thermal stress cracking of sliding gate platesccc.illinois.edu/s/publications/12...
TRANSCRIPT
Thermal Stress Cracking of Sliding Gate Plates
Hyoung-Jun Lee1, Seong-Mook Cho1, Seon-Hyo Kim1, Brian G. Thomas2 Sang-Woo Han3, Tae-In Chung3, Joo Choi3
1Department of Materials Science and Engineering,
Pohang University of Science and Technology, San 31, Hyoja-Dong, Nam-Gu, Pohang, Gyeongbuk, 790-784, South Korea
Phone: +82-54-279-5968 Fax: +82-54-279-2399
E-mail: [email protected], [email protected], [email protected]
2Department of Mechanical Science and Engineering, University of Illinois at Urbana-Champaign, 1206 W. Green St., Urbana, IL, 61801, USA
Phone: +1-217-333-6919 Fax: +1-217-244-6534
E-mail: [email protected]
3Steel Making Research Group, POSCO Technical Research Laboratories, POSCO
Dongchon-Dong, Nam-Gu, Pohang, Gyeongbuk, 790-785, South Korea Phone: +82-54-220-2890 Fax: +82-54-220-6832
E-mail: [email protected], [email protected], [email protected]
Key words: Refractory, Friction force, Pre-heating, Finite-element, Computational model
ABSTRACT
The sliding gate plate that controls steel flow through the tundish nozzle sometimes cracks leading to air aspiration and safety concerns. To evaluate possible mechanisms for crack formation, this research applies a 3-D finite-element model of the thermal and mechanical stress in a sliding-gate plate during preheating and casting induced by thermal expansion and/or mechanical movement. The thermal model is first validated with previous temperature histories measured during preheating and casting in a ladle plate. The model of a tundish sliding gate nozzle is then used to investigate thermal-mechanical behavior and cracking due to the temperature variations during preheating and casting. The model predictions of the maximum stress location and orientation match well with the crack location observed in used plates from POSCO. Different mechanisms for the formation of two different types of common through-thickness cracks are identified and explained. The first involves exterior tensile stress during heating stages, and the second is due to excessive compression from non-optimal placement of guide points on the steel cassette.
INTRODUCTION
Flow between the tundish and mold in continuous-steel casting is often controlled by an assembly of three refractory plates, as shown in Figure 1. The middle plate is moved to adjust the opening to control the flow rate through the nozzle to maintain a stable mold meniscus level. Cracking of the sliding gate refractories is important because it poses a great potential safety hazard, in addition to steel quality problems. Even if cracks are rare, the precautionary limits put on lifetimes and productivity to avoid potential problems
may baspirab). Thindicashow Previtheir differtempemech[1]. Apressu
be very costly.ation through the white, inverates that graphs some less-co
ous research hinherent brittle
rent stress souerature during
hanical load proAlso, friction fure due to heig
. Furthermore,the cracks. Radrted triangular-hite in the slidiommon transve
has investigatedeness, refracto
urces have receg pre-heating aovided to the pforces are geneght difference b
Plat
a) Bo
c) Schem
through-thickdial through-thshaped area ciring gate plate rse cracks.
d refractory coories are subjeceived little attand casting, th
plates and casseerated by the hbetween tundis
Bol
te
Cassette
Figure
ottom view of
matic of rare cr
Figu
kness cracks mhickness cracksrcled in red onhas been oxid
omposition effect to cracking tention. When he resulting thette by the bolthorizontal (bacsh free surface
UTNlt
e
SEN
e 1. Schematic
used plate
rack locations
ure 2. Different
may lead to incls are common,
n the fracture sudized, suggestin
ects to extend sproblems, butrefractories a
hermal stress ts, which may bk and forth) mand sliding gat
Ar Injection
c of sliding gate
b) Co
In
d) Cros
t cracks in slidi
lusion problem, such as showurface of the thng that the ste
service life andt the detailed mare subjected t
may cause crbe non-uniform
movements of tte location pro
M
In
Line Contac
e nozzle system
ommon through
ss section view
ing gate plates
ms in the steel,wn in the slide ghrough-thickneeel was likely
d achieve highmechanisms anto high heat trracks [3]. Anom and create hthe middle slid
ovides addition
Metal Band
nsulation
ct
m
h-thickness cra
O
w of common c
, due to re-oxigate plates in F
ess crack surfacoxidized as w
her productivitynd relative impransfer and raother source oigh localized sding plate. Final load.
