Impo ance o he expe imen al in es iga ion o
a concas ing echnology
Jose Š ě ina1, F an išek Ka ička1,*, Ja osla Ka olický1, Tomáš Maude 1, and Lubomí
Klimeš1
1B no Uni e si y o Technology, Facul y o Mechanical Enginee ing, Czech Republic
Abs ac . The solidi ica ion and cooling o a con inuously cas bille , slab
o cylinde , gene ally o a concas ing and he simul aneous hea ing o he
mold is a e y complica ed p oblem o h ee-dimensional (3D) ansien
hea and mass ans e . The sol ing o such a p oblem is impossible
wi hou nume ical models o he empe a u e ield o he concas ing i sel
which i is being p ocessed h ough he concas ing machine (cas e ).
The applica ion o he nume ical model equi es sys ema ic
expe imen a ion and measu emen o ope a ional pa ame e s on a eal
cas e as well as in he labo a o y. The measu emen esul s, especially
empe a u es, se e no only o he e i ica ion o he exac ness o he
model, bu mainly o op ímiza ion o he p ocess p ocedu e: eal p ocess
inpu da a nume ical analyses op imiza ion co ec ion o eal
p ocess. The mos impo an pa o he in es iga ion is he measu emen o
he empe a u es in he walls o he mold and he su ace o he slab in he
zones o seconda y and e ia y cooling.
1 In oduc ion
The p oduc ion o s eels, alloys and me allu gical p oduc s in gene al is cons an ly
de eloping. Ma e ials wi h high u ili y pa ame e s a e mo e in demand and adi ional
p oduc ion is being eplaced by highe quali y s eel. Mo e and mo e sophis ica ed
agg ega es using mo e sophis ica ed echnological p ocedu es a e being implemen ed.
In o de o main ain compe i i eness, di e si y p oduc ion and expand o o he ma ke s, i is
necessa y o moni o echnological de elopmen .
In he case o concas ing, i is no possible o ul ill hese equi emen s wi hou he
applica ion o models o all cas e p ocesses dependen on he mal-mechanical
ela ionships. The success ul applica ion o a nume ical model ( o example o a s eel slab)
and i s con inuous imp o emen is necessa ily condi ioned by ca ying ou demanding
expe imen s and hei e alua ion [1-3]. The measu ed pa ame e s se e ei he as di ec
inpu da a o he nume ical model o hey a e used o he c ea ion o his inpu da a. This
is, o example, he case o de e mining he hea ans e coe icien s benea h he cooling
je s ha has o be ca ied ou in he labo a o y (see below). The ange o he measu emen s
ca ied ou and applied was ela i ely conside able. They can he e o e be di ided in o
* Co esponding au ho : ka icka@ me. u b .cz
© The Au ho s, published by EDP Sciences. This is an open access a icle dis ibu ed unde he e ms o he C ea i e Commons
A ibu ion License 4.0 (h p://c ea i ecommons.o g/licenses/by/4.0/).
MATEC Web o Con e ences 168, 07009 (2018) h ps://doi.o g/10.1051/ma eccon /201816807009
XXI. AEaNMiFMaE-2018
expe imen al esea ch ca ied ou on an ope a ional de ice (cas e ) and expe imen al
esea ch ca ied ou in labo a o y condi ions [4]. Resea ch conduc ed on he ope a ional
de ice comp ises pe manen moni o ing o he necessa y physical and echnological
quan i ies du ing he en i e p ocess. Some o hem a e pa o a so-called b eakou sys em.
This en ails ela i ely ex ensi e iles o alues con aining indi idual da a as well as ime
eco ds – o example empe a u e in he walls o he mould, su ace empe a u es o he
slab in he place o he bend, he pou ing empe a u e, he cas ing speed, he empe a u e o
he cooling wa e , e c. Fu he mo e, o he expe imen s we e also conduc ed o he
pu poses o esea ch. This includes, o example, measu emen o he su ace empe a u es
in o he places o he seconda y-cooling zone (i.e. he cage), a e exi ing he cage (in he
so-called e ia y-cooling zone) [5] and mo e.
