scieee Science in your language
[en] (orig)

Steel heat treating: Mathematical modelling and numerical simulation of a problem arising in the automotive industry

Abstract

We describe a mathematical model for the industrial heating and cooling processes of a steel workpiece representing the steering rack of an automobile. The goal of steel heat treating is to provide a hardened surface on critical parts of the workpiece while keeping the rest soft and ductile in order to reduce fatigue. The high hardness is due to the phase transformation of steel accompanying the rapid cooling. This work takes into account both heating-cooling stage and viscoplastic model. Once the general mathematical formulation is derived, we can perform some numerical simulations.

Read accessible full text

Steel heat treating: Mathematical modelling and numerical simulation of a problem arising in the automotive industry

Author: Díaz Moreno, José Manuel; García Vázquez, Concepción; González Montesinos, María Teresa; Ortegón Gallego, Francisco; Viglialoro, Giuseppe
Publisher: ASTES Publishers
Year: 2017
DOI: 10.25046/aj020510
Source: https://idus.us.es/bitstreams/553fd937-886f-490d-aed0-2e4b2d599d38/download
Ad ances in Science, Technology and Enginee ing Sys ems Jou nal
Vol. 2, No. 5, 55-62 (2017)
www.as esj.com
P oceedings o In e na ional Con e ence on Applied Ma hema ics
(ICAM’2017), Taza, Mo occo
ASTES Jou nal
ISSN: 2415-6698
S eel hea ea ing: ma hema ical modelling and nume ical
simula ion o a p oblem a ising in he au omo i e indus y
Jos´
e Manuel D´
ıaz Mo eno1, Concepci´
on Ga c´
ıa V´
azquez1, Ma ´
ıa Te esa Gonz´
alez Mon esinos2,
F ancisco O eg´
on Gallego*,1, Giuseppe Viglialo o3
1Depa amen o de Ma em´
a icas, Facul ad de Ciencias, Uni e sidad de C´
adiz, 11510 Pue o Real, SPAIN,
[email p o ec ed], [email p o ec ed], [email p o ec ed].
2Depa amen o de Ma em´
a ica Aplicada I, ETS de Ingenie ´
ıa In o m´
a ica, Uni e sidad de Se illa, 41012 Se illa,
SPAIN, [email p o ec ed].
3Dipa imen o di Ma ema ica ed In o ma ica, Uni e si `
a degli S udi di Caglia i, iale Me ello 92 – 09123
Caglia i, ITALY, [email p o ec ed].
A R T I C L E I N F O A B S T R A C T
A icle his o y:
Recei ed: 10 June, 2017
Accep ed: 15 July, 2017
Online: 10 Decembe , 2017
We desc ibe a ma hema ical model o he indus ial hea ing and
cooling p ocesses o a s eel wo kpiece ep esen ing he s ee ing ack o
an au omobile. The goal o s eel hea ea ing is o p o ide a ha dened
su ace on c i ical pa s o he wo kpiece while keeping he es so and
duc ile in o de o educe a igue. The high ha dness is due o he phase
ans o ma ion o s eel accompanying he apid cooling. This wo k
akes in o accoun bo h hea ing-cooling s age and iscoplas ic model.
Once he gene al ma hema ical o mula ion is de i ed, we can pe o m
some nume ical simula ions.
Keywo ds :
S eel ha dening
Phase ansi ions
Po en ial Maxwell equa ions
Nume ical Simula ions
Fini e Elemen Me hods
1 In oduc ion
In he au oma i e indus y, many wo kpieces such
gea s, bea ings, acks and pinions, a e made o s eel.
S eel is an alloy o i on and ca bon. Gene ally, indus-
ial s eel has a ca bon con en up o abou 2 w %.
O he alloying elemen s may be p esen , such as C
and V in ools s eels, o Si, Mn, Ni and C in s ain-
less s eels. Mos s uc u al componen s in mechanical
enginee ing a e made o s eel. Ce ain o hese com-
ponen s, such as oo hed wheels, be el gea s, pinions
and so on, engaged each o he s in o de o ansmi
some kind o ( o a ional o longi udinal) mo emen .
