Thermal Behavior of Hot Copper Plates for Slab Continuous Casting Mold with High Casting Speed

Transcription

1 , pp Thermal Behavior of Hot Copper Plates for Slab Continuous Casting Mold ith High Casting Speed Xiangning MENG and Miaoyong ZHU School of Materials & Metallurgy, Northeastern University, Shenyang , P. R. China. (Received on February 12, 2009; accepted on April 24, 2009) A three-dimensional finite-element heat-transfer model as established to predict temperature of hot copper plates in a slab continuous casting mold ith high casting speed and the temperature distribution and effect of casting speed on thermal behavior ere simulated in detail. The results sho that the calculated temperature agree ell ith the measured ones and the temperature of hot copper surface is influenced by the flash elded chrome (Cr) layer to a certain extent. The temperature profile of hot copper surface is determined by heat flux, material properties, solidifying shell shrinkage and mold taper and presents a certain shape. Temperature distributions in different transverse sections along effective casting height of mold are all similar and depend on mold structure and contact state beteen mold all and slab. The centre temperature of hot copper surface at casting speed 1.8 m min 1 and 2.0 m min 1 are higher than that of 1.6 m min 1 casting speed C and C respectively and temperature is not increased linearly ith casting speed. Temperature difference adjacent to meniscus beteen mold all and shell surface is influenced obviously by casting speed and increased C ith increment of casting speed 0.2 m min 1. Fluctuation of temperature difference in meniscus should be a main reason to deteriorate casting effectiveness as increasing casting speed. KEY WORDS: slab continuous casting mold; high casting speed; copper plate temperature; finite element model. 1. Introduction In conventional continuous casting of steel, a large amount of sensible and latent heat of molten steel dissipates in primary cooling zone and continuous casting mold becomes the most critical component of continuous caster to govern steel quality. In recent years, the thermal load on mold has been increasing constantly and the orking environment also been more abominable since high productivity and quality have been required and relevant high-effective technologies been popularized and applied. Therefore an in-depth understanding thermal behavior for continuous casting mold ith high casting speed is essential to design mold and maintain casting stability and it is also significant to other relevant research. Many studies have been carried out to shed light on thermal behavior of continuous casting mold over the past years. Cho et al. 1) applied a mathematical billet thermal model to the heat transfer results from a high casting speed instrumented mold trial to provide a ne mold taper for minimizing mold strand interaction. Thomas and Liu et al. 2,3) investigated the relation beteen copper plate geometry of mold and temperature distribution on the basis of a finite-element heat-transfer model and evaluated the mold design. Lu et al. 4) indicated that the main factors on mold temperature are thickness of copper plate and plated nickel layer through a heat transfer model for slab casting. Samarasekera and Brimacombe 5 7) calculated temperature of copper plates for both billet and slab molds based on heat flo in mold all and the effect of a large number of variables ranging from ater quality to casting speed, moreover applied mathematical models to predict temperature field as a function of operating design variables combined ith inplant measurements. Park, Thomas and Meng et al. 8 10) developed a finite-element thermal-stress model to determine temperature and thermal distortion and stress in both thin and conventional slab molds and O Connor and Dantzig 11) applied a elastic-plastic-creep finite-element model in a funnel shaped mold for casting thin slabs in order to achieve the same goals. Cicutti et al. 12) evaluated mold heat transfer and temperature fluctuation in a slab mold and proposed an expression to predict global heat flux as a function of process parameters and a variability index to quantify mold thermal instability. All orks mentioned above have shon that thermal behavior is not only important to mold life and casting stability, it is also the premise to further investigate the mechanical behavior in mold. At present, the detailed analysis to the heat transfer in mold in combination ith actual production operation has seldom involved, especially for high casting speed. The present article aims to quantity temperature of hot copper plates of a slab mold ith a high casting speed using a three-dimensional heattransfer model and combines operating conditions and primary cooling mechanism to provide an insight into effect of 2009 ISIJ 1356