2 An o iginal o -line model o he empe a u e ield o he
concas s eel slab 1 530 × 250 mm (wid h × hickness)
The 3D model had i s been designed as an o -line e sion and la e as an on-line e sion
so ha i could wo k in eal ime. The nume ical model akes in o accoun he empe a u e
ield o he en i e slab ( om he meniscus o he le el o he mel in he mould o he
cu ing o ch) using a 3D mesh con aining mo e han a million nodal poin s.
The solidi ica ion and cooling o a concas slab 1 530 x 250 mm is a global p oblem o
3D ansien hea and mass ans e . I hea conduc ion wi hin he hea ans e in his
sys em is decisi e, he p ocess is desc ibed by he Fou ie -Ki chho equa ion. I desc ibes
he empe a u e ield o he solidi ying slab in all h ee o i s s a es: a he empe a u es
abo e he liquidus (i.e. he mel ), wi hin he in e al be ween he liquidus and solidus
(i.e. in he mushy zone) and a he empe a u es below he solidus (i.e. he solid s a e).
In o de o sol e hese i is con enien o use he explici nume ical me hod o ini e
di e ences. Nume ical simula ion o he elease o la en hea s o phase o s uc u al
changes is ca ied ou by in oducing he en halpy unc ion dependen on empe a u e T,
p e e ably in he o m o en halpy ela ed o uni olume H . The la en hea s a e con ained
he e. A e he au oma ed gene a ion o he mesh (p e-p ocessing) ies on he en y o he
he mophysical ma e ial p ope ies o he in es iga ed sys em, including hei dependence
on empe a u e – in he o m o ables o using polynomials. They a e namely he hea
conduc i i y k, he speci ic hea capaci y c and densi y
o he cas me al.
The empe a u e dis ibu ion in he slabs desc ibed by he en halpy balance equa ion.
The simpli ied equa ion, sui able o applica ion on adial-cas e s wi h a g ea adius, whe e
only he speed (o he mo emen o he slab) componen w in he z-di ec ion is conside ed,
is:
z
H
w
z
T
y
T
x
T
k
τ
H
ν
2
2
2
2
2
2
ν
(1)
En halpy H as a he modynamic unc ion o empe a u e mus be known o each speci ic
s eel. I is dependen on he composi ion o he s eel and on he a e o cooling.
The expe imen was conduc ed on a 1530 x 250 mm s eel slab, whe e he pou ing
condi ions we e cha ac e ised by he empe a u e in he undish (1 550 C), he empe a u e
o he liquid (1 521 C), and he shi a e o he slab (0.82 m.min-1). The liquidus and
solidus empe a u es a e de i ed om he composi ion o he s eel.
The exac ness o he p esen ed nume ical model depends no only on he spa ial and
empo al disc e iza ion, bu also on he accu acy wi h which he he mophysical p ope ies
o he ma e ials o all pa s o he sys em a e de e mined. I also depends on he de i a ion
o bounda y condi ions, i.e. he alues o he hea ans e coe icien on all cas e
bounda ies.
2
MATEC Web o Con e ences 168, 07009 (2018) h ps://doi.o g/10.1051/ma eccon /201816807009
XXI. AEaNMiFMaE-2018
expe imen al esea ch ca ied ou on an ope a ional de ice (cas e ) and expe imen al
esea ch ca ied ou in labo a o y condi ions [4]. Resea ch conduc ed on he ope a ional
de ice comp ises pe manen moni o ing o he necessa y physical and echnological
quan i ies du ing he en i e p ocess. Some o hem a e pa o a so-called b eakou sys em.
This en ails ela i ely ex ensi e iles o alues con aining indi idual da a as well as ime
eco ds – o example empe a u e in he walls o he mould, su ace empe a u es o he
slab in he place o he bend, he pou ing empe a u e, he cas ing speed, he empe a u e o
he cooling wa e , e c. Fu he mo e, o he expe imen s we e also conduc ed o he
pu poses o esea ch. This includes, o example, measu emen o he su ace empe a u es
in o he places o he seconda y-cooling zone (i.e. he cage), a e exi ing he cage (in he
so-called e ia y-cooling zone) [5] and mo e.
2 An o iginal o -line model o he empe a u e ield o he
concas s eel slab 1 530 × 250 mm (wid h × hickness)
The 3D model had i s been designed as an o -line e sion and la e as an on-line e sion
so ha i could wo k in eal ime. The nume ical model akes in o accoun he empe a u e
ield o he en i e slab ( om he meniscus o he le el o he mel in he mould o he
cu ing o ch) using a 3D mesh con aining mo e han a million nodal poin s.