ack
Out
crack
idation from aiFigure 2-a) andce in Figure 2-
well. Figure 2-c
y [1, 2]. Due tportance of th
apid changes inof stress is thsurface pressurnally, ferrostati
ir d d)
c)
o e n e e c
COMPUTATIONAL MODEL
To explore thermal and mechanical stress in a sliding gate plate during preheating and casting, a three-dimensional thermal-stress model was developed. Temperature, ( )T x , of the components of the sliding gate assembly is found by solving the heat-conduction equation [4].:
2 0k T! = (1)
where k is the thermal conductivity, and x are the three coordinate directions. The mechanical behavior is found by solving the differential equations of force equilibrium:
F!" =# (2)
where F if the force vector from thermal, mechanical, and ferrostatic pressure loads, ( )x# is the Cauchy stress tensor, computed by Hooke’s law of elasticity:
: el=# $C (3)
where C is the fourth-order tensor containing 81 elastic coefficient.:
( ) ( ) ( )( )2 1 1 1 2ijkl ik jl ij klE EC %& & & &
% % %= + +
+ + ' (4)
where E is Young’s modulus, 65 GPa for refractory and 206 GPa for the steel cassette; % is Poisson’s ratio, 0.2 for refractory and 0.3 for the steel cassette; and ij& is the Kronecker delta. The elastic strain tensor ( )el x$ is computed from an additive decomposition of the strains:
el th= '$ $ $ (5) where ( )x$ is the total strain tensor, computed from the gradient of the displacement field ( )xu :
( )( )12
T$ = ! + !u u (6)
and ( )th x$ is the thermal strain tensor, calculated based on the coefficient of thermal expansion ( and reference temperature 0T :
( )0
th T T$ (= ' I (7) where ij&I = is the second-order identity tensor, and ( is thermal expansion coefficient, 8.2 × 10-6 °C-1 for refractory. For the thermal problem, the flame and molten steel heat convection boundary conditions in the ladle and tundish sliding gate nozzle:
( )k T h T T)' ! " = 'n (8)
where ( )xn is the surface normal to outside, ( )h x is the convection heat transfer coefficient and ( )T x) is the sink temperature. During the preheating stage, internal gas flame temperature and heat transfer coefficient are needed. Based on liquefied natural gas (LNG), containing 88% Methane (CH4), 5% Ethane (C2H6), 5% Propane (C3H8) and 2 % Butane (C4H6) [5], and stoichiometric air, a flame temperature model [6] gives a gas temperature of 1518 °C with no excess air. However, the flames do not directly touch the sliding gate plates and there are heat losses due to excess air entrainment, so the internal gas temperature was assumed to be 750 °C. During preheating, the heat transfer coefficient for free heat convection from the exterior of the cylinder-shaped nozzle to atmosphere is given by the relation for turbulent flow by Churchill et al. [7]:
21 6
8 270.3870.825
[1 (0.492 Pr) ]D
RaNu ½° °= +® ¾
+° °¯ ¿ 2
Nu khr"= (9)
where Nu is Nusselt number, Ra is Rayleigh number, Pr is Prandtl number, h is free heat transfer coefficient, k is thermal conductivity and r is nozzle outside radius, of 0.225 m. This gives a free heat transfer coefficient (ho,preheat) of 8.8 W/m2·K. During preheating, the heat transfer coefficient for forced heat convection from the turbulent flowing combustion products to the contact surfaces on the nozzle interior are given by an empirical equation for smooth cylinders by Petukhov et al.[8]:
1 2 2 3
[( 8) Re Pr][1.07 12.7( 8) (Pr 1)]
DfNuf
=+ ' 2
Nu khr"= (10)