In his si ua ion, he con ac su aces o hese compo-
nen s a e pa icula ly s essed. The goal o hea ea -
ing o s eel is o a ain a sa is ac o y ha dness. P io
o hea ea ing, s eel is a so and duc ile ma e ial.
Wi hou a ha dening ea men , and due o he su -
ace s esses, he gea ee h will soon ge damaged and
hey will no longe engage co ec ly.
In his wo k we a e in e es ed in he ma hema ical
desc ip ion o he ha dening p ocedu e o a ca s ee -
ing ack (see Figu e 1). This pa icula si ua ion is one
o he majo conce ns in he au omo i e indus y. In
his case, he goal is o inc ease he ha dness o he
s eel along he oo h line and a he same ime keep-
ing he es o he wo kpiece so and duc ile in o de
o educe a igue. This p oblem is go e ned by a non-
linea sys em o pa ial di e en ial equa ions coupled
wi h a ce ain sys em o o dina y di e en ial equa-
ions. Once he ull sys em is se we pe o m some
nume ical simula ions.
Figu e 1: Ca s ee ing ack.
Solid s eel may be p esen a di e en phases,
namely aus eni e, ma ensi e, baini e, pea li e and e -
i e. The phase diag am o s eel is shown in Fig-
u e 2. Fo a gi en w % o ca bon con en up o 2.11,
*Co esponding au ho . Depa amen o de Ma em´
a icas, Facul ad de Ciencias, Uni e sidad de C´
adiz, 11510 Pue o Real, SPAIN,
[email p o ec ed]
www.as esj.com 55
h ps://dx.doi.o g/10.25046/aj020510
J. M. D´
ıaz e al. / Ad ances in Science, Technology and Enginee ing Sys ems Jou nal Vol. 2, No. 5, 55-62 (2017)
all s eel phases a e ans o med in o aus eni e p o-
ided he empe a u e has been aised up o a ce -
ain ange. The minimum aus eniza ion empe a u e
(727◦) is a ained o a ca bon con en o 0.77 w %
(eu ec oid s eel). Upon cooling, he aus eni e is ans-
o med back in o he o he phases (see Figu e 3), bu
i s dis ibu ion depends s ongly on he cooling s a -
egy ([4, 12]).
Ma ensi e is he ha des cons i uen in s eel, bu
a he same ime is he mos b i le, whe eas pea li e
is he so es and mo e duc ile phase. Ma ensi e de-
i es om aus eni e and can be ob ained only i he
cooling a e is high enough. O he wise, he es o he
s eel phases will appea .
The ha dness o he ma ensi e phase is due o a
s ong supe sa u a ion o ca bon a oms in he i on la -
ice and o a high densi y o c ys al de ec s. F om
he indus ial s andpoin , hea ea ing o s eel has
a colla e al p oblem: ha dening is usually accom-
panied by dis o ions o he wo kpiece. The main
easons o hese dis o ions a e due o (1) he mal
s ains, since s eel phases unde go di e en olume -
ic changes du ing he hea ing and cooling p ocesses,
and (2) expe imen s wi h s eel wo kpieces unde ap-
plied loading show an i e e sible de o ma ion e en
when he equi alen s ess co esponding o he load
is in he elas ic ange. This e ec is called ans o ma-
ion induced plas ici y.
The hea ing s age is accomplished by an
induc ion-conduc ion p ocedu e. This echnique has
been success ully used in indus y since he las cen-
u y. Du ing a ime in e al, a high equency cu -
en passes h ough a coil gene a ing an al e na ing
magne ic ield which induces eddy cu en s in he
wo kpiece, which is placed close o he coil. The eddy
cu en s dissipa e ene gy in he wo kpiece p oducing
he necessa y hea ing.
2 Ma hema ical modeling
We conside he se ing co esponding o Figu e 4.
The domain Ωc ep esen s he induc o (made o cop-
pe ) whe eas Ωss ands o he s eel wo kpiece o be
ha dened. He e, he coil is he domain Ω=Ωs∪Ωc∪
S0. In his way, he wo kpiece i sel akes pa o he
coil.
