Spray-Cooling Control for Maintaining Metallurgical …ccc.illinois.edu/s/2015_Presentations/12_CHEN-Z_JFE_Spray...Spray-Cooling Control for Maintaining Metallurgical Length During

CCC Annual ReportUIUC, August 19, 2015

Zhelin Chen, M.S. Student

Spray-Cooling Control for Maintaining

Metallurgical Length During Speed Drop

Department of Mechanical Science & Engineering

University of Illinois at Urbana-Champaign

Steelmaking Research Dept.

JFE Steel Corporation

University of Illinois at Urbana-Champaign • Metals Processing Simulation Lab • Zhelin Chen • 2

Motivation

• Maintaining ML during speed changes within

certain bounds is helpful for many operations:

- Unbending to prevent cracks

- Staying inside support zone to prevent whales

- Soft reduction to prevent centerline segregation

• Objective: explore the potential to avoid

metallurgical length change during a casting

speed drop for a thick-slab caster using different

control methods.


‘Whale’ Formation

• Ferrostatic pressure is transmitted from the meniscus via the

liquid pool acting internally on the steel shell.

• If ML exceeded the machine length, final portion of

solidifying steel is not supported by rolls, steel will bulge out.

From CCC


Soft Reduction

• In soft reduction

operation, roll gap

should match the

shrinkage to

prevent centerline

defects

• During speed

change, ML should

stay within soft

reduction region.

Rolls deform strand

to match shrinkage

Centerline is susceptible

to segregation and other

defects if the roll gap

does not satisfy the

desired shrinkage

Solid

Liquid


Objective

• Choose realistic caster and conditions to explore control methods to maintain ML during a speed drop

• Caster studied: thick slab caster at JFE Steel

• Steel grade: not sensitive to surface cracks but may be sensitive to centerline defects.

• Small speed drop: 1.7m/min to 1.5m/min.

- maintaining constant ML during steady state is

achievable by changing water flow rates.


Cononline: On-line Control System for

Secondary Cooling Water Sprays

Concontroller: tries to keep

Consensor

predictions

at setpoints

Consensor:predicts

surface

temperature

and shell

thickness

[1] Petrus, B., K. Zheng, X. Zhou, B.G. Thomas, and J. Bentsman, “Real-Time Model-Based Spray-Cooling Control System for Steel Continuous Casting”, Metallurgical and Materials Transactions B, Vol. 42B:2, 87- 103, 2011


Conoffline

• Currently runs on one Linux server.

Recorded or specified casting

conditions

∑

SetpointGenerator

Surface temperature setpoint

Shell thickness setpoint

User specified control loop

Caster

data

Spray water

flow rate

Shell thickness and surface temperature estimation

Controller:• Concontroller: Automatic PI

control loop

• Spray table control

• “Time-constant” based control

• User specified control

methods

Consensor


Example of Conoffline “replay” File

• Generate scenarios by editing CSV files in Excel:

-- each row contains all (83) CON1D casting

conditions at specified time.

Time Casting conditions

Casting speed

sample_timedistance from meniscus to mold top

tundish nozzle submergence depth

speed width thicknessPouring

temp.Slab Chemistry

mm mm m/min mm mm deg_C

21:30:00 90 250 1.7 2095 221 1557

21:40:09 90 250 1.7 2095 221 1557

21:40:52 90 250 1.5 2095 221 1557


Validation of Cononline with Roll Force Measurements [4]

• Predicted thermal linear expansion of strand thickness (depends on liquid shrinkage & metallurgical length) matches closely with timing of changes in measured roll loads during speed changes at Burns Harbor caster [5].


Thermal Linear Expansion of Carbon Steels

Liquid data from: Jimbo & Cramb, Met. Trans. B, 24B, 1993, 5-10.Solid data from: Harste, Jablonka & Schwerdtfeger, 4th Int.Conf.

on Continuous Casting, CRM, 1988, Brussels, 633-644

1.5

2

2.5

3

3.5

4

1300 1350 1400 1450 1500 1550

TL

E (

%)

Temperature (C)