where ReD is the Reynolds number and f is Darcy friction factor. This gives hi,preheat of 65 W/m2·K During the steel casting stage, the combustion gases are replaced by molten steel, flowing through the sliding gate bore at 1550°C. The forced convection relation for turbulent metal flow from Sleicher and Rouse equation[9] is used: 2Re ur
*=
Pr *
(=
0.240.884 Pr
a = '+
1 0.5exp( 0.6Pr)3
b = + '
5 0.015Re Pra bNu = + 2Nu kh
r"= (11)
where u is an average flow velocity in cross sectional area of cylinder, % is kinematic viscosity of molten steel and a and b vary with Nusselt number. This gives hi,steel of 28.7 kW/m2·K. The heat conduction boundary conditions of the sliding plates are:
k T q' ! " = "n n (12) where ( )q x is the heat flux vector. The boundary conditions of the radiation heat loss from the heated sliding gate plates:
( )4 4surk T T T$#' ! = ' (13)
where $ is the emissivity of materials and # is the Stefan-Boltzmann constant. Mechanical loading is applied by constraining the horizontal displacements of the refractory plates where they touch the guide points on the steel cassette. Additional mechanical loading pressure in the vertical direction is provided by tension in the bolts where they contact the plates, but this was ignored in the current work. Ferrostatic pressure is imposed on the inside refractory surfaces exposed to the flowing steel:
pF gh+= (14)
where h is height difference between the tundish free surface and the sliding gate location, 1.8 m, + is molten steel density, 7020 kg/m3, g is gravitational acceleration 9.81 m/s2 and pF is the resulting average ferrostatic pressure, 0.124 MPa. Relative to the typical stress from thermal expansion, E T(' " ", , which for 1000 °C temperature change is ~540 MPa, this ferrostatic pressure is negligible and so can be neglected. The above equations were solved using the finite-element method with the commercial software ABAQUS 6.9-1. The heat transfer model used standard linear three-dimensional wedge-shaped (DC3D6) 6-node brick elements for the sliding plates, hexahedral (DC3D8) 8-node brick elements for the steel band. The stress model used wedge-shaped (C3D6) 6-node linear brick elements for the sliding plates and hexahedral (C3D8R) 8-node linear brick elements for the steel band. This linear thermal and stress problem required about 10 hours to solve on a computer with an 8-core 2.99 GHz Intel Xeon Processor and 16 GB of RAM.
The mrepresambiethe ouuser fcastinof 3-D The ttwenteverytime s
To evcondu
Figur
Figur
model is first vsenting preheaent temperaturutside (ambienfriendly spreadng, that was vaD elements wit
two solutions mty elements bey one second anstep sizes were
valuate the chouction in a lad
t = 120 miTinitial = 20 ° hi = 70 W/mTi = 600 °C *Other surfa
re 3. Schematitest probl
re 5. Geometrycouples lo
HEAT TR
validated by sating of a typicres are all 20 °nt air). The modsheet-based toalidated previouth ABAQUS, w
match very clotween inside and every hunde used for the s
HE
oice of heat trale-nozzle plate
in C
10�
m2·K
aces are insulat
ic of boundarylem
y and thermo-ocation [10]
RANSFER M
olving a simplcal nozzle walC; the heat tra
odel geometry, ool that modelusly [6]. The cwith all surface
osely, as shownand outside of tdred seconds alsliding gate noz
EAT TRANSFE
ansfer boundare, and the resu
ho = 20 To = 20
40 mm
k = 20 W ȡ = 2460 Cp = 150
ted
y condition an
Pa
Ladle
Figure 6
MODEL VALID
le one-dimensill. A schematic
ansfer coefficieproperties and
ls heat transfercurrent model ses insulated exc
n in Figure 4. Tthe nozzle is elso match exaczzle model.