In o de o desc ibe he hea ing-cooling p ocess,
we will dis inguish wo subin e als o ming a pa i-
ion o [0,T ], namely [0,T ] = [0,Th)∪[Th,Tc], Tc> Th>
0. The i s one [0,Th) co esponds o he hea ing p o-
cess. All along his ime in e al, a high equency
elec ic cu en is supplied h ough he conduc o
which in i s u n induces a magne ic ield. The com-
bined e ec o bo h conduc ion and induc ion gi es
ise o a p oduc ion e m in he ene gy balance equa-
ion (14), namely b(θ)|A +∇φ|2. This is Joule’s hea ing
which is he p incipal e m in hea p oduc ion. In ou
model, we will only conside h ee s eel phase ac-
ions, namely aus eni e (a), ma ensi e (m), and he
es o phases ( ). In his way, we ha e a+m+ = 1
and 0 ≤a,m, ≤1 in Ωs×[0,T ]. A he ini ial ime we
ha e (0) = 1 in Ωs. Upon hea ing only aus eni e can
be ob ained. In pa icula m= 0 in Ωs×[0,Th] and he
ans o ma ion o aus eni e is de i ed a he expense
o he o he phase ac ions ( ).
A he ins an =Th, he cu en is swi ched o
and du ing he ime in e al [Th,Tc] he wo kpiece is
se e ely cooled down by means o aqua-quenching.
The hea ing model
The cu en passing h ough he se o conduc o s
Ω=Ωc∪Ωs∪S0is modeled by he elec ic po en-
ial di e ence, ϕ0, applied on he su ace Γ2⊂Ωc(see
Figu e 4). No ice ha he applied po en ial on Γ1is
ze o. In he sequel, we pu Γ=Γ1∪Γ2.
The hea ing model in ol es he ollowing un-
knowns: he elec ic po en ial, φ; he magne ic ec-
o po en ial, A= (A1,A2,A3); he s ess enso ,
σ= (σij)1≤i,j≤3,σij =σji o all 1 ≤i,j ≤3; he dis-
placemen ield u= (u1,u2,u3); he aus eni e phase
ac ion, a; and he empe a u e, θ. Among hem, only
Ais de ined in he domain Dcon aining he se o
conduc o s Ω. On he o he hand, since he induc o
and he wo kpiece a e in close con ac , bo h φand θ
a e de ined in Ω. Since phase ansi ions only occu
in he wo kpiece, we may neglec de o ma ions in Ωc.
This implies ha σ,uand aa e only de ined in he
wo kpiece Ωs.
Since elec omagne ic ields gene a ed by high e-
quency cu en s a e sinusoidal in ime, bo h he elec-
ic po en ial, φ, and he magne ic po en ial ield, A,
ake he o m ([1, 2, 14, 15]) M(x, ) = Reheiω M(x)i,
whe e Mis a complex- alued unc ion o ec o ield,
and ω= 2π is he angula equency, being he
elec ic cu en equency. In gene al, Malso depends
on , bu a a ime scale much g ea e han 1/ω. In
his way, we may in oduce he complex- alued ields
ϕand Aas
φ= Re[eiω ϕ(x, )],A= Re[eiω A(x, )].(1)
As a a as he nume ical simula ion o a sys em
like (2)-(15) is conce ned, he in oduc ion o he new
a iables ϕand Ais qui e con enien since he ime
scale desc ibing he e olu ion o bo h ϕand Ais much
smalle han ha o he empe a u e θ. In he case o
s eel hea ea ing, is abou 80 KHz.
The hea ing model eads as ollows ([3, 9, 10, 7]):
∇ · (b(θ)∇ϕ) = 0 in ΩTh=Ω×(0,Th),(2)
∂ϕ
∂n = 0 on (∂Ω Γ)×(0,Th),(3)
ϕ= 0 on Γ1×(0,Th), ϕ=ϕ0on Γ2×(0,Th), (4)
b0(θ)iωA+∇ × 1
µ∇ × A!−δ∇(∇ · A)
=−b0(θ)∇ϕin D×(0,Th),(5)
A= 0 on ∂D ×(0,Th),(6)
www.as esj.com 56
J. M. D´
ıaz e al. / Ad ances in Science, Technology and Enginee ing Sys ems Jou nal Vol. 2, No. 5, 55-62 (2017)
Tempe a u e ◦C
400
600
800
1000
1200
1400
1600
012345 6 7
Fe % C Fe3C (C)
Hypoeu ec oide
Hype eu ec oide
α
(Fe)
Fe i e + Cemen i e
727 ◦C
1148 ◦C
912 ◦C
1394 ◦C
0.77%
(Eu ec oid)
2.11% 4.30%
Aus eni e + Cemen i e
Liquid
γ+ Liquid
γ
Aus eni e
Figu e 2: I on-ca bide phase diag am.
