Ecole de Mécanique des Matériaux Mécamat, Aussois 2008 Modelling techniques for (semi-) continuous casting processes Outils numériques appliqués aux procédés de coulée continue J.-M. Drezet and M. Rappaz Computational Materials Laboratory Ecole Polytechnique Fédérale de Lausanne MXG CH 1015 Lausanne, Switzerland
Table of Contents 1. Overview of the Al, Cu and Steel casting processes 2. Status of solidification simulation capabilities 3. State-of-the-art and challenges at the macro scale Heat flow fluid flow Stresses, mushy behavior and residual stresses Macrosegregation 4. State-of-the-art and challenges at the micro scale Grain structures Microstructures Porosity Hot tearing 5. Conclusion
Table of Contents 1996-2000 EU project EMPACT (European Modeling Program on Aluminium Casting Technology) 2000-2004 EU project VIRACT (Virtual Cast House) 2003-2006 project POST (Porosity and Stress) 2005-2006 Project Alcan Fund (Alufond)
1. General Overview of Continuous Casting liquid Sump depth as a function of ingot format and alloy properties: x sump depth l V = 1 mm/sec 2w = 0.5 m solid 2w Copper k (W/mK) 350 density (-) 9000 Sump depth (m) 0.62 speed v Aluminium Steel 200 31 2700 7850 0.93 10.13
1. Overview of the Al. semi-continuous casting process 1 Convection Macrosegregation 2 Exudation 5 Grain movement S+L 6 Microstructure formation 4 Lateral faces pull in 8 Hot cracking Solid 7 Microporosity x 3 Butt curl Butt swell
1. Overview of Copper Continuous Casting Copper: horizontal or vertical casting Coarse columnar microstructure Almost all macro and micro aspects similar to Al. but with different scales owing to different physical properties
1. General Overview of Steel Continuous Casting Computational Fluid Dynamics (CFD) Casting simulation Finite Element Structural Analyis
2. Solidification Calculations : Evolution over 25 years 10 9 10 8 10 7 # Nodes 10 6 10 5 10 4 1000 20.6667 (Year-1980) 10 3 10 2 0 5 10 15 20 25 Year-1980 3 largest solidification simulations (number of nodes) reported in each volume of the MCWASP conference follow Moore s law (double every 16 months)! V.R. Voller and F. Porte-Agel, J. Computat. Physics 179, 698-703 (2002).
2. Solidification Calculations Whatever the power of the computers, scientists and engineers are always going to use them at the limit. From simple thermal 2D computations done in 1980, we see nowadays complex multi-scale, multi-component alloys, multi-phenomena simulations. Although simulations are still limited by the computers performances, who could imagine of not using them? Simulation is saving money not only through direct cost but through a LEARNING process where concomitant phenomena can be separated. Validation and determination of model input parameters are really an important step to get confidence in the results and to prospect new directions When defining a strategy of simulation, Moore s law should be accounted for (eg. molecular dynamic computations).
3. State-of-the-art and challenges at the macro scale Heat flow simulations do not pose any challenge, except maybe boundary conditions. Fluid flow and thermo-mechanical simulations are more delicate but can rely on advanced commercial software. There is no direct coupling between fluid flow in the liquid and stresses in the solid. But fluid flow changes the shape of the isotherms, and thus modifies thermal stress development. Combined heat-fluid-stresses calculations require mixed Lagrangian- Eulerian representations. Achieved only in few codes. Macrosegregation calculations are still very difficult : due to the various origins (convection, shrinkage, stresses, grain sedimentation) often very localised (e.g., in regions where v l c 0).
Heat Flow 6 Heat flux [MW/m 2 ] 5 4 3 2 1 Water impingement point 0 0 50 100 150 200 250 300 Vertical distance [mm] Heat flux as a function of the distance to the top liquid level for an AA5182 alloy, as determined by inverse modeling. J.-M. Drezet et al, Met. Mater. Trans. 31A (2000) 1627
Mould Thermo-Mechanical Aspects Water flow T S T L S+L Solid Liquid Bottom block Lateral faces pull-in G Liquid S+L Solid G Butt curl V g G Butt swell A non uniform contraction of the whole ingot leads to distortions. J.-M. Drezet et al, Met. Mater. Trans. 31A (2000) 1627
Thermo-Mechanical Aspects Shrinkage between mould and ingot. W. Schneider et al, in Continuous Casting Conf., Neu-Ulm, Ed. Mueller, 2005
Thermo-Mechanical Aspects Ingot rolling faces pull-in J.-M. Drezet, PhD thesis no. 1509, EPF-Lausanne, 1996
Thermo-Mechanical Aspects mould rectangular convex ingot Steady state (500 to 9000 mm) swell Butt (0 to 500 mm) Optimized mould shape W. Schneider et al, in Continuous Casting Conf., Neu-Ulm, Ed. Mueller, 2005
Thermo-Mechanical Aspects 8 Shrinkage [%] 6 4 810 mm 2 1500 mm Measured 0 1200 1000 800 600 400 200 0 Lateral width [mm] Measured and simulated contraction along the rolling faces of a 2200 600 mm 2 AA1050 slab cast at 60 mm/min. W. Droste et al, in Light Metals (TMS Publ. 2002) p. 703.