ER MODEL V
ry conditions, lts were comp
W/m2·K 0 °C
W/mK 0 kg/m3 00 J/kgK
nd properties o
art Plate (
6. Finite elemeladle plate
10
20
30
40
Tem
pera
ture
(Deg
ree
C)
DATION WIT
ional transientc of the test pents are 70 W/md constants arer in submergedsimulated a wecept for the thi
The solutions enough for gooctly, so that a h
VALIDATION
the heat transared with mea
of Figure 4
Element(Wedge) 92
ent mesh and d
0 200
0
0
0
0
TH ANALYTI
t heat transfer problem is shom2·K for the in
e given in Figud entry nozzleedge-shaped poin inner and ou
with twenty anod spatial resolhundred secon
N WITH EXP
fer model wasasurements by
4. Temperature
2,378
details of Fi
40 60Time (m
ICAL SOLUT
problem of coown in Figure nside (combus
ure 3. The exaces during the portion of the n
uter boundary s
nd forty elemelution. The sol
nds time step is
PERIMENTS
s next applied K. V. Simonov
variation of A
hi,preheat = 65 Whi,steel = 29 ൈ İ = 0
ho,preheat = 7 Who,steel = 7 W/mİ = 0.92
igure 7. Bound
80 100min.)
20 Elements-In 20 Elements-Ex 40 Elements-In 40 Elements-Ex Every 1 Sec.-In Every 1 Sec.-E Every 100 Sec. Every 100 Sec. Singh VBA-Int Singh VBA-Ex
TION
onduction thro3. Initial, insid
stion gas) and ct solution waspreheating, coonozzle wall withsurfaces labeled
ents match, thuutions with tims reasonable. T
to predict tranv et al. during
ABAQUS and S
W/m2·K, Ti,prehe103 W/m2·K, T
W/m2·K, To,prehem2·K, To,steel =
dary conditions
120
nternal Temp.xternal Temp.
nternal Temp.xternal Temp.nternal Temp.
External Temp..-Internal Temp..-External Temp.ternal Temp.xternal Temp.
ough a cylinderde, and outsid20 W/m2·K fos found using oling down andh a single layed with arrows.
us showing thame step sizes oThese mesh an
nsient 3-D heatop teeming o
S. VBA model
Insulatedh = 0 İ = 0
eat = 750 °C Ti,steel = 1590 °C
eat = 25 °C 150 °C
s of ladle plate
r, e
or a d
er
at of d
at of
d
C
molten steel into the mold [10]. Temperatures were measured with four thermocouples installed on the non-working side of the upper plate, and plates were subjected to a preheating thermal treatment at 170 °C. The geometry and thermocouple locations are shown in Figure 5. Thermocouples are aligned along the plate symmetry plane to measure temperature variation with distance from the inner bore of the plate. A wedge shaped finite element mesh is designed as shown in Figure 6. During the preheating stage, the ladle plate is heated from 25°C with an ambient temperature of 25°C and inside gas temperature of 750 °C, followed by steel casting at 1590 °C, using the same convection coefficients as proposed for the sliding gate nozzle model, and given in Table I. The preheating time was estimated to be 50 minutes, based on when thermocouple number one reached 150°C. Half of the ladle-nozzle plate is simulated, so the symmetry plane is insulated. The contact surfaces between the upper surface of the upper plate and the upper tundish nozzle, the lower surface of the upper plate and the upper surface of the lower plate are also insulated, because it is assumed that the heat exchanged between these surface pairs is negligible (Figure 7). The thermal properties [11] and constants for this ladle plate validation problem are given in Table I. The upper ladle plate temperature contours predicted by the current ABAQUS model after ninety minutes are shown in Figure 8. Temperature in the plate decreases radially from the plate inner bore to the outside surface with a large, nonlinear temperature gradient due to the large heat capacity of the refractory. Experimental measured and predicted ladle plate temperature histories are compared in Figure 9. Temperature (°C)
Preheating Molten steel pouring
Figure 8. Ladle plate temperature contour, 90 min Figure 9. Comparison of measured and predicted temperature in ladle plate During the forty minutes of casting where thermocouples were recording, the experimental and predicted temperatures in the ladle plate match well with the model simulation, especially near the plate bore. This suggests that assuming an internal gas temperature of 750 °C during preheating is reasonable, in addition to the heat transfer coefficients used for the tundish sliding gate nozzle application.