0.003%C

0.1%C

0.2%C

0.4%C

0.6%C

0.8%C

γγγγ or δδδδ

L


Casting Condition

Liquidus Temperature 1516.10 oC

Solidus Temperature 1468.37 oC

Density of solid steel 7400 kg/m3

Steel emissivity 0.8 -

Fraction solid for shell

thickness location0.3 -

Specific heat of solid steel 670 J/kgK

Thermal conductivity of solid

steel30 W/mK

Thermal diffusivity of solid steel 6.0508 e-6 m2/s

Latent heat of fusion 271 kJ/kg

Initial cooling water temperature 29.67 oC

Pour temperature 1545 oC

Mold conductivity (WF/NF) 418.7/355 W/mK

Time step 0.01 s

Mesh size 0.55 mm

Slab thickness 221 mm

Slab width 2095 mmSlab Size


Surface Heat Removal Eqs

• Mold heat flux: varies with casting speed, based

on empirical correlation for a thin slab caster [2]

• Heat flux due to water sprays: empirical relation

of Nozaki [3]

[ ]( )0.47

2

mMW/m 1.2154 m/min

cQ v =

( )

( )( )sw sw surf sw

0.55

sw sw surf sw0.3925 1 0.0075

q h T T

Q T T T

= −

= ⋅ ⋅ − ⋅ −

Average mold heat flux Casting speed

Spray water flux in L/m2/min

Steel surface

temperature Spray water

temperature


Shell Thickness at Steady State

0 1 2 3 4 5

x 104

0

20

40

60

80

100

120

Distance from meniscus(mm)

Shell

thic

kne

ss(m

m)

Casting speed: 1.7m/min

Casting speed: 1.5m/min

ML=22.29 m

ML=19.98 m


Calibration of Conoffline

• Calculated surface temperatures agree with measurements under steady state.

0 1 2 3 4 5

x 104

0

200

400

600

800

1000

1200

1400

1600


Sla

b s

urf

ace

te

mp

era

ture

(de

gC

)

Vc=1.7m/min

Simulation result

Measured result

0 1 2 3 4 5

x 104

0

200

400

600

800

1000

1200

1400

1600


Sla

b s

urf

ace t

em

pera

ture

(de

gC

)

Vc=1.5m/min

Simulation result

Measured value


• Control setpoints: Controller setpoints is generated based on the Cononline result of steady state at 1.7 m/min.

Zone # Setpoint (deg)

1 1214.62

2 1122.15

3 1071.99

4 932.26

5 840.31

6 795.21

7 771.47

8 780.88

9 822.50

10 816.52

11 792.28

12 726.04

Zone # Kp (/60)

1 0.5

2 2

3 2

4 2.5

5 2.5

6 3.5

7 3.5

8 3.5

9 2.5

10 2.5

11 4.5

12 4.5

Zone # Ki (/60)

1 0.1

2 0.1

3 0.25

4 0.25

5 0.25

6 0.25

7 0.25

8 0.25

9 0.25

10 0.25

11 0.35

12 0.35

First 2 zone is very short and right after the mold exit and their behavior will

largely affect the surface temperature for the following zones, so choose smaller

kp and smaller damping (ki)

Example Concontroller Setup for

Cononline/Conoffline PI Control


-500 0 500 1000 1500900

950

1000

1050

1100

1150

1200

1250

Time after speed change(Sec)

Avera

ge

Su

rfa

ce T

em

pera

ture

(de

g)

Average Surface Temperature Change During Sudden Slow Down

zone 1

zone 1 setpointzone 2

zone 2 setpoint

zone 3


zone 4 setpoint

Simulation Results for PI Control:

Surface Temperature Histories (12 zones)

-500 0 500 1000 1500750

760

770

780

790

800

810

820

830

840

850


Avera

ge S

urf

ace T

em

pera

ture

(deg)


zone 5


zone 6 setpoint

zone 7


zone 8 setpoint

-500 0 500 1000 1500700

720

740

760

780

800

820

840


Avera

ge S

urf

ace T

em

pera

ture

(deg)


zone 9


zone 10 setpoint

zone 11


zone 12 setpoint

-1000 -500 0 500 1000 15000

20

40

60

80

100

120

140


Wate

r flow

rate

(l/m

in/r

ow

)

Spray Flow Rate Change During Sudden Slow Down

zone 1zone 2

zone 3

zone 4zone 5

zone 6

zone 7

zone 8zone 9

zone 10

zone 11zone 12


ML Behavior during Speed Change under

PI Control

Cononline can maintain constant surface temperature during speed change,

but its performance on maintaining ML is not good.