The solidi ica ion and cooling o a concas slab 1 530 x 250 mm is a global p oblem o
3D ansien hea and mass ans e . I hea conduc ion wi hin he hea ans e in his
sys em is decisi e, he p ocess is desc ibed by he Fou ie -Ki chho equa ion. I desc ibes
he empe a u e ield o he solidi ying slab in all h ee o i s s a es: a he empe a u es
abo e he liquidus (i.e. he mel ), wi hin he in e al be ween he liquidus and solidus
(i.e. in he mushy zone) and a he empe a u es below he solidus (i.e. he solid s a e).
In o de o sol e hese i is con enien o use he explici nume ical me hod o ini e
di e ences. Nume ical simula ion o he elease o la en hea s o phase o s uc u al
changes is ca ied ou by in oducing he en halpy unc ion dependen on empe a u e T,
p e e ably in he o m o en halpy ela ed o uni olume H . The la en hea s a e con ained
he e. A e he au oma ed gene a ion o he mesh (p e-p ocessing) ies on he en y o he
he mophysical ma e ial p ope ies o he in es iga ed sys em, including hei dependence
on empe a u e – in he o m o ables o using polynomials. They a e namely he hea
conduc i i y k, he speci ic hea capaci y c and densi y
o he cas me al.
The empe a u e dis ibu ion in he slabs desc ibed by he en halpy balance equa ion.
The simpli ied equa ion, sui able o applica ion on adial-cas e s wi h a g ea adius, whe e
only he speed (o he mo emen o he slab) componen w in he z-di ec ion is conside ed,
is:
z
H
w
z
T
y
T
x
T
k
τ
H
ν
2
2
2
2
2
2
ν
(1)
En halpy H as a he modynamic unc ion o empe a u e mus be known o each speci ic
s eel. I is dependen on he composi ion o he s eel and on he a e o cooling.
The expe imen was conduc ed on a 1530 x 250 mm s eel slab, whe e he pou ing
condi ions we e cha ac e ised by he empe a u e in he undish (1 550 C), he empe a u e
o he liquid (1 521 C), and he shi a e o he slab (0.82 m.min-1). The liquidus and
solidus empe a u es a e de i ed om he composi ion o he s eel.
The exac ness o he p esen ed nume ical model depends no only on he spa ial and
empo al disc e iza ion, bu also on he accu acy wi h which he he mophysical p ope ies
o he ma e ials o all pa s o he sys em a e de e mined. I also depends on he de i a ion
o bounda y condi ions, i.e. he alues o he hea ans e coe icien on all cas e
bounda ies.
3 Expe imen al measu emen o he applica ion o he model o
he empe a u e ield o a s eel slab 1 530 × 250 mm
3.1 Measu ing empe a u es in he mould (c ys allize ) and in he seconda y
cooling zone
Figu e 1 illus a es he posi ions o 44 he mo-couples in wo lines along all walls o he
mould [6] (so called he b eakou sys em). The esul s o hese measu emen s a e indica ed
in Figu e 2, oge he wi h he cou se o he mean alue and dis ibu ion, and a e en e ed
in o he nume ical model o he mould. Figu e 2 shows an example o eal immedia e
empe a u es on he mould.
In o de o make he model o he empe a u e ield o he mould mo e accu a e and o
e i y i , i is use ul o ca y ou expe imen al measu emen o he su ace empe a u es
immedia ely below he mould, i.e. in he a ea o he e aining olle s. The measu emen is
ca ied ou using op ical py ome e s posi ioned a he cen e o he slab, hal -way be ween
he cen e and he edge and nea he co ne (Figu e 1). This measu emen also makes i
possible o assess whe he he empe a u e ield is symme ical along he axis o he c oss-
sec ion. Figu e 3 shows he measu emen o he su ace empe a u es along he la ge adius
in he a ea o he e aining olle s. The g aph illus a es ha i is necessa y o p ocess he
da a be o e i is used, i.e. i is necessa y o ind a sui able me hod o il e ing. The gene ally
used me hod o il e ing is loa ing a e aging. I s disad an age is ha la ge de ia ions a e
sp ead along a la ge a ea. The median ecu si e il e b ings he s onges il a ion – i s
ou pu is, o example, usable o he egula ion o he seconda y cooling. I is possible o
conside ha in e e ence o he signal om he py ome e is caused by a quali y su ace o
he slab (scales and solidi ied cas ing powde ). Tha is why he maximal alues p o ided by
he maximum il e a e conside ed as co ec [7].