APPLICATION TO TUNDISH SLIDING GATE NOZZLE PLATES
The half of the symmetrical tundish sliding gate nozzle system assembly simulated in this work is shown in Figure 10. The upper and lower cassette and three sliding-gate plates are all clamped together by four bolts. Each of the plates is surrounded with a steel band and is restrained within the cassette by guide points. The middle plate moves back and forth between such guides which are connected to a hydraulic cylinder to control the molten steel flow. Details of the finite element mesh are shown in Figure 11 and Table II. Three plates and three steel bands are modeled and number of the total elements is 52,871. Starting initially at ambient temperature, all of the parts are preheated to 750 °C at 100 percent (fully open) for 3.5 hours. Then, the middle plate is moved to zero percent open at 25 mm/sec, taking 5 seconds and held for 12 more minutes while steel fills the tundish.
Table I. Properties[11] and constants for ladle plate model validation problem
Property or Constant Value Initial nozzle temp., Tinitial 25 °C Initial gas temp., Ti,preheat 750 °C Internal conv. heat transfer coeff., hi,preheat 65.24 W/m2·K External ambient temp., To,preheat 25 °C External conv. heat transfer coeff., ho,preheat 7 W/m2·K Molten steel temp., Ti,steel 1590 °C Internal conv. heat transfer coeff., hi,steel 2.87ൈ104 W/m2·K External ambient temp., To,steel 150 °C External conv. heat transfer coeff., ho,steel 7 W/m2·K Density, ȡ 3200 kg/m3 Thermal conductivity, k 8.26 W/m·K Specific heat, Cp 1004.64 J/kg·°C Stefan-Boltzmann const., ı 5.37ൈ10-8 W/m2·K4 Emissivity[12], İ 0.92
0 20 40 60 80 1000
200
400
600
800 Measured 1-a Measured 1-b Measured 2-a Measured 2-b Measured 3 Measured 4-a Measured 4-b Predicted 1 Predicted 2-a Predicted 2-b Predicted 3 Predicted 4-a Predicted 4-bTe
mpe
ratu
re (D
egre
e C
)
Time (min.)
Then,the co
After replacmode The bcastinthe toW/mW/mFigurupper(SENzero. the cdirectblue rthe x FigurdurinTempoutsidespecdiffergeneris cooutsidnear tensiogenerpointstensil
, the middle plontinuous casti
Figure
r 3.5 hours casced in the castel. The properti
boundary condng stages respeop of the uppe·K and betwe·K. For the sres 12 and 13r tundish nozz
N) so the displGrey regions
assette and pltion displacemregions) touchand y-direction
re 14 and 15 sng the preheperature in thede surface to cially during rences betweenrates great theronstrained by de of the platethe inner bor
on hoop stressral expansion s, generating le hoop stress.
late is moved aing process are
10. Schematic
sting, the openter every 3.5 hies and constan
ditions of the hectively. The bor plate, bottomeen each platestress model, , the purple rzle (UTN) anlacement of z-are the contac
lates, which arment. The side
the UTN and ns.
show contoursating and co
e plates increathe inner bothe casting
n inside and rmal expansionthe small am
e. This generatre, which is s towards the
of the plate radial compre
again, increasine neglected.