Fe i e
Pea li e
Baini e
Ma ensi e
Aus eni e
Fe i e
Pea li e
Baini e
Ma ensi e
Hea ing Cooling
Figu e 3: Mic ocons i uen s o s eel. Upon hea ing, all phases a e ans o med in o aus eni e, which is ans o med back o he o he
phases du ing he cooling p ocess. The dis ibu ion o he new phases depends s ongly on he cooling s a egy. A high cooling a e
ans o ms aus eni e in o ma ensi e. A slow cooling a e ans o ms aus eni e in o pea li e.
www.as esj.com 57
J. M. D´
ıaz e al. / Ad ances in Science, Technology and Enginee ing Sys ems Jou nal Vol. 2, No. 5, 55-62 (2017)
Ωs(s eel)
Ωc(coppe ) Ωc
Γ1Γ2
S0
S0
D
Figu e 4: Domains D,Ω=Ωs∪Ωc∪S0and he aces Γ1,Γ2⊂Ωc. The induc o Ωcis made o coppe . The wo kpiece con ains a oo hed
pa o be ha dened by means o he hea ing-cooling p ocess desc ibed below. I is made o a hypoeu ec oid s eel. The domain is aken as
a big enough ec angle con aining bo h he induc o and he ack
−∇ · σ=Fin Ωs×(0,Th),(7)
σ=Kε(u)−A1(a,m,θ)I
−Z
0
γ(a,m,a ,m ,θ)Sdτ,(8)
u= 0 on Γ0×(0,Th),(9)
σ·n= 0 on (∂Ωs Γ0)×(0,Th),(10)
a =1
τa(θ)(aeq(θ)−a)H(θ−As) in Ωs×(0,Th),(11)
a(0) = 0 in Ωs,(12)
α(θ,a,m,σ)θ − ∇ · (κ(θ)∇θ)
+3 ¯
κq(a,m)θ(∇ · u −3A2(a ,m ,θ))
=b(θ)|A +∇φ|2−ρLaa
+A2(a ,m ,θ) σ
+γ(a,m,a ,m ,θ)|S|2in ΩTh,(13)
∂θ
∂n = 0 on ∂Ω×(0,Th),(14)
θ(0) = θ0in Ω.(15)
He e, b(θ) is he elec ical conduc i i y (by b(θ)
we mean he unc ion (x, )7→ b(x,θ(x, )), and also o
κ(θ), e c.); ϕ0 ep esen s he po en ial ex e nal sou ce.
The domain Dcon aining he se o conduc o s is
aken big enough so ha he magne ic ec o po en-
ial A anishes on i s bounda y ∂D (in ou model, is
aken o be a 2D ec angle o a 3D cube). Since bo h
σand aa e only de ined in Ωs, when hey appea in
a e m e e ed in Ω, we mean ha his e m anishes
ou side Ωs( o ins ance, −ρLaa appea ing in (13));
b0(x,s) = b(x,s) i x∈Ω,b0(x,s) = 0 elsewhe e; µ=µ(x)
is he magne ic pe meabili y; δ > 0 is a small cons an ;
Fis a gi en ex e nal o ce (usually F= 0); K=Kijkl,
1≤i,j,k,l ≤3 is he s i ness enso . S eel can be con-
side ed as an iso opic and homogenous ma e ial so
ha
Kijkl =¯
λδijδkl+¯
µ(δikδjl+δilδjk), o all i,j,k,l ∈ {1,2,3}
whe e ¯
λ≥0 and ¯
µ > 0 a e he Lam´
e coe icien s
o s eel; ε(u) = 1
2(∇u+∇uT) is he s ain enso ;
A1(a,m,θ)Imodels he he mal s ain, Ibeing he 3×3
uni y ma ix, whe eas A1(a,m,θ) is de ined as
A1(a,m,θ) = qaa(θ−θa) + qmm(θ−θm)
+q (1 −a−m)(θ−θ ),
and in i s u n qa,qmand q a e he he mal expan-
sion coe icien s o he phase ac ions a,mand , e-
spec i ely, and θa,θmand θ a e e e ence empe a-
u es (no ice ha du ing he hea ing s age is m= 0);
R
0γ(a,m,a ,m ,θ)Sdτgi es he model, h ough he
unc ion γ, o he ans o ma ion induced plas ici y