The expe imen al empe a u es could be used om he b eakou sys em (and no he
da a om he py ome e s), which, wi h espec o he densely a anged he mocouples in
wo ho izon al planes (Figu e 1) co e s he asymme y be e . In o de o apply he dynamic
model o he empe a u e ield, he in es iga ion o only one hal o he slab is su icien .
Figu e 3 illus a es he a ou able compa ison o empe a u es measu ed by h ee
py ome e s wi h a calcula ed cu e o he su ace empe a u es o he slab a e 4 000 s.
In o de o make he model o he empe a u e ield mo e accu a e and o e i y i , i was
necessa y o use wo pe manen ly ins alled py ome e s wi h bo h measu ing he uppe
su ace o he slab along he small adius. The i s was ins alled by Mannesmann Demag
al eady a he s age o s a ing up he cas e . I is posi ioned jus be o e he unbending poin .
This py ome e was o iginally used o checking he unc ion o he p ocess seconda y
cooling. Du ing he in eg a ion o he on-line model in o he sys em, i is now used o
compa ing he empe a u es calcula ed by he model wi h he eal ones. Fu he mo e, i was
joined by ano he py ome e o he same ype a he exi om he cage. This a angemen o
bo h models also includes a p ocessing uni ha il e s he measu ed da a. This acili a es
hei u ilisa ion because he so wa e o he empe a u e model no longe has o deal wi h
he il e ing o he signal. The py ome e s measu e wi hin he ange om 750 o 1 200 °C
and in he case o a measu emen ailu e i shows a empe a u e o 700 °C. The posi ioning
o he py ome e s in he wo men ioned poin s is necessa y o he pu poses o checking.
The empe a u e o he slab in he poin o he i s py ome e a he beginning o he
unbending (poin ) mus be highe han he empe a u e p esc ibed o his poin (i.e. a slab
ha is oo cold mus no be s aigh ened ou ). On he o he hand, he empe a u e a he exi
o he cage should be lowe han he p esc ibed empe a u e.
3
MATEC Web o Con e ences 168, 07009 (2018) h ps://doi.o g/10.1051/ma eccon /201816807009
XXI. AEaNMiFMaE-2018
Fig. 1. A angemen o measu emen senso s.
Fig. 2. Immedia e empe a u es o he b eakou sys em ( ed – uppe line, yellow – bo om line).
Fig. 3. A compa ison o he cou se o he calcula ed and measu ed empe a u es a e 4 000 s.
3.2 Measu ing o cooling e ec o nozzles
A eal cas e con ains a o al o 8 ypes o ai -je s and geome ical layou s. Since i is no
possible o de e mine he in ensi y o he ai -wa e je s on an ac ual cas e , i is necessa y o
4
MATEC Web o Con e ences 168, 07009 (2018) h ps://doi.o g/10.1051/ma eccon /201816807009
XXI. AEaNMiFMaE-2018
Fig. 1. A angemen o measu emen senso s.
Fig. 2. Immedia e empe a u es o he b eakou sys em ( ed – uppe line, yellow – bo om line).
Fig. 3. A compa ison o he cou se o he calcula ed and measu ed empe a u es a e 4 000 s.
3.2 Measu ing o cooling e ec o nozzles
A eal cas e con ains a o al o 8 ypes o ai -je s and geome ical layou s. Since i is no
possible o de e mine he in ensi y o he ai -wa e je s on an ac ual cas e , i is necessa y o
ans e he in es iga ion – o each je indi idually – o he expe imen al labo a o y de ice
(Figu e 4), which is capable o simula ing he su ace o a concas slab [8-11]. This de ice
also allows he measu emen o empe a u es benea h he su ace wi hin he slab.
The empe a u es measu ed a e con e ed o cooling in ensi ies by means o an in e se
ask, which, in u n, a e con e ed o he cou ses o he hea ans e coe icien s using an
expanded nume ical model. The nume ical model also de e mines he e ec o adia ion
which is dependen on he su ace empe a u e.