2 increasing casting speed. 2. Mathematical Models 2.1. Finite Element Model The physical model folloing design diagram for calculating temperature of mold copper plates during steady operation has been established using a commercial finite-element analysis package ANSYS TM and provided as a representative three-dimensional quarter-mold model because of symmetry shon in Fig. 1. Five cooling ater slots are arranged evenly beteen each bolt in ide face of mold and thirteen slots are arranged in narro face including deep ones adjacent to bolts and other shallo ones and make up of primary cooling system together ith cooling ater apertures in steel backups. Copper plates surface is coated ith a nickel layer in loer portion (above mold exit 500 mm), here mold ear is usually greater due to solid friction cause by ferrostatic pressure of molten steel. The hole mold hot face is flash elded a very thin chrome layer (0.05 mm) in order to avoid copper element being permeating into casting slab to produce stellated crack on slab surface. In addition, thermocouples are embedded under hot face 23 mm and to ros of ide face far from mold top 200 mm and 400 mm and that of narro face are 200 mm and 440 mm respectively. The meshed finite-element model is described using three-dimensional 10-node tetrahedral thermal solid element hich referred to in the ANSYS TM manual as SOLID87 and shon in Fig. 2. The local mesh scale of chrome and nickel layers and cooling ater slots has been conducted for guaranteeing calculation precision. The minimum mesh size is 10 mm and free mesh pattern is used Heat Flo Model In order to simulate thermal behavior of hot copper plates of slab mold, the folloing assumptions have been made: (1) heat transfer in mold is steady and symmetrical; (2) thermal properties of copper plate and steel backup are isotropic and density and heat capacity are constant; (3) ater in cooling channel is in plug flo and nuclear boiling of cooling ater is neglected due to semicircular bottom of slot and (4) top and bottom of mold are considered to be adiabatic and heat absorption by mold poder above meniscus is negligible. As thermal conductivity is dependent on temperature, temperature is calculated by solving steady heat conduction equation using non-linear finite elements. T T T λ( T ) λ( T ) x x y y z λ( T ) z 0...(1) Where, l is thermal conductivity of copper and steel. The heat flux on mold hot faces as a function of distance both across and don the mold and function coefficients may be determined by heat flux equilibrium of cooling ater. 3,9) Thus the relevant boundary conditions can be expressed as Eqs. (2) and (3) and the heat flux beneath meniscus from 30 mm along off-corner to corner region of ide face and narro face decreased to 67% of standard ide Fig. 1. Schematic diagram of slab mold geometry: (a) top vie of quarter-mold, (b) transverse section of cooling ater slots, (c) longitudinal section of hot copper plates. Fig. 2. Meshed model for finite-element analysis. face and narro face values to simulate large interfacial gaps near mold corner. And the top and bottom of mold and cold face of steel backup are considered to be adiabatic. Thus the relevant boundary conditions can be expressed as Eqs. (4) and (5). In addition, the convection heat transfer beteen cooling ater and copper plates are described as Eq. (6) and its heat transfer coefficient expressed as Eq. (7). q a a 1 a 2 z...(2) q a a z b (3) ISIJ

3 To top and bottom of mold: T q λ( T) 0...(4) z To cold face of steel backup: T T q λ( T) λ( T ) 0...(5) x y q h (T T )...(6) Table 1. Table 2. Function coefficients for different casting speed. Properties of copper, nickel, steel and cooling ater. h λ d dρv Cμ μ λ...(7) Where, q is heat flux (subscript marks a, b and represent above and belo meniscus and cooling ater respectively), a i (i 1, 2, 3) is function coefficient, h is heat transfer coefficient, l is thermal conductivity, d is hydraulic diameter of slot, r is density, v is flo rate, m is viscosity and C is specific heat. Table 3. Operation conditions of continuous casting mold. 3. Results and Discussion 3.1. Models Availability The function coefficients of boundary conditions of heat flux for different casting speed have been determined and listed in Table 1. The thermal behavior of copper plates of slab mold as casting SPHC ([C] 0.08%) steel at high casting speed 1.6 m min 1 has been calculated and used thermal properties and operating conditions given in Table 2 and Table 3. 13) The predicted temperature on positions of thermocouples have been compared ith in-plant measurement to verify exactness of above-mentioned models and shon in Fig. 3. All available temperature measured fluctuates up and don near the calculated temperature and the compatible ones that accord ith calculating value appear frequently, it proves the mathematical models are available. Through analyzing on-line process data, it orth to be pointed that the real temperature of thermocouples has a greater fluctuation during actual production even under steady casting state and the greatest temperature difference beteen adjacent monitoring moments is nearly 20 C. It is because the stochastic change of temperature at measuring positions caused by periodic contact beteen mold all and slab surface originating from mold oscillation, shell shrinkage and ferrostatic pressure. A regular pattern, lo nearby cooling ater slots and high nearby binding bolts, presents in temperature distribution of ide face midst. Lo temperature of off-corner region due to lo heat flux and the rapid temperature reducing near ide face corner still oing to the sufficient cooling and hindrance to heat transfer caused by ider corner gap. Uniformity and compactness of cooling ater slots make the temperature of narro face midst change gently and the donard trend in off-corner region is also determined by lo heat flux and ider gap. Temperature on off-corner bolt has rising of a certain degree and temperature close to narro face corner also goes up due to ithout cooling ater slot to reduce the influence of ide Fig. 3. Temperature at measuring points of thermocouples: (a) ide face, (b) narro face. face. Thus cooling ater slot in the edge of narro face is usually designed to slope to enhance cooling capacity Longitudinal Distribution Figure 4 shos the distribution of calculated temperature and heat flux at the centre of hot copper surface in the direction of mold length. The surface temperature of copper plates ith flash elded chrome layer is higher than that ithout chrome layer and maximum difference appears in meniscus is approximately C. Though this difference seems unobvious, it is still very important to investigate microscopic interfacial phenomena in mold. The thin 2009 ISIJ 1358