c of tundish slid
ning is moved hours. The heants are shown i
heat transfer aroundaries inclu
m of the lower e and its steeillustrated on regions touch d submerged -direction is ccting guide-pore constrainedsurfaces of theSEN, so are c
s of temperatuontinuous cas
ases sharply frre in the radstage, when outside are g
n near the platemount of expa
es compressivbalanced by outside. At thpushes again
ession, and ex
ng the opening
ding gate nozz
back to 0 percat transfer modin Table III.
re illustrated inude the inner bplate and sym
el band is 7 the right of the cassette, entry nozzle onstrained to
oints between d in x and y-e plates (dark constrained in
ure and stress sting stages. rom the plate ial direction,
temperature greatest. This e bore, which ansion in the e hoop stress simultaneous e same time,
nst the guide xtra localized
to 60 percent,
le assembly
cent. For safetdel is run for a
n the left side ore (red region
mmetry plane (v
Table IIP
Initial nozzlInitial gas teInternal conExternal amExternal conMolten steelInternal conExternal amExternal conDensity of rThermal conSpecific heaEmissivity oDensity of sThermal conSpecific heaEmissivity oStefan-Boltz
, for steady cas
Fig
Table II. Par
Upper PlaMiddle PlLower PlaUpper BanMiddle BaLower BaTotal
ty and productall of these pro
of Figures 12ns), outer surfaviolet regions)
II. Properties[1Property or Conle temp., Tinitial
emp., Ti,preheat nv. heat transfermbient temp., Tnv. heat transfel temp., Ti,steel
nv. heat transfermbient temp., Tnv. heat transferefractory, ȡref
nductivity of reat of refractory,of refractory[12steel, ȡsteel nductivity of stat of steel, Cp,st
of steel[12], İst
zmann const., ı
sting. The mult
gure 11. Finite
Computationartate (Wlate (Wate (W
and and
and
t quality reasoocess steps, fo
2 and 13, for thaces (blue regio). The gap con
11] and constannstant
r coeff., hi,prehea
o,preheat er coeff., ho,preh
r coeff., hi,steel o,steel er coeff., ho,steel
efractory, kref y, Cp,ref 2], İref
teel, ksteel teel teel ı
tiple small mov
element mesh
l model mesh dElements
Wedge) Wedge) Wedge)
(Hex.) (Hex.) (Hex.)
ons, the slidingollowed by run
he preheating ons), and insulnductance betw
nts for sliding g
7at 65.
heat 15
28719.1
l 328.
1004.0.7848
4180.
5.37 ൈ 10
vements during
details
15,133 18,008 18,782
512 512 384
52,871
g gate nozzle inning the stres
and continuouated surfaces a
ween plates is
gate model Value 25 °C 50 °C 24 W/m2·K25 °C
7 W/m2·K50 °C 63 W/m2·K50 °C 7 W/m2·K
00 kg/m3 26 W/m·K 64 J/kg·°C 92 60 kg/m3
8.6 W/m·K 8.6 J/kg·°C 75 0-8 W/m2·K4
g
is s
us at 0
The pcommthe loconfirtensiostrengthe ou
Temp.(°C)
Figur
predicted princmon through-thocation of maxrms one mechaon in the exterigth of the refrautside to the in
Insuh = 0 h0 = T0 =İ = 0
Insuh = 0
h0 = T0 =İ = 0
re 14. Thermostage
cipal stresses arhickness crackimum tensile sanism of crack ior of the refraactory. The cranner bore. Ten
ulated 0, İ = 0
7 W/m2·K 25°C
0.92
Figure 12
ulated 0, İ = 0
7 W/m2·K 150°C
0.92
Figure 13. B
S(
o and mechan
re shown in a s below. The rstress on the pl
formation. Thctory plate. Th
acks likely formnsile strength o
*Gap condu
= 7 W/m·K
h = 65 W Ti = 750 İ = 0
T = 25 °C İ = 0.75
2. Boundary co
*Gap cond
= 7 W/m·K
h = 29 ൈ Ti = 1550 İ = 0
T = 150 °C İ = 0.75
oundary condi
Stress (Pa)
nical behavior
top-view of thresults suggestate exterior at
hermal expansiohe maximum prmed at these poof the refractor
uctance K
W/m2·K °C
*Gap conducta = 0 W/m·K
onditions of he
uctance K
103 W/m2·K 0 °C
*Gap conducta = 0 W/m·K
tions of heat tr
in preheating
he middle platet two different the location ofon during heatrincipal hoop soints and propary is not availa
nce
TounozzUX =
eat transfer and
ance
TouwithUX =
ransfer and stre
Temp. (°C)
g Figure 15.