s ain enso , whe e S=σ−1
3 σIis he de ia o o σ,
ha is, he ace ee pa o he s ess enso ; Γ0is a
ce ain smoo h enough pa o ∂Ωs;nis he uni ou e
no mal ec o o he e e ed bounda y; he unc ions
τa(θ), aeq(θ) a e gi en om expe imen al da a (see
Figu e 5), and His he Hea iside unc ion; κ(θ) is he
he mal conduc i i y; he unc ions appea ing in (13)
a e gi en as ollows
α(θ,a,m,σ) = ρcε−9¯
κq(a,m)2θ−q(a,m) σ,
whe e ρand cεa e he s eel densi y and he spe-
ci ic hea capaci y a cons an s ain, espec i ely, ¯
κ=
1
3(3 ¯
λ+ 2 ¯
µ) is he bulk modulus, and q(a,m) is de ined
as
q(a,m) = qaa+qmm+ (1 −a−m)q ;
A2(a ,m ,θ) = qaa (θ−θa) + qmm (θ−θm)
−q (a +m )(θ−θ ).
Finally, La>0 is he la en hea ela ed o he aus en-
i e phase ac ion. No ice ha , in a mo e gene al si u-
a ion ρ,cεand Lamay also depend on a,mand/o θ.
www.as esj.com 58
J. M. D´
ıaz e al. / Ad ances in Science, Technology and Enginee ing Sys ems Jou nal Vol. 2, No. 5, 55-62 (2017)
AsA θ
τa(θ)
aeq(θ)
Figu e 5: Func ions aeq and τa.
Equa ions (2) and (5) de i e om Maxwell’s equa-
ions. In [9], i is assumed he Coulomb gauge condi-
ion o he magne ic ec o po en ial, namely, ∇ · A=
0. He e, we do no impose his condi ion since his
makes appea an undesi ed p essu e g adien in he
equa ion o A. In i s u n, we include a penal y e m
in his equa ion o he o m −δ∇(∇ · A). In doing so,
bo h he heo e ical analysis and he nume ical simu-
la ions a e simpli ied.
Equa ion (7) is a quasis a ic balance law o mo-
men um and (8) is Hooke’s law. The ans o ma ion
o aus eni e om he ini ial phase (0) = 1 is desc ibed
in (11).
Finally, equa ion (13) de i es om he balance law
o in e nal ene gy. As i has been poin ed ou abo e,
Joule’s hea ing is he main esponsible in hea p o-
duc ion. Since γ(a,m,a ,m ,θ)|S|2≥0, he con ibu-
ion o he ans o ma ion induced plas ici y o he
ene gy balance is also a p oduc ion e m. On he o he
hand, du ing he hea ing s age we ha e a ≥0 so ha
−ρLaa ≤0. This means ha he ans o ma ion o
aus eni e abso bs ene gy, which is eleased du ing he
cooling s age.
The cooling model
The hea ing p ocess ends, he high equency cu en
passing h ough he coil is swi ched-o and aqua-
quenching begins. The quenching is jus modeled ia
he Robin bounda y condi ion gi en in (25).
We pu aTh=a(Th), ha is, aThis he aus eni e
phase ac ion dis ibu ion a he inal hea ing ins an
Thob ained om (11). In he same way, we de ine
θTh=θ(Th). Ob iously, hese unc ions will be aken
as he ini ial phase ac ion dis ibu ion and empe -
a u e, espec i ely, in he cooling model. He e we
use he Kois inen-Ma bu ge model ([11, 13]) o he
desc ip ion o he ans o ma ion o ma ensi e om
aus eni e.