2.31 m

-500 -100 300 700 1100 1500 1900

1.5

1.6

1.7

Time(Sec)

Casting s

peed(m

/min

)

Metallurgical Length Change During Sudden Slow Down

-500 -100 300 700 1100 1500 190019

20

21

22

23

Meta

llurg

ical Length

(m)

2.66 m


Reduce water flow rates at lower speed: match ML at steady-state

0 1 2 3 4 5

x 104

600

700

800

900

1000

1100


Sla

b s

urf

ace tem

pera

ture

(de

gC

)

1.7-Conv.1.5-Conv

1.5-Reduced

No. 1.7-Conv. 1.5-Conv. 1.5-Reduced.

vc 1.7 m/min 1.5 m/min 1.5 m/min

Spray pattern Conventional Conventional

1-2Z: Conv.

3Z: Conv. x 0.7

4Z: Conv. x 0.45

5-8Z: Conv. x 0.35

8-12Z: Conv. x 0.7

ML (fs=0.7) 22.29 m 19.98 m 22.17 m

0 1 2 3 4 5

x 104

0

20

40

60

80

100

120


Sh

ell

thic

kn

ess (

mm

)

1.7-Conv.

1.5-Conv1.5-Reduced


Control Methods for ML in Transient Conditions (i.e. speed drop)

• Spray table control

- conventional (conv.) to conv.

- conv. to reduced

• Time-constant spray control

• PI control for ML (60X surf temp gains)

(1.7 to 1.5 m/min)


Spray Table Control (conv. to conv.)

-ML

• Spray pattern:

conventional (1.7m/min) to conventional (1.5 m/min)

• ML drop = 2.31 m

= ML change (defined as maximum change in ML)

2.31 m


Spray Sable Control (conv. to conv.)

-Surface Temperature

• Spray pattern = conventional

to conventional

zone1

zone2

zone3

zone4

zone6

zone5

zone8zone7

zone10

zone9

zone11

zone12

Time after speed change (sec)

Ave

. S

urf

. Te

mp

.


Ave

. S

urf

. Te

mp

.


Ave

. S

urf

. Te

mp

.


Spray Table Control (conv. to reduced) -ML

• Spray pattern = conventional to reduced

• ML change = 0.92 m

0.92 m


Spray Table Control (conv. to reduced)-Surface Temperature

• Spray pattern = conventional

to reduced

zone1zone2

zone3

zone4

zone6

zone5

zone7

zone8

zone10zone9

zone11

zone12


Ave

. S

urf

. Te

mp

.


Ave

. S

urf

. Te

mp

.


Ave

. S

urf

. Te

mp

.


“Time-constant” Spray Control(conv. to reduced spray pattern)

• Spray table control: spray water rate changes immediately when speed changes.

• Intuitive idea: spray water rate changes gradually as speed changes, (i.e. spray water rate changes according to time instead of velocity).

• For a typical caster, the relation between spray water rate and casting speed is assumed to be:

• Above equation can be transferred to:

• Can be found solving the inverse of equation:

[ ] [ ]/ / / minspray cSW l m row A Bv m= +

[ ]( )

[ ]/ / / minspray

zSW l m row A B m

zτ= +

( )( )t

t Z v s ds zτ−∫ =


“Time-constant” Spray Control - ML

(conv. to reduced spray pattern)

• ML change = 1.61 m

1.61 m


“Time-constant” Spray Control – Surface

temperature (conv. to reduced spray pattern)

• The transition between two

steady state is more

smooth than conv. to

reduced spray table control

zone1zone2

zone3

zone4

zone6

zone5

zone7

zone8 zone10

zone9

zone11

zone12


Ave

. S

urf

. Te

mp

.


Ave

. S

urf

. Te

mp

.


Ave

. S

urf

. Te

mp

.


• Control setpoints: Controller setpoints is generated based on the CONONLINE result of steady state at 1.7 m/min.

Zone # Setpoint (mm)

3 23.65

4 33.45

5 44.32

6 58.80

7 76.20

8 91.25

9 105.01

10 110.5

11 110.5

12 110.5

Zone #

Water flow rate upper bound (l/min/row)

3 180

4 70

5 70

6 63

7 63

8 42

9 26

10 48

11 36

12 36

Zone #

Water flow rate lower bound (l/min/row)

3 36

4 10

5 10

6 4.5

7 4.5

8 3

9 1.53

10 1.5

11 1.5

12 1.5

PI Control based on Shell Thickness

Setpoints


PI control based on Shell Thickness Setpoints-ML

• ML change = 0.87 m, BEST

-500 -100 300 700 1100 1500 1900

1.5

1.6

1.7

Time(Sec)

Casting s

peed(m

/min

)

Metallurgical Length Change During Sudden Slow Down

-500 -100 300 700 1100 1500 190019

20

21

22

23

24

25

X: -38

Y: 22.29

Meta

llurg

ical Length

(m)

0.87 m


PI Control based on Shell Thickness Setpoints - Ave. Surface Temp.