Thermo-Mechanical Aspects Thermal and deformation field within the mould (external view of ingot butt) Deformation during the start-up phase including fluid flow, as calculated with Alsim/Alspen for an AA5182 alloy using a mixed Lagrangian-Eulerian approach. D. Mortensen, Met. Mater. Trans. 30B (1999) p.119.
Thermo-Mechanical Aspects Measured and computed butt curl and curling speed (Abaqus). W. Droste et al., in Continuous Casting Conf., Frankfurt, Ed. Ehrke, p. 175, 2000.
Thermo-Mechanical Aspects: solid behaviour Modified Ludwik s law for as-cast behaviour: Gleeble tests 0.7 0.6 0.5 0.4 AA2024 n(t) m(t) K(T) 700 600 500 400 n(t) p p p 0 ε& p ( 0 ) σ=k(t) ε +ε ε& m(t) 0.3 0.2 0.1 0-0.1 0 50 100 150 200 250 300 350 400 450-100 300 200 100 0 K(T) alloy consistancy (MPa) n(t) strain hardening coefficient m(t) strain rate sensitivity (-)
Fluid Flow 0.02 m/s T L T S 660 600 500 Feathery grains 300 Velocities and temperatures at the surface and in the vertical symmetryplane of an AA1050 DC cast billet as calculated with Fidap. S. Henry et al, Met. Mater. Trans. 35A (2004) pp. 2495-2501.
Thermo-Mechanical Aspects: cold and hot cracking Solidification cracking (hot tearing) during casting Cold cracking during ingot storage Productivity and safety issues
Thermo-Mechanical Aspects: behaviour of the mush Von Mises stress vs strain curves for an Al-2 wt pct Cu alloy, tested in simple compression in the fully solid state and in shear condition in the mushy state O. Ludwig et al., Met. Mater. Trans. 36A (2005) 1525.
Thermo-Mechanical Aspects: behaviour of the mush Compressible constitutive model with internal variables (cohesion). Drained compression tests O. Ludwig et al., Met. Mater. Trans. 36A (2005) 1525.
Thermo-Mechanical Aspects: residual stresses Ingot sawing operation are delicate for high strength alloys. Saw pinching and ingot cracking are predicted using strain energy release and FE modeling J.-M. Drezet et al., in Solidification Processing 2007, Ed. H. Jones, p. 690., Sheffield, 2007.
Thermo-Mechanical Aspects: residual stresses Cracking might occur if lip tends to open and rate of energy release is higher than a critical value which depends on the fracture toughness of the alloy. J.-M. Drezet et al., in Int. J. of Cast Metals Research 2007, vol. 20, p. 163.
Thermo-Mechanical Aspects: residual stresses Cold cracking after casting during storage. J-type crack (courtesy Alcan CRV) Interaction of residual stresses with a given population of stress risers i.e. defects such as oxydes, particles, inclusions, microcraks, pores, (fracture mechanics)
Mixed Lagrangian-Eulerian description feeding velocity fixed level 4 Fluid flow and stress development 3 casting velocity v c 4 1 2 2 3 Fixed mesh: Eulerian description Moving mesh owing to deformation and/or transport of solid: Lagrangian description
Coupled fluid flow and solid deformation Automatic generation of the accordion domain. The thickness of the elements is set here to a small value to make the elements visible. One layer of elements of the accordion domain is already stretched (ProCAST).
Coupled fluid flow and solid deformation Fluid flow and stress are computed simultaneously (ProCAST distributed by ESI, www.calcom.ch).
Macrosegregation Empact measurement 0.01 0.1 0.99 <ω > %pds Mg CALCOSOFT Mg concentration AA5182 alloy with 4.05 wt% Mg T. Jalanti, PhD Thesis 2145, EPFL (2000).
4. State-of-the-art and challenges at the micro-scale Modelling of grain structure formation (i.e., nucleation/ fragmentation, growth and transport) in Al and Cu casting is still a challenge. Modelling of microstructure formation can be coupled with phase diagram calculations. It is restricted to a small region (typically mm 2 ) in 2D (Phase field, Pseudo front tracking, ). Remaining challenges : CPU time, nucleation of phases, kinetics of eutectics. Microporosity modelling, including pipe shrinkage and macroporosity, can now be predicted in 3D. Model being extended to multicomponent alloys and more than one gas. Challenge of the curvature contribution, but hard to measure so hard to validate. Hot tearing has made substantial progress over the past five years, thanks to the VIRCAST project. Two-phase approaches (i.e., deformation of the solid mush and liquid feeding) now exist, as well as an approach to coalescence for the prediction of bridging of the primary phase.