The cooling model eads as ollows
−∇ · σ=Fin Ωs×(Th,Tc),(16)
σ=Kε(u)−A1(a,m,θ)I
−Z
0
γ(a,m,a ,m ,θ)Sdτ,(17)
u= 0 on Γ0×(Th,Tc),(18)
σ·n= 0 on (∂Ωs Γ0)×(Th,Tc),(19)
a =1
τa(θ)(aeq(θ)−a)H(θ−As) in Ωs×(Th,Tc),(20)
a(Th) = aThin Ωs,(21)
m =cm(1 −m)H(−θ )H(Ms−θ) in Ωs×(Th,Tc),(22)
m(Th) = 0 in Ωs,(23)
α(θ,a,m,σ)θ − ∇ · (κ(θ)∇θ)
+ 3 ¯
κq(a,m)θ(∇ · u −3A2(a ,m ,θ))
=−ρLaa +ρLmm +A2(a ,m ,θ) σ
+γ(a,m,a ,m ,θ)|S|2in Ω×(Th,Tc),(24)
∂θ
θn =β(x, )(θ−θe) on ∂Ω×(Th,Tc),(25)
θ(Th) = θThin Ω.(26)
In (22) cm>0 is a cons an alue. Also, in (24),
Lm>0 is he la en hea ela ed o he ma ensi e
phase ac ion. The unc ion β(x, ) in (25) is a hea
ans e coe icien and is gi en by
β(x, ) = (0 on ∂Ω∩∂Ωc,
β0( ) on ∂Ω∩∂Ωs.
whe e β0( )>0 (usually aken o be cons an ). Finally,
θeis he empe a u e o he quenchan .
The ma hema ical analysis o a sys em simila
o (16)-(26) can be seen in [3]. In his e e ence, an
exis ence esul is shown assuming ha he da a a e
smoo h enough and Tc−This su icien ly small.
Dh
Figu e 6: Domain iangula ion. The mesh con ains 61790 ian-
gles and 30946 e ices.
www.as esj.com 59

J. M. D´
ıaz e al. / Ad ances in Science, Technology and Enginee ing Sys ems Jou nal Vol. 2, No. 5, 55-62 (2017)
3 Nume ical simula ion
Using he F ee em++ package ([8]), we ha e pe -
o med some nume ical simula ions o he app oxi-
ma ion o he solu ion o he sys ems (2)-(15) and (16)-
(26). We wan o desc ibe he ha dening ea men o
a ca s ee ing ack du ing he hea ing-cooling p ocess.
The goal is o p oduce ma ensi e along he oo h line
oge he wi h a hin laye in i s neighbo hood inside
he s eel wo kpiece ([5, 6]).
Dh
Figu e 7: Domain iangula ion. Elemen densi y nea h ee
ee h.
Figu e 4 shows he open se s D,Ω=Ωs∪Ωc∪S
and he aces Γ1and Γ2( hey appea s ick oge he
in his igu e) which in e ene in he se ing o he
p oblem. The wo kpiece con ains a oo hed pa o
be ha dened by means o he hea ing-cooling p ocess
desc ibed abo e. I is made o a hypoeu ec oid s eel.
The open se D ¯
Ωis ai . The magne ic pe meabili y µ
in (5) is hen gi en by
µ(x) = 








µ0i x∈D ¯
Ω,
0.99995µ0i x∈Ωc,
2.24 ×103µ0i x∈Ωs,
whe e µ0= 4π×10−7(N/A2) is he magne ic cons an
( acuum pe meabili y).
The ma ensi e phase can only de i e om he
aus eni e phase. Thus we need o ans o m i s
he c i ical pa o be ha dened ( he oo h line) in o
aus eni e. Fo ou hypoeu ec oid s eel, aus eni e
only exis s in a empe a u e ange close o he in-
e al [1050,1670] (in K). Du ing he i s s age, he
wo kpiece is hea ed up by conduc ion and induc ion
(Joule’s hea ing) which ende s he oo h line up o
he desi ed empe a u e. In o de o ans o m he
aus eni e in o ma ensi e, we mus cool i down a a
e y high a e. This second s age is accomplished by
aquaquenching.
In his simula ion, he inal ime o he hea ing
p ocess is Th= 5.5 seconds and he cooling p ocess
ex ends also o 5.5 seconds, ha is Tc= 11.
We ha e used he ini e elemen s me hod o he
space app oxima ion and a C ank-Nicolson scheme
o he ime disc e iza ion. Figu es 6 and 7 show he
iangula ion o Din ou nume ical simula ions. We
ha e used P2-Lag ange app oxima ion o ϕ,Aand θ
and P1 o aand m.