4 chrome layer should not be neglected, but its influence had never been considered in past reports. In addition, metal chromium possesses higher hardness and effect of chrome layer on thermal distortion and residual stress of mold all remains to be further researched. The heat flux of narro face is higher than that of ide face along the mold casting height in Fig. 4 through calculating heat equilibrium and there is a peak heat flux at meniscus here a peak temperature also can be achieved, and another peak temperature is at boundary of copper and nickel due to a loer nickel thermal conductivity hich induces a higher temperature in loer mold portions. There are to valley temperature on the hot copper surface, one corresponds to minimum heat flux in higher mold portions above nickel layer and another appears nearby mold bottom due to terminal of cooling ater slots can not completely arrive mold exit. Temperature of ide face is higher than that of narro face ithin 100 mm belo meniscus and maximum difference may be reach 6 C, it is because ider gap beteen mold and slab in narro face caused by greater solidifying shell shrinkage has reduced heat transfer. In direction far aay from meniscus, temperature of narro face is higher than that of ide face gradually due to improved heat-transfer capability induced by mold taper of 1.1% per unit length in narro face but ithout taper in ide face. Fig. 4. Calculated temperature and heat flux at centre of hot copper surface in direction of mold length: (a) ith Cr layer, (b) ithout Cr layer, (c) heat flux Transverse Distribution The temperature distribution of transverse sections of hot copper plates is shon in Fig. 5. The selected sections lay in positions of 50 mm belo meniscus and 30 mm above nickel layer (i.e. 270 mm belo meniscus) and calculation results are taken from representative regions A, B and C shon in Fig. 1(a) and idth of regions B and C are all 144 mm. The calculation results indicate that temperature distributions in different heights are similar and positions of Fig. 5. Temperature distributions of transverse sections of hot copper plates: (a) belo meniscus 50 mm of narro face, (b) belo meniscus 50 mm of ide face midst, (c) belo meniscus 50 mm of ide face corner, (d) belo meniscus 270 mm of narro face, (e) belo meniscus 270 mm of ide face midst, (f) belo meniscus 270 mm of ide face corner ISIJ

5 coating nickel layer are also like this. As shon in Figs. 5(a) and 5(d), temperature of narro face from centre to off-corner deep slot is reduced gradually and temperature of bolt position is higher than that of cooling ater slots due to large slots interval of both sides of bolt. The temperature of copper plate in bolt position decreased rapidly due to loer heat flux and ider gap beteen mold and slab but temperature from corner deep slot to narro face corner raised oing to profile plane of narro face covered by ide face (i.e. copper plate and steel backup of narro face all keep in touch ith hot surface of ide face). Temperature of ide face midst regards bolt as centre and distributes symmetrically and an almost linear temperature gradient is established beteen hot surface and roots of cooling ater slots as shon in Figs. 5(b) and 5(e). As shon in Figs. 5(c) and 5(f), temperature of corner bolt position of ide face is loer than that of midst due to loer heat flux and temperature close to ide face corner reduced obviously oing to the part that ide face touched ith narro face is still assigned cooling ater slots and ide face corner just under cooling ater slots beteen to binding bolts. Fig. 6. Temperature at centre of hot copper surface under different casting speed: (a) ide face, (b) narro face Effect of Casting Speed Though it is almost impossible to adjust mold structure that put into production, understanding regularity that temperature of mold copper plates varying ith casting speed combined ith actual operating conditions is essential and significant. The principia and characteristic of mold primary cooling in real production can be described as: (1) initial temperature of imported cooling ater should be loer than 38 C; (2) temperature difference beteen imported and exported cooling ater in ide face should be loer than 8 C and that of narro face should be loer than 10 C; (3) flo rate of primary cooling ater under maximum casting speed is invariable and (4) temperature difference beteen imported and exported cooling ater is linearly increased 0.3 C ith increment of casting speed 0.1 m min 1. On the basis of above-mentioned, temperature of mold copper plates at casting speed 1.8 m min 1 and 2.0 m min 1 ere calculated repeating previous mathematical models and shon in Fig. 6, here 2.0 m min 1 is maximum steady casting speed for lo-carbon steel and making use of temperature of centre of hot copper surface to reflect effect of casting speed due to similarity and continuity of temperature distribution. The temperature distributions at casting speed 1.8 m min 1 and 2.0 m min 1 are similar to that of casting speed 1.6 m min 1 and regularities of them can be described according to descriptions for Fig. 4 and Fig. 5. The centre temperature increased C and C as casting speed increased to 1.8 m min 1 and 2.0 m min 1 from 1.6 m min 1 respectively and maximum temperature differences all appear at meniscus. The centre temperature is not increased linearly ith casting speed because heat flux based on heat equilibrium calculation is not linearly. Temperature of hot copper surface directly influences contact state beteen mold and casting metal and then decides surface quality and castability of various steel grades. Combining calculation results above and slab surface temperature from on-line monitoring system of continuous Fig. 7. Temperature difference at centre under different casting speed: (a) slab surface temperature, (b) temperature difference. caster, temperature difference at centre beteen mold all and slab under different casting speed ithin 200 mm belo meniscus has been determined and shon in Fig. 7. The slab surface temperature increased ith casting speed and maximum promotions appear in meniscus are 10.8 C and 22.1 C as casting speed increased to 1.8 m min 1 and 2.0 m min 1 from 1.6 m min 1 respectively and promotions far from meniscus greater than 50 mm in range of C as shon in Fig. 7(a). The effect of casting speed on temperature difference ithin 50 mm belo meniscus is obvious and difference promotions ith increment 0.2 m min 1 of casting speed in range of C as shon in Fig. 7(b) and this fluctuation of temperature difference in meniscus should be an important influence factor to deteriorate casting effectiveness. 4. Conclusions (1) The thin flash elded chrome layer in order to avoid permeating copper element has influence on temperature of mold copper plates. The surface temperature of hot copper plates ith chrome layer is higher than that of ithout chrome layer and maximum temperature difference in range of C in meniscus. (2) There are to peak temperatures on hot copper surface due to higher heat flux and different material properties and to valley temperatures due to loer heat flux and 2009 ISIJ 1360

