE MICROSTRUCTURE AND PROPERTIES ...
Materiali in tehnologije / Materials and technology 44 (2010) 3, 157–161 159
Figure 5: Results of the analysis of the carbide particles’ sizes and the
number in the as-received material
Slika 5: Rezultati dolo~itve velikosti in {tevila karbidnih zrn v preje-
tem jeklu
Figure 4: Microstructure of Vanadis 6 after the heat-treatment B
Slika 4: Mikrostruktura jekla Vanadis 6 po toplotni obdelavi B
Figure 3: Microstructure of Vanadis 6 after the heat-treatment A
Slika 3: Mikrostruktura jekla Vanadis 6 po toplotni obdelavi A
Figure 6: Results of the analysis of the carbide particles’ sizes and the
number in the material after the heat treatment
Slika 6: Rezultati dolo~itve velikosti in {tevila karbidnih zrn v
toplotno obdelanem jeklu
in Table 3. The bending strength was calculated from the
Fmax using the formula: R
F L
aMo
=
⋅3
2 3
max
, where: Fmax/N –
load at the fracture, L/mm – length of the sample, a/mm
– width of the sample with square intersection.
Table 3: Results of the three-point bending-strength test
Tabela 3: Rezultati trito~kovnega upogiba
Heat treatment
Average value of
bending strength
/MPa
Average value of
the maximum load
/kN
A 2436 16.24
B 2961 19.74
C 3217 21.45
The relationships between the load and the defor-
mation (F-s) of the samples are shown in the following
diagrams. Figure 7 shows the F-s diagram for the sample
processed with the processing route A (no sub-zero
treatment). The F-s relation is practically linear, there-
fore only a minimum plastic deformation until fracture
can be expected to occur. Also, the three-point bending
strength and the maximum load Fmax were relatively low
(Table 3); the fracture appeared when the load reached
only 16.24 kN. In Figure 8, the behavior after the B heat
treatment (–196 °C, 4 h sub-zero treatment) is shown.
The F-s relation is not a completely linear shape,
especially at high loading, and indicates that some
plastic deformation took place during the testing. The
fracture in this case appeared at a load of 19.74 kN.
Figure 9 shows the results of the three-point bending test
of the sample processed with the C treatment (–196 °C, 7
h sub-zero treatment). The dependence of the load
deformation is even less linear because of the greater
portion of plastic deformation. The fracture occurs at
21.45 kN.
From these results, it seems that the sub-zero period
has a relatively small, but positive, effect on the
three-point bending strength and it agrees with the
changes in hardness, as with a higher bending strength,
the hardness is lower.
3.4 Fractography
The fracture of the specimen processed using the
processing route A is shown in Figure 10. The fracture
was initiated on the tensile-strained side of the specimen
and propagated throughout the specimen. Step-like
propagation tracks are visible in the region near the
surface of the specimen. At a depth of approx. 1 mm, a
great step is also found.
The micrograph in Figure 11 at a higher magnifica-
tion shows that the fracture surface presents a dimpled
morphology with some areas of microcleavages and
indicates that a very low energy was spent for the crack
propagation.
S. KRUM ET AL.: EFFECT OF HEAT TREATMENT ON THE MICROSTRUCTURE AND PROPERTIES ...
160 Materiali in tehnologije / Materials and technology 44 (2010) 3, 157–161
Figure 8: Dependence of F-s for the three-point bending test of the
sample processed using the B treatment (with –196 °C, 4 h sub-zero
treatment)
Slika 8: F-s-odvisnost za trito~kovni upogib po toplotni obdelavi B
(podhladitev –196 °C, 4 h)
Figure 9: Dependence of F-s for the three-point bending test of the
sample processed using the C treatment (with –196 °C, 7 h sub-zero
treatment)
Slika 9: F-s odvisnost za trito~kovni upogib po toplotni obdelavi C
(podhladitev –196 °C, 7 h)
Figure 7: Dependence of F-s for the three-point bending test of the
sample processed using the A treatment (without the sub-zero
treatment)
Slika 7: F-s-odvisnost za trito~kovni upogib po toplotni obdelavi A
(brez podhladitve)
The fracture surface of the sample processed using
the C treatment sample at 35-times magnification is
presented in Figure 12, and at 1000-times magnification
is shown Figure13. The dimples in the fracture surface
of the sample processed using the C treatment are deeper
and the step-like propagation tracks are longer than on
the sample processed using the A treatment (Figure 10).
Considering the F-s diagrams, showing that for the
sample processed using the C treatment a greater amount
of plastic deformation energy was consumed, it is possi-
ble to say that performed sub-zero treatment improves
slightly the ductility of the tool steel.
4 CONCLUSIONS
1) After all of the performed heat-treatment processes
the microstructure of he material consisted of tem-
pered martensite and undissolved carbides.
2) The hardness was slightly higher for the steel without
the sub-zero treatment than for the steel that under-
went this process.
3) The sub-zero processed samples had a slightly higher
three-point bending strength. The reason is that the
crack propagation is connected with a higher con-
sumption of plastic energy until the fracture.
4) The results of the mechanical tests are rather un-
expected and so further, and more accurate, tests and
examinations will be necessary to determine and
verify the causes of this unexpected observation.
Acknowledgements
This paper has originated thanks to the support
obtained from the Czech Universities Development Fund
(IGS) within the grant CTU0908212.
5 LITERATURE
1 Rapatz, F.: Die Edelstaehle, 5. Aufl. Springr Verlag, Berlin/Gottin-
gen/Heidelberg, 1962, p. 862
2 Hoyle, G.: High Speed Steels, Butterworth, London – Boston –
Durban – Singapore – Sydney – Toronto – Wellington, University
Press Cambridge
3 Berns, H.: HTM 29 (1974), 4, s.236
4 Company literature Uddeholm AB Vanadis 6
S. KRUM ET AL.: EFFECT OF HEAT TREATMENT ON THE MICROSTRUCTURE AND PROPERTIES ...
Materiali in tehnologije / Materials and technology 44 (2010) 3, 157–161 161
Figure 13: Detail of the fracture surface of the sample processed
using the C treatment
Slika 13: Detajl povr{ine preloma po toplotni obdelavi C
Figure 12: Fracture surface of the sample processed using the C
treatment
Slika 12: Povr{ina preloma po toplotni obdelavi C
Figure 11: Detail of the fracture surface of the sample processed
using the A treatment
Slika 11: Detajl povr{ine preloma po toplotni obdelavi A
Figure 10: Fracture surface of the sample processed using the A
treatment
Slika 10: Povr{ina preloma po toplotni obdelavi A