In Figu e 8 we can see he empe a u e dis ibu-
ion o he ack along he oo h line a di e en in-
s an s o he he hea ing-cooling p ocess. The ini ial
empe a u e is θ0= 300K. A = 5.5 he hea ing p o-
cess ends and he compu ed empe a u e shows ha
he empe a u e along he ack oo h line lies in he
in e al [1050,1670] (K).
= 1s
= 3s
= 5.5s
= 6s
= 7s
Figu e 8: Tempe a u e e olu ion a ins an s = 1, = 3, = 5.5
(end o he hea ing s age, aqua-quenching begins), = 6 and = 7
seconds, espec i ely. A = 5.5s he empe a u e along he oo h
pa has eached he aus eniza ion le el in his pa o he ack. The
empe a u e is measu ed in Kel in.
= 5.5s (le ), = 6.5s ( igh ),
= 7s (le ), = 8s ( igh )
= 9s (le ) and = 11s ( igh ).
Figu e 9: T ans o ma ion o he aus eni e phase ac ion du ing
he aquaquenching a ime ins an s =5.5, 6.5, 7, 8, 9, and 11 sec-
onds, espec i ely.
www.as esj.com 60
J. M. D´
ıaz e al. / Ad ances in Science, Technology and Enginee ing Sys ems Jou nal Vol. 2, No. 5, 55-62 (2017)
Figu e 9 shows he aus eni e e olu ion om he
beginning o he cooling s age. Blue co esponds o
0% while ed is 100%. We obse e ha ma ensi e
s a s o appea , app oxima i ely, one second a e he
beginning o he cooling s age. A he inal ins an ,
all he amoun o aus eni e has been ans o med in o
ma ensi e as i is shown in Figu e 10 .
= 5.5s (le ), = 6.5s ( igh ),
= 7s (le ), = 8s ( igh )
= 9s (le ) and = 11s ( igh ).
Figu e 10: T ans o ma ion o he ma ensi e phase ac ion om
aus eni e du ing he aquaquenching a ime ins an s =5.5, 6.5, 7,
8, 9, and 11 seconds, espec i ely.
Figu e 11 shows he aus eniza ion along he oo h
line a he end o he hea ing p ocess T= 5.5 seconds.
Figu e 12 shows he inal dis ibu ion o ma en-
si e om aus eni e along he ack oo h line h ough
he cooling s age = 11 seconds. We ha e good ag ee-
men e sus he expe imen al esul s ob ained in he
indus ial p ocess.
Du ing he hea ing-cooling p ocess, he wo k-
piece is de o med so ha an indus ial ec i ica ion is
needed (o o he wise he ack would be useless). Fig-
u es 13 and 14 shows he di e en de o ma ions un-
de gone by he wo kpiece.
Figu e 11: Hea ing p ocess. Aus eni e a = 5.5 along he ack
oo h line.
Figu e 12: Cooling p ocess. Ma ensi e ans o ma ion a he i-
nal s age o he cooling p ocess = 11 seconds.
Figu e 13: Dis o ed mesh (wi h a scale ac o o 10) a e he
hea ing s age. The aus eni e ans o ma ion along he oo h line
changes he o iginal p o ile.
Figu e 14: Dis o ed mesh (wi h a scale ac o o 10) a e he
cooling s age. The o iginal con igu a ion is pa ially eco e ed. Due
o he plas ici y e ec and he lack o he uppe suppo s, he ack
bends down.
www.as esj.com 61
J. M. D´
ıaz e al. / Ad ances in Science, Technology and Enginee ing Sys ems Jou nal Vol. 2, No. 5, 55-62 (2017)
Con lic o In e es The au ho s decla e no con lic
o in e es .
Acknowledgmen s This esea ch was pa ially sup-
po ed by Minis e io de Educaci´
on y Ciencia unde
g an s MTM2010-16401 and TEC2014-54357-C2-2-
R wi h he pa icipa ion o FEDER, and Conseje ´
ıa
de Educaci´
on y Ciencia de la Jun a de Andaluc´
ıa, e-
sea ch g oup FQM–315.
