1

Thermal Analysis of Solidifying steel shell in Continuous Casting Process

Tran Hoang-Son, Castiaux Etienne,

Habraken Anne Marie

4th - 8th May, 2020

University of Liège

Recherche et Technologie en Wallonie DG06

2

Steel Continuous Casting

²

Real world: liquid and solid states –thermo-mechanical model should predict the steel shell behaviour

Current model: thermal approach (liquid steel is present however no fluid mechanic analysis), however temperature measurements within the mould validate the results

slag

3

Sticking Break-Out phenomenon (BO)

²

Effect of sticking event and appearance of tearing

oscilla

tion

Meniscus

Mould

Solidifiedshell

=shell

Break- Out or BO

Stickingpart

(Lee et al., 2009)

Coper mould

Mensicus

Liquid steel

solidified shell

Research objectives:

Analysis of the correlation between the mouldtp° field and the BO defect

Warn to the operator to adapt casting process features

EBDS product= tp° mould measurement by optical fibers= BO detection system called “Emerald”

http://www.ebds-engineering.com/en/Tensile force

4

Mould TP° measurement & Model challenges

²

• Multiple solid to liquid contact with the mould• Crack detection criterion crack propagation• Lagrangian code however liquid steel movement have to be implemented• Thermo-physical data and boundary conditions

Variation of temperature with time for some positions

2D Thermal Finite Element model

Broken shell in the top of the mould

CMAM Pascon, F., Habraken, A. (2007).

Tp°Data

Non calibrated tp° values just relative evolution

5

2D FEM Thermal simulation

²

2D finite element model of the continuous casting

Cooling

system

Copper

plates

MeniscusLiquid

steel

Ferrostatic

pressure

Schematic representation of CC problem

Meniscus

Mouldall elementsalwaysactive

OFF Elementsin the slab

ON Elementsin the slab

Casting

Direction

In whiteliquidsteel

Convection and

Radiation

Element

In blacksolidsteel

6

Thermal simulation for steady state

²

6

Finite Element mesh - 2D

Mold

Liquid

Steel

Solid

+

Mushy

+

Liquid

Steel

tend

700mm

Thermal properties

of low carbon steel

fromPascon, F., Habraken, A. (2007). Finite element study of the effect of some local defects on the risk of transverse crackingin continuous casting of

steel slabs. Computer Methods in Applied Mechanics and Engineering, 196, 2285-5599.

18969 elements

18846 nodes

1800mm

Menicus

Initial stateFinal state.Predictions are stationary

Thermal simulation for steady state

²

7

Speed casting

(m/min)

Convection h

(mW/mm²K )

Thermal resistance of

Contacts R (mW/mm²K )

1.3 200 2.35

1520

1507

1497

1487

1477

1470

(°C)227

182

138

93

49

27

(°C)

P-2

P-3

P-1

The chosen parameters

steady state after 50s

10 20 30 40 50 60 70 80

50

100

150

200

250

Times (s)

Tem

pera

ture

(°C

)

Times-Temperature

P-1

P-2

P-3

8

Remeshing step

Old meshNew mesh

4164 elements

4051 nodes

Thermal field at the

end of steady

simulation at 83s

Finer mesh

near sticking

zone

9

2D Model Methodology: start with an accurate thermal field

Mould

Sticking

Step 2 (crack in a refined mesh close to meniscus)

1st crack

Schematic approachThe thermal stationary field in a refined mesh (closed to meniscus)

= initial state where the user decide

1st sticking part location

1st crack position

10

Model Methodology: add liquid elements within 1st crack

²

MouldMould

New

element

layers

Step 3

1st crack1stcrack

Temperature of liquid steel

Birth element technique to add liquid elements inside the crack

Sticking zone

11

Links of craks and mould oscillations

²

Rupture criteria for 2D FE thermal model= thinnest section of the shell

L

L< or = Vcasting× Tperiod of Oscillation

Negative Strip Time or NST:

The mold go down faster than the slab

high risk of cracks

Mould Oscillation to enhance slab displacement

Sticking Part

Descending part

NST =1

π𝑓arccos

𝑉0

2π𝑓𝐴

NST: Negative Strip Time (s)

𝐴 : Oscillation Amplitude

𝑉0 : Casting velocity

𝑓 : Oscillation Frequency

T : Oscillation Period

𝑵𝑺𝑻

𝑻=28%→35%

In practice

↑ higher oscillation Frequency ↓ NST (K. C. Mills et al. 1999)

Model constraints

Time for solidification < Negative Strip Time

12

Model Methodoly: formation of a new solid shell

²

MoldMould2nd crack

Solidification time = “negative strip time”(link with mould oscillation)

Step 5: 2nd crack eventStep 4: Liquid elements solidify

Crack: when high tension is applied due to relative mold-strand velocity (mould oscillation) and where the temperature is the hottest

13

1st result of simulation

P-2

P-1

P-3

1st sticker

1st crack

3 measurement points

Thermal history computed by 2D FE thermal model

14

Simulation results

0 10 20 30 40 50170

180

190

200

210

220

230

Times (s)

Tem

pera

ture

(°C

)

Times-Temperature

P-2P-1

Speed casting

(m/min)

Oscillation

Period (Hz)

Amplitude

(mm)

Solidification

Time -NST (s)

1.3 2 3.6 0.17

V 1→2= 46% Vcast V 2→3= 51% Vcast

Thermal peak velocity

Crack preview

Crack preview

15

1st result of simulation

0 10 20 30 40 50170

180

190

200

210

220

230

Times (s)

Tem

pera

ture

(°C

)

Times-Temperature

P-3

Speed casting

(m/min)

Oscillation

Period (Hz)

Amplitude

(mm)

Solidification

Time -NST (s)

1.3 2 3.6 0.17

Crack preview

P-1

Crack velocity Vcrack= 62% Vcast

∆T= 5°C -20°C

Crack preview

16

Influence of Casting speed

Speed

casting

(m/min)

Oscillation

Period (Hz)

Amplitude

(mm)

Solidification

Time -NST

(s)

NST/T Vcrack/VCast

0.7 2 2.0 0.173 34.6% 59.8%

1.0 2 2.9 0.174 34.8% 62,1%

1.3 2 3.6 0.170 34.2% 63.9%

10 20 30 40 50

170

180

190

200

210

220

Times (s)

Tem

pera

ture

(°C

)

Times-Temperature

10 20 30 40 50

170

180

190

200

210

220

Times (s)

Tem

pera

ture

(°C

)

Times-Temperature

10 20 30 40 50

170

180

190

200

210

220

Times (s)

Tem

pera

ture

(°C

)

Times-Temperature

Point 1

Point 2Point 3

Casting Speed = 0.7m/min

Casting Speed = 1.0m/min

Casting Speed = 1.3m/min

Convection h

(mW/mm²K )

Thermal resistance of

Contacts R (mW/mm²K )

200 2.35

Temperature Peak and Amplitude depends on Casting Speed

17

Conclusions and Perspective

Perspective: local 3D thermo-mechanical study of the crack

2D thermal FE of the mould and the slab + calibration

Tp° field validated by optical fiber measurement

Sticking Break-Out phenomenon simulation

• Crack location is based on the thermal field

• Crack event is linked with the Negative Strip Time within oscillation cycle

• Birth element technique is used to fill the crack by liquid steel

Simulation results:

• Thermal peak velocities, peak shapes and amplitudes

• Fracture propagation velocity

(crack propagation = 62% of the casting speed)

• Sensitivity analysis confirms practice:

A Slower casting velocity decreases the crack propagation velocity

recover trends

observed by EBDS