e in Figures 16cracking mec
f common obseting causes comstress near the agated inwardsable, owing to
ching with zle = 0, UY = 0
Y Sy
Touching witUX = 0, UY =
d stress model i
ching h nozzle = 0, UY = 0
Y S
Touching witUX = 0, UY =
ess model in co
Thermo and mstage
6 and 17, and cchanisms. The erved through-
mpression nearthickest part o
s, perpendiculaits high brittle
ymmetry UZ
th cassette 0
in preheating s
Symmetry
th cassette 0
ontinuous casti
Stress (Pa)
mechanical beh
compared withexcellent agre
-thickness cracr the inner boreof the plate excar to the stress eness. Refracto
= 0
stage
UZ = 0
ing stage
havior in contin
h the location oeement betweencks in Figure 1e, and balancingceeds the tensil
direction, fromory hand book
nuous casting
of n 6 g e
m ks
suggeonly the ca A secsurfacThis certailarge geom
SDist(DirMa
Phoof u
This castinusingand wdurinmodesuggeare thdesigcassepreveinves
The ain theKwangratef
est that the tens10 minutes of pasting is even h
cond common ce in Figure 17second crackinin, however. Ittensile stress,
metry design an
Stress tribution rection & agnitude)
otograph used plate
TStൎ
Figu
paper investigng, using a thrg previous modwith previous eng the preheatinel predictions oests two differehe first importan of the cassette is an impor
ent cracks due tigate crack fo
authors wish toe Clean Steel angyang works fully acknowle
sile strength is preheating. Ashigher.
location of thr7, where the png mechanismt is possible th
and accompand clamp forces
MiddPreh
Tensile trength 12 MPa
ure 16. Commomechan
ates the thermree-dimensionadels and theoryexperimental teng stage, the tof the maximument cracking mant mechanism tte appears to rtant part of thto excessive crmation.
o thank the Conand Non-ferrou
and Kwon-Medged.
about 5 percens time and temp
rough-thickneslate sides cont
m depends on thhat this highly-nying tensile cs must be optim
dle section of midheating (750 °C), 5
Cra
on through-thicnism of tensile
mal and mechanal finite-elemey. The heat-tranemperature meatensile and comm stress locati
mechanisms whicausing crackconstrain the
he model prediccompression. F
ntinuous Castinus Metals Procyung Lee – P
nt of the comprperature increa
ss cracks matctact the cassetthe plate shape-compressed recracks, during
mized to lower
dle plate, 502 sec
Crack directio
ack
ckness crack foe stress
CON
nical behavior ent model. Thensfer model anasurements of mpressive streions and orientich are both ob
ks to form in thplates to applyction of crack
Further work is
ACKNOW
ng Consortiumcessing LaboratOSCO Pohang
ressive strengthase during preh
hes with the lote guide pointse and cassette egion generatethe final coolthis compressi
on
ormation F
NCLUSIONS
of tundish slide internal gasnd its boundarya ladle-nozzle sses exceed botation match wbserved in prache plate due to ty excessive prformation, and
s needed to inc
WLEDGEME
m at the Universtory, POSTECg works. Finan
th in refractoryheating, the stre
ocation of maxs. Here, the hooguide point lo
ed excessive crling-down stagive stress to pr
Compressive Strength ൎ 245 MPa
Figure 17. Commec
ding gate nozztemperature an
y conditions ar plate. The tunoth the refract
well with the crctice. Thermal tensile stress d
ressure at the cd the geometrycorporate the e
ENTS
sity of Illinois CH. Thanks arencial support f
y materials [13]ess gradually in
ximum comprop stress excee
ocation. The timreep in comprge after castingrevent this mec
Middle section ofPreheating (750°C C
mmon through-tchanism of com
zle plates durinnd heat transfre validated windish sliding gatory tensile andrack locations expansion due
during heating cassette guide y of its guide peffects of creep
at Urbana-Chae also given to from POSCO
]. This stress isncreases. Tensi
essive stress aeds the compreme of crack fo
ression, which g. It is clear th
chanism of plat
f middle plate, C), 12,600 sec
Crack direction
Crack
thickness crackmpressive stress
ng preheating fer coefficientsith an analyticaate nozzle modd compressiveobserved in us
e to the temperstages. Secondpoints during
points should bp and residual
ampaign as weSung-Kwang (Grant No. 4.