-500 0 500 1000 1500750

800

850

900

950

1000

1050

1100

1150


Ave

rag

e S

urf

ace

Te

mp

era

ture

(deg

)


zone 5

zone 6zone 7

zone 8

-500 0 500 1000 1500700

750

800

850

900

950

1000

1050

1100


Ave

rag

e S

urf

ace

Te

mp

era

ture

(deg

)


zone 9

zone 10zone 11

zone 12

-500 0 500 1000 1500900

950

1000

1050

1100

1150

1200

1250


Ave

rag

e S

urf

ace

Te

mp

era

ture

(deg

)


zone 1

zone 2zone 3

zone 4

zone1zone2

zone3zone4

zone6

zone5

zone7zone8

zone10

zone9

zone11

zone12

• Change of water flow rate have

immediate response on average

surface temperature.


Ave

. S

urf

. Te

mp

.


Ave

. S

urf

. Te

mp

.


Ave

. S

urf

. Te

mp

.


Conclusions: Conoffline

• Has been calibrated to match steady casting conditions at a

commercial thick-slab caster

• Using PI controller, surface temperature can successfully be

controlled during a speed drop at this caster.

• Has been applied to investigate potential for ML control

during speed drop from 1.7 to 1.5 m/min:

– Conventional spray table control causes metallurgical length drop of 2.31 m (fs=0.7).

– Spray table control with conv. to reduced water flow rates at low speed causes metallurgical length (ML) drop of 0.92 m (fs=0.7).

– “Time constant” spray control (based on conv. to reduced spray table control) causes ML change of 1.61 m (fs=0.7)

– PI control based on shell thickness causes ML change of 0.87 m.

30


Future work

• Improve control methods to maintain ML during speed change.

• Tune the PI controller gains for shell thickness feed back control method.

• Find better (fundamental) methods for ML control

• Investigate other scenarios?

• Any other suggestions?


Acknowledgments

• Continuous Casting Consortium Members (ABB, AK Steel,

ArcelorMittal, Baosteel, JFE Steel Corp., Magnesita

Refractories, Nippon Steel and Sumitomo Metal Corp.,

Nucor Steel, Postech/ Posco, SSAB, ANSYS/ Fluent)

• JFE Steel for providing caster data and casting conditions

• Akitoshi Matsui (JFE Steel) for his excellent work and

continuous support on this project.

• Dr. Petrus, (former grad student, currently at Nucor) for

helping with Cononline program development and advice.

• Prof. Thomas and Prof. Bentsman for their patience,

knowledge and guidance on this project.


References

• [1] Petrus, B., K. Zheng, X. Zhou, B.G. Thomas, and J. Bentsman, “Real-Time Model-Based Spray-Cooling Control System for Steel Continuous Casting”, Metallurgical and Materials Transactions B, Vol. 42B:2, 87- 103, 2011

• [2] P. Duvvuri, B. Petrus, B.G. Thomas, “Correlation for Mold Heat Flux Measured In A Thin Slab Casting Mold,” AISTech 2014, Indianapolis, IN, 2014

• [3] T. Nozaki: Iron Steel Inst. Jpn. Trans., vol. 18, pp. 330-38, 1978.

• [4] Petrus, Bryan, Danny Hammon, Megan Miller, Bob Williams, Adam Zewe, Zhelin Chen, Joseph Bentsman, Brian G. Thomas, “New Method to Measure Metallurgical Length and Application to Improve Computational Models”, AISTech 2015, Cleveland, OH, May 4-6, 2015, Assoc. Iron Steel Technology, Warrendale, PA, pp. 3238-3248.

• [5] Gregurich, N., Flick, G., Moravec, R., and Blazek, B. “In-Depth Analysis of Continuous Caster Machine Behavior During Casting With Different Roll Gap Taper Profiles”, Iron & Steel Technology, Dec. 2012

Documents

Spray-Cooling Control for Maintaining Metallurgical …ccc.illinois.edu/s/2015_Presentations/12_CHEN-Z_JFE_Spray...Spray-Cooling Control for Maintaining Metallurgical Length During