6 geometry of cooling ater slots. Temperature on hot copper surface of ide face is higher than that of narro face ithin 100 mm belo meniscus caused by greater shell shrinkage and it just opposites far from meniscus caused by mold taper in narro face. (3) Under steady casting state ith a high casting speed, temperature distributions in different transverse sections of hot copper plates along effective casting height of mold are similar and its characteristic is decided by co-action of heat flux, contact state beteen mold all and slab and mold structure, for example, interval and depth of cooling ater slots and arrangement of binding bolts. (4) The centre temperature of hot copper surface increased C and C as casting speed increased to 1.8 m min 1 and 2.0 m min 1 from 1.6 m min 1 respectively and temperature distributions under different casting speed are all similar and surface temperature is not increased linearly ith casting speed. (5) The effect of casting speed on temperature difference ithin 50 mm belo meniscus is obvious and difference promotions ith increment 0.2 m min 1 of casting speed in range of C. The fluctuation of temperature difference in meniscus should be an important influence factor to deteriorate casting effectiveness as improving casting speed. Acknoledgements The authors are especially grateful to China Postdoctoral Science Foundation ( ) and Postdoctoral Science Foundation of Northeastern University ( ). REFERENCES 1) C. Cho, I. V. Samarasekera, B. N. Walker and G. Lockhart: Ironmaking Steelmaking, 29 (2002), 61. 2) B. G. Thomas, M. Langeneckert, L. Castella, M. Dziuba, G. D. Gresia and W. Balante: Steelmaking Conf. Proc., Vol. 85, ISS, Warrendale, PA, (2002), 87. 3) X. D. Liu and M. Y. Zhu: ISIJ Int., 46 (2006), ) M. J. Lu, Y. Y. Kuo, T. H. Ong and C. H. Lin: Steelmaking Conf. Proc., Vol. 79, ISS, Warrendale, PA, (1996), ) I. V. Samarasekera and J. K. Brimacombe: Can. Metall. Q., 18 (1979), ) J. K. Brimacombe and I. V. Samarasekera: Iron Steelmaker, 10 (1979), 20. 7) I. V. Samarasekera and J. K. Brimacombe: Ironmaking Steelmaking, 9 (1982), 1. 8) J. K. Park, I. V. Samarasekera, B. G. Thomas and U. S. Yoon: Steelmaking Conf. Proc., Vol. 83, ISS, Warrendale, PA, (2000), 9. 9) B. G. Thomas, G. Li, A. Moitra and D. Habing: Steelmaking Conf. Proc., Vol. 80, ISS, Warrendale, PA, (1997), ) Y. Meng and B. G. Thomas: Metall. Mater. Trans. B, 34B (2003), ) T. G. O Connor and J. A. Dantzig: Metall. Mater. Trans. B, 25B (1994), ) C. Cicutti, M. Valdez, T. Perez, G. D. Gresia, W. Balante and J. Petroni: Steelmaking Conf. Proc., Vol. 85, ISS, Warrendale, PA, (2002), ) X. N. Meng and M. Y. Zhu: Ironmaking Steelmaking, 36 (2009), ISIJ