1. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 1 LES Simulation of Transient Fluid Flow and Heat Transfer in Continuous Casting Mold Bin Zhao Department of Mechanical Engineering University of Illinois at Urbana-Champaign 10/ 2002

2. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 2 Acknowledgements l Professor B.G.Thomas & Professor S.P. Vanka l Accumold l AK Steel l Columbus Stainless Steel l Hatch Associates l National Science Foundation l National Center for Supercomputing Applications

3. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 3 Previous Work LES flow and heat transfer simulations of an unconstrained impinging cylindrical jet. LES flow simulations had insufficient grid refinement near walls for heat transfer. High grid density needed near the impingement plate in order to get correct prediction of heat transfer rate without wall models.

4. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 4 Objectives Study transient fluid flow in continuous casting mold region: (eg. AK Steel thin slab mold). Study turbulent heat transfer in the casting mold liquid pool.

5. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 5 Simulation of Fluid Flow and Heat Transfer in the AK Steel Thin Slab Mold ¼ mold simulation

6. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 6 Computational details Solving 3D transient Navior-Stokes Equations Second Order accuracy in space and time Non-Structured Cartesian collocation grid Algebraic Multi-Grid (AMG) solver is used to solve pressure Poisson Equation No sub-grid model (Coarse Grid DNS) 3D flux-limited advection scheme 852,442 finite volume cells Time step 0.0005s

7. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 7 Simulation domain

8. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 8 Boundary conditions

9. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 9 Simulation parameters * Nozzle geometry based on blueprints

10. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 10 Mesh for the nozzle part AA BB CC DD EE A-A B-B C-C D-D E-E *note: mesh of half of the mold, ¼ actually used in the simulation.

11. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 11 Mesh for the mold part AA BB A-A B-B *note: mesh of half of the mold, ¼ actually used in the simulation.

12. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 12 Instantaneous velocity and temperature

13. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 13 Instantaneous velocity field at 50s Instantaneous velocity vs. dye injection experiment Dye injection experiment by R. O’Malley (Armco, Inc)

14. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 14 Instantaneous heat flux on narrow face

15. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 15 Instantaneous heat flux on wide face

16. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 16 Time-averaged fields (averaged from 30s to 50s)

17. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 17 Velocity field averaged from 30s to 50s Time-averaged velocity vs. dye-injection experiment Dye injection experiment by R. O’Malley (Armco, Inc)

18. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 18 Time-averaged heat flux on narrow face

19. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 19 Time-averaged heat flux on wide face

20. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 20 Temperature comparison (I) SEN Profile Location Measurement Position NF 50 mm

21. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 21 Temperature comparison (II) Position SEN Profile Location Measurement NF 125 mm

22. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 22 Temperature comparison (III) SEN Profile Location Measurement Position NF Middle Point

23. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 23 Temperature comparison (IV) SEN Profile Location Measurement Position NF 125 mm

24. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 24 Temperature comparison (V) SEN Profile Location Measurement Position NF 50 mm Probe broken

25. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 25 Observations NF heat flux may be a little lower than reality. Heat flux peak at NF impingement is 800 kw/m 2, instantaneously reaches 1,300 kw/m 2 ; peak at small jet impingement point on WF is 500 kw/m 2, instantaneously reaches 700 kw/m 2. No upper roll in the flow field. Temperature in the upper corner region is too low. Jed diffuse a lot in the vertical direction, resulting in weak impingement and low NF heat flux.

26. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 26 Simulation of Fluid Flow and Heat Transfer in the AK Steel Thin Slab Mold ½ mold simulation

27. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 27 Computational details Mesh: doubled (reflected) ¼ mold domain & mesh. 1,651,710 finite volume cells. Use the result of ¼ mold simulation as initial condition. Other parameters and conditions are identical to the ¼ mold simulation.

28. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 28 Simulation domain

29. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 29 Instantaneous velocity and temperature

30. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 30 Instantaneous velocity field v.s. dye injection experiment Dye injection experiment by R. O’Malley (Armco, Inc) Instantaneous velocity field at 9s

31. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 31 Temperature comparison (I) SEN Profile Location Measurement Position NF 50 mm

32. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 32 Temperature comparison (II) Position SEN Profile Location Measurement NF 125 mm

33. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 33 Temperature comparison (III) SEN Profile Location Measurement Position NF Middle Point

34. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 34 Temperature comparison (IV) SEN Profile Location Measurement Position NF 125 mm

35. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 35 Temperature comparison (V) SEN Profile Location Measurement Position NF 50 mm Probe broken

36. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 36 Instantaneous heat flux on narrow face

37. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 37 Instantaneous heat flux on wide face

38. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 38 Observations The flow field now shows tendency towards an upper roll. The jet has less diffusion in the vertical direction compared to the ¼ mold simulation: ie. has more penetration power. Applying WF-WF symmetry (cutting through a jet) is inappropriate for transient simulations.

39. University of Illinois at Urbana-Champaign Computational Fluid Dynamics Lab Bin Zhao 39 Future Work Continue the ½ mold simulation. Add shell shape into the simulation. Investigate thermal buoyancy