Re e ences
[1] A. Be m ´
udez, J. Bull´
on, F. Pena and P. Salgado, “A nume -
ical me hod o ansien simula ion o me allu gical com-
pound elec odes”, Fini e Elem. Anal. Des., 39, 283–299, 2003.
h ps://doi.o g/10.1016/S0168-874X(02)00069-0
[2] A. Be m ´
udez, D. G´
omez, M. C. Mu˜
niz and P. Salgado, “T an-
sien nume ical simula ion o a he moelec ical p oblem in
cylind ical induc ion hea ing u naces”, Ad . Compu . Ma h.,
26, 39–62, 2007. h ps://doi.o g/10.1007/s10444-005-7470-9
[3] K. Chełminski, D. H¨
ombe g and D. Ke n, “On a he -
momechanical model o phase ansi ions in s eel”,
Ad . Ma h. Sci. Appl., 18, 119–140, 2008.
[4] J. R. Da is e al. “ASM Handbook: Hea T ea ing”, ol. 4, ASM
In e na ional, USA, 2007.
[5] J. M. D´
ıaz Mo eno, C. Ga c´
ıa V´
azquez, M. T. Gonz´
alez Mon-
esinos and F. O eg´
on Gallego, “Nume ical simula ion o
a Induc ion-Conduc ion Model A ising in S eel Ha dening
model a ising in s eel ha dening”, Lec u e No es in Enginee -
ing and Compu e Science, Wo ld Cong ess on Enginee ing
2009, Volume II, July 2009, 1251–1255.
[6] J. M. D´
ıaz Mo eno, C. Ga c´
ıa V´
azquez, M. T. Gonz´
alez Mon-
esinos, F. O eg´
on Gallego and G. Viglialo o, “Ma hema ical
modeling o hea ea men o a s ee ing ack including me-
chanical e ec s”, J. Nume . Ma h. 20, no. 3-4, 215–231, 2012.
h ps://doi.o g/10.1515/jnum-2012-0011
[7] J. Fuh mann, D. H¨
ombe g and M. Uhle, “Nume ical simula-
ion o induc ion ha dening o s eel”, COMPEL, 18, No. 3,
482–493, 1999.
h ps://doi.o g/10.1108/03321649910275161
[8] F. Hech , “New de elopmen in eeFem++”, J. Nume . Ma h.
20, no. 3-4, 251–265, 2012. h ps://doi.o g/10.1515/jnum-
2012-0013
[9] D. H¨
ombe g, “A ma hema ical model o induc ion ha d-
ening including mechanical e ec s”, Nonlinea Anal.-Real
Wo ld Appl., 5, 55–90, 2004. h ps://doi.o g/10.1016/S1468-
1218(03)00017-8
[10] D. H¨
ombe g and W. Weiss, “PID con ol o lase su ace ha d-
ening o s eel”, IEEE T ans. Con ol Sys . Technol., 14, No. 5,
896–904, 2006. h ps://doi.o g/10.1109/TCST.2006.879978
[11] D. P. Kois inen, R. E. Ma bu ge , “A gene al equa ion p e-
sc ibing he ex en o he aus eni e-ma ensi e ans o ma-
ion in pu e i on-ca bon alloys and plain ca bon s eels”,
Ac a Me all., 7, 59–60, 1959. h ps://doi.o g/10.1016/0001-
6160(59)90170-1
[12] G. K auss, “S eels: Hea T ea men and P ocessing P inci-
ples”, ASM In e na ional, USA, 2000.
[13] J. B. Leblond and J. De aux, “A new kine ic model o
aniso he mal me allu gical ans o ma ions in s eels includ-
ing e ec o aus eni e g ain size”, Ac a Me all., 32, No. 1, 137–
146, 1984. h ps://doi.o g/10.1016/0001-6160(84)90211-6
[14] F. J. Pena B age, “Con ibuci´
on al modelado ma em´
a ico de al-
gunos p oblemas en la me alu gia del silicio”, Ph. hesis, Uni-
e sidade de San iago de Compos ela, 2003.
[15] H. M. Yin, “Regula i y o weak solu ion o Maxwell’s equa-
ions and applica ions o mic owa e hea ing”, J. Di e . Equ.,
200, 137-161, 2004.
h ps://doi.o g/10.1016/j.jde.2004.01.010
www.as esj.com 62