3.3 Measu ing o slab su ace empe a u es in he e ia y cooling zone
Th ee 2 mm insula ed he mocouples had been used o he measu emen . The ho end o
each he mocouple was ixed inside he hollow ip o a s ud, ca e ully posi ioned on he
su ace o he slab a e cooling (Figu e 5) and all we e hamme ed in o he same dep h
du ing he o ch-cu ing p ocess [12]. The measu emen was ini ia ed on a sepa a ed slab, 5-
o-6 m om he exi o he cage, and con inued un il he slab was cold – in he cooling ield.
The co ec ion o he e o o he measu emen esul , om he hea ans e h ough he
he mocouple insula ion and he p o uding end, is conduc ed ia compa ison
measu emen s. The measu ed empe a u es o h ee su ace poin s a e in Figu e 5.
Fig. 5. Tempe a u es o he h ee su ace poin s.
Fig. 4. The labo a o y de ice in use.
5
MATEC Web o Con e ences 168, 07009 (2018) h ps://doi.o g/10.1051/ma eccon /201816807009
XXI. AEaNMiFMaE-2018
4 Conclusion
Resea ch in o he he mokine ics o solidi ica ion and cooling o concas slabs equi es
sys ema ic expe imen al measu emen on a eal cas e . I s esul s a e u ilised no only o
imp o ing he nume ical model o he empe a u e ield bu also o assessing he exac ness
o his model. This dynamic model, which wo ks non-s op in eal ime, ensu es con inuous
co ec ion o he eal p ocess o he cas e in ques ion. The main measu ed quan i ies a e
he empe a u e in he walls o he mould, he su ace empe a u es o he slab upon exi
om he mould, a he unbending poin o he slab and upon exi om he cage o he
seconda y cooling. Fu he mo e, i is he cas ing empe a u e, he cas ing speed, he
empe a u e o he cooling wa e , he me allu gical leng h, e c. A specialised labo a o y
conduc s measu emen s o he sp aying cha ac e is ics o indi idual cooling nozzles. I s
ou pu es ablishes he hea ans e coe icien s benea h each o he wa e o wa e -ai
nozzles.
This esea ch was suppo ed h ough NETME CENTRE PLUS (LO1202) by he Minis y o
Educa ion, You h and Spo s o he Czech Republic unde he „Na ional Sus ainabili y P og amme
I“.
Re e ences
1. J.K. B imacombe, Me all. and Ma e . T ans. B, 30B, 553 (1999)
2. J. Mie inen, S. Louhenkilpi, J. Laine, P oc. o Gen. COST 512 Wo kshop (1996)
3. R.A. Richa d, R. Ha ding, Ch. Becke mann, Me all. & Ma e . T ansac ions, 34B,
297 (2003)
4. J. Š ě ina, Ph.D. Thesis, TU Os a a, Czech Rep. (2007)
5. B.G.Thomas, R.J. O’Malley, D.T. S one, Model. O Cas . Weld., Solidi . P oc. VIII,
San Diego, CA, TMS (1998)
6. F.Ka ička, e al., P oc. o 4 h In e . Con . on Compu . Fluid Dyn. in he Oil, Gas,
Me all. and P oc. Ind., T ondheim, No way (2005)
7. F. Ka ička, e al., Communica ions, 4A2012, 14, 13 (2012)
8. J. Ho ský, M. Raudenský, P oc. o 14 h In e . Me al. & Ma e . Con . METAL, H adec
n/M., Czech Rep. (2005)
9. J. Ho ský, e al., AISTech, I on&S eel Techn. Con . & exp., Cha lo e, USA (2005)
10. M. Raudenský, J. Ho ský, I onm. & S eelm. 32, 2 (2005)
11. J .Š ě ina, Assoc. p o esso . & habili . hesis, TU Os a a, Czech Rep. (2008)
12. J. Š ě ina, e al., Comp. Me h. & Exp. Meas. XII, 749, WIT P ess (2005)
13. F. Ka ička e al., P oc. o 5 h Pac. Rim In e . Con . on Ad . Ma e . & and P oces.,
Zue ich, T ans Tech Pub. (2004)
6
MATEC Web o Con e ences 168, 07009 (2018) h ps://doi.o g/10.1051/ma eccon /201816807009
XXI. AEaNMiFMaE-2018