s exceeded afteile stress during
at the refractoryessive strength
formation is nothen generate
hat guide-pointe cracking.
k formation s
and continuous are calculatedal test problemdel predicts thae strengths. Thsed plates. Thiature variationdly, geometrica
operation. Thbe optimized tostress, to bette
ell as colleagueKim – POSCO0007764.01) i
er g
y h. ot d
nt
us d
m, at e is ns al e o
er
es O is
REFERENCES 1. Munteanu, V., Steel and Refractory Chemical Interactions and Mechanical Behavior of Plates for Sliding Gate during Steel
Continuous Casting. The Annals of "Dunarea de Jos" University of Galati. Fascicle IX. Metallurgy and Materials Science, 2008. No.2-2008: p. 32-36.
2. Jayanta Chaudhuri, G.C., Satyendra Kumar, Visvanathan V. Rajgopalan, New Generation Ladle Slide Gate System for Performance Improvement. MPT International, 2007. 30(No. 6): p. 38-42.
3. Zhi-He Jin, Y.-W.M., Effects of Damage on Thermal Shock Strenth Behavior of Ceramics. Journal of American Ceramic Society, 1995. 78(7): p. 1873-1881.
4. Lance C. Hibbeler, B.G.T. Thermal Distortion of Funnel Molds. in AISTech. 2011. Indianapolis, Ind., USA. 5. Korea Gas Safety Corporation. 6. Varun Kumar Singh, B.G.T., User Friendly Model of Heat Transfer in Submerged Entry Nozzles during Preheating, Cool
Down and Casting, in Continuous Casting Consortium2010: University of Illinois at Urbana-Champaign. 7. Stuart W. Churchill, H.H.S.C., Correlating Equations for Laminar and Turbulent Free Convection from a Vertical Plate.
International Journal of Heat and Mass Transfer, 1975. 18(9): p. 1323-1329. 8. Incropera, F.P., and Dewitte, P. D., Fundamentals of Heat and Mass Transfer2002, New York: John Wiley and Sons. 9. C. A. Sleicher, M.W.R., A convenient correlation for heat transfer to constant and variable property fluids in turbulent pipe
flow. International Journal of Heat and Mass Transfer, 1975. 18(5): p. 677-683. 10. K. V. Simonov, S.I.R., A. A. Kortel', L. A. Reinov, Thermal Loading of Periclase Plates in Sliding Gates of Steel-teeming
Ladles. Refractories and Industrial Ceramics, 1979. 20: p. 742-746. 11. Chosun Refractories Co. Ltd. Research Center. 12. Monarch Instrument "Table of Emissivity". 13. Schacht, C.A., Refractories Handbook2004, Cimarron Rd., Monticello, NY 12701, USA 800-228-1160: Marcel Dekker, Inc