4. State-of-the-art and challenges at the micro-scale In situ observation of columnar and equiaxed grain structure formation (X-Ray imaging at ERSF in Grenoble) Time Resolved X-Ray Imaging of Dendritic Growth in Binary Alloys, R. H. Mathiesen and L. Arnberg in Physical Review Letters, vol. 83, no. 24, 1999.
As-cast Grain Structures Velocity Calculation of nucleation, growth and transport of equiaxed grains using a granular approach. Enthalpy (T,fs) H. Combeau et al, INPL-Nancy
As-cast Grain Structures AA5182 1000 mm 188 µm 250 mm liquid metal inlet 385 µm Grain size distribution in a quarter section at the outlet of a DC casting N. Houti et al, Aluminium (2004)
Cu as-cast grain structures: electromagnetic stirring Influence of the position of the induction coil on the convection in the liquid as calculated for the BZ4 alloy (Cu-4wt%Zn- 4wt%Sn-4wt%Pb) Th. Campanella et al. in Met. Trans. A. vol. 35 (2004), p. 3201.
Cu as-cast grain structures: electromagnetic stirring Schematics of remelting (1) and fragmentation (2) of secondary dendrite arms to give birth to new nuclei (promotion of the CET). Th. Campanella et al. in Met. Trans. A. vol. 35 (2004), p. 3201.
Cu as-cast grain structures: electromagnetic stirring C R d u l, z = > 1 V T Schematics of local fluid flow in the liquid, u l, and in between the dendrites, u ld. Criterion for remelting.
Cu as-cast grain structures: electromagnetic stirring Optical micrographs of longitudinal cross-sections of the solidified C97 and BZ4 alloys without EMS (P = 0% at Vc = 4 mm/min) and with EMS (P = 10 and 30% at Vc = 4 mm/min) for two inductorliquidus distance, 23 and 13 mm. Th. Campanella et al. in Met. Trans. A. vol. 35 (2004), p. 3201.
As-cast Microstructures: Phase field methods AA5182 Calculated Mg concentration field at 27 mm from the rolling face. Microstructure at 27 mm from the rolling face. A. Jacot et al, in Modelling of Casting, Welding and Advanced Solidification Processes, Eds D. Stefanescu et al (TMS Publ. 2003), p. 229.
Microporosity Al-4.5wt%Cu 1 mm/s 0.34% 0.15% Ch. Pequet et al, Met. Mater. Trans. 33A (2002) 2095.
Hot Tearing gravity z Liquid feeding Integrated critical strain (ICS) feeding Solidification Coalescence p l p c Solid deformation Two-phase approach to hot tearing, using the rheology of a compressible porous solid and a Darcy-type equation for liquid feeding. M. M Hamdi et al, in Light Metals (TMS Publ., Warrendale, USA, 2004)
Hot tearing Casting speed [mm/min] 130 120 110 100 90 T2 T4 T3 T5 T1 T2 Pressure drop [kpa] 160 140 120 100 80 60 40 T1 T3 T2 T4 T5 AA 6060 φ= 228 mm 80 20 0 50 100 150 200 250 300 Cast length [mm] 0 10 20 30 40 50 60 70 Cast length [mm] J.-M. Drezet et al, in Aluminium Alloys (Trans Tech Pub., CH, 2002) p. 59.
Hot tearing : coalescence and percolation 2 mm s 600 ºC Al - 1%Cu 550 ºC Red : wet grain boundaries 500 ºC Evolution of a population of equiaxed grains from a continuous liquid film network to a fully coherent solid V. Mathier et al, MSMSE 12 (2004) 479. Extended by S. Vernède
Hot tearing : morphology of the liquid phase In situ 3D investigation of solidification (Al -4%Cu) Evolution of liquid morphology 635 C (59%) 604 C (85%) ESRF-Grenoble In O. Ludwig et al., Met. and Mat. Trans. 36A, June 2005, p. 1515. 582 C (92%) 560 C (95%)
Conclusion Macroscopic models of solidification are fairly mature for Al. and steel casting (slightly less for Cu casting) Macrosegregation calculations are still a challenge for two reasons : difficulty to maintain the mass balance, various origins (shrinkage, convection, sedimentation, deformation). Grain structure modelling still a topic of active research. Microstructure modelling coupled with phase diagrams can be done for a small 2D domain. Defect predictions have really moved forward over the past 4-5 years. Porosity : the challenge is to validate the model predictions by accurate comparison. Hot tearing : the challenge is to combine coalescence, mechanical deformation and liquid feeding.