Flow Behavior in the Slab Mold under Optimized Swirling Technology in Submerged Entry Nozzle

Transcription

1 ISIJ International, Vol. 58 (2018), ISIJ International, No. 7 Vol. 58 (2018), No. 7, pp Flow Behavior in the Slab Mold under Optimized Swirling Technology in Submerged Entry Nozzle Jin CHEN, Zhijian SU,* Donggang LI, Qiang WANG and Jicheng HE Key Laboratory of Electromagnetic Processing of Materials (Ministry of Education), Northeastern University, Shenyang, P. R. China. (Received on February 1, 2018; accepted on March 27, 2018) Flow behavior in the slab mold by optimized swirling flow generation in the submerged entry nozzle (SEN) is investigated. Therefore, hydromechanical model experiment was carried out to simulate this metallurgical process, the effect of turning around the nozzle with swirling and different bottom shape of nozzle is especially analyzed. The flow behavior in the mold was recorded by dye tracer and camera; The two dimension average velocity distribution in the slab mold was measured by the Ultrasound Doppler Velocimetry (UDV). The velocity distribution agrees well with flow pattern recorded by camera. The swirling flow in SEN can reduce the impinging to the narrow face and slag entrapment is inhibited by swirling. Furthermore, the flow in the mold was optimized with swirling in SEN as well as turning around the nozzle outlet in reverse, which avoids the impinging to both the wide and narrow faces of the mold at the same time. It is a potential technology for continuous casting. KE WORDS: swirling flow; UDV; flow pattern; submerged entry nozzle; slab billet. 1. Introduction It is essential as well as vital to require the higher quality of the billet today. One of the major measures to raise casting efficiency and enhance the quality of billet is effective control flow behavior of molten steel in the mold. It indicates that the swirling flow in the submerged entry nozzle (SEN) has a great effect on improving the uniformity and stability of the outlet flow of nozzle, due to reducing the meniscus fluctuation and homogenizing the distributions of flow and temperature in the mold. 1 5) And outlet flow angle and outlet port design of swirling flow SEN for slab caster is done for further improvement of flow pattern in the mold but it cannot solve the problem the flow impacting the wide face. 6,7) Furthermore, the commercial test of swirling flow continuous casting is conducted, in which a refractory swirling blade was installed in SEN, in the case of slab, subsurface pinholes and sliver defects are remarkably reduced, but it is liable to cause clogging in SEN, which limits the further application. 8 10) A new method for swirling flow generation in nozzle, without the problem of clogging and easy for controlling swirling velocity has been proposed by our team in ) In previous work, the flow and temperature fields in the mold and SEN are studied numerically when this kind of electromagnetic swirling flow in SEN. 12) However, not only the velocity was only partly measured near the meniscus and nozzle outlet in the experiments above but also both two swirling methods cause flow impacting the wide face, which is not good for flow control. Besides, * Corresponding author: zhijian_su@epm.neu.edu.cn DOI: other flow control technologies such as M-EMS, EMBR, new type SEN are effective and the flow patterns in the mold with these technologies are different ) But these technologies cannot solve this problem either. It needs to be solved by other optimization of the swirling method. And it is difficult to measure the velocity distribution in practical continuous casting process, then model experiments using low temperature alloy or water are widely adopted until this day ) Therefore, hydromechanical model experiment for slab continuous casting is carried out to understand how to optimize the swirling technology. The swirling flow is generated by mechanical blade as okoya et al. in this study. 1 5) Meanwhile, the Ultrasonic Doppler Velocimetry (UDV) is widely used for velocity measurement in model experiment ) It is adapt for both opaque and transparent flow. In the swirling case and the optimized case-turning around the nozzle outlet in reverse (TNO), the 2D average velocity distribution in the slab mold is measured by UDV and the flow behavior in the mold is investigated by tracer and camera. The study reveals the flow behavior in the new metallurgical technology-swirling process, especially the optimized method and provides the optimum parameters. 2. Experimental Procedure According to the practical metallurgical process at certain Steel, a series of 1:2 hydromechanical model experiments are carried out, it includes tundish, mold, nozzle, circulating line, pump, rotor flow meter and valve. The major experimental setup (the slab mold and SEN) is shown in Fig. 1. The swirl blade is designed as that done by okoya to generate the swirling flow in nozzle. 1 5) The inner size of the 1242

2 Fig. 1. Schematic of experimental setup ((a) Structure of the nozzle and (b) Arrangement of UDV probes in experimental process). (Online version in color.) Table 1. UDV Parameters and Sensor Arrangement in the Experiment. Frequency Time resolution Space resolution Sound velocity in water 0.5 MHz ms 4.43 mm m/s Flow of water, m 3 /h Casting speed, m/min x y z Unit: mm Inner size of the mold below the meniscus 25, 0, mm 25, 0, mm Measurement of the sensor arrangement, for measuring x velocity component, the measure points are totally 14 For measuring z velocity component, half of the wide face , 0, 25 25, 0, mm (upward edge of nozzle outlet) 25, 0, mm the center of nozzle outlet) 25, 0, mm (the bottom edge of nozzle outlet) 25, 0, mm (the bottom of the nozzle) 25, 0, From mm and the measure points are 8 the sensors between every measure point is 10 mm, and the measure points are totally 30. mold is mm, the wall thickness is 10 mm, the inner diameter of the upper part above the blade is 46 mm, the outer diameter is 66 mm, the inner diameter of the under part below the blade is 27.5 mm, the outer diameter is 55 mm, the length of nozzle is 445 mm, the immersion depth is mm. The swirl blade is fixed at the position 152 mm from the top edge of nozzle outlet. In the experimental process, the flow is controlled by pump, valve and stopper. According to dimensionless number Fr ( =v 2 /(gl), v is velocity, g is the gravity, l is half wide face length of the mold), the scale coefficient of velocity in model experiment is 0.7. The flow rates of water is product of velocity and cross section of the mold ( mm) in model experiment. Then flow rates of water in model experiment and corresponding prototype casting speed is listed in Table 1. The video camera records flow behavior from the front view and the top view. UDV sensors is arranged (in x, y, z directions) to measure the velocity distribution in the mold (shown in Fig. 1(b)). The velocity distribution in xz plane (front view) is measured by UDV, one profile is along the center line of narrow face ( =0), and other two profiles are symmetrically parallel to the center line (=25 mm, = 25 mm). The UDV parameters and the arrangement of sensors are shown in Table 1. (the sensor arrangement, for velocity x component measurement, the measure points are totally fourteen; for velocity z component measurement, half of the wide face, the sensors are arranged starting 10 mm from the narrow face to the nozzle outlet. Distance between every measure point is 10 mm, and the measure points are totally thirty) 3. Experimental Results and Discussion The parameters & experimental conditions and visualization results are shown in Table 2. The flow is visible by dye tracer and is recorded from the front view and top view (shown in Figs. 2 and 3). The SEN bottom shape (convex and concave), the casting speed( m/min)/flow rates 1243

3 ISIJ International, Vol. 58 (2018), No. 7 Table 2. Parameters &Experimental Conditions and Visualization Results. Exp. Nozzle type swirling Casting speed, m/min Flow rates of water, m3/h Immense depth, mm ATNO α and β, θ, No. 1 N No No. 3 Clockwise No. 4 N No No. 6 Clockwise No. 7 N Clockwise 5.0 No Clockwise 4.0 No. 9 N Clockwise 10.0 No Clockwise 10.0 No Clockwise 10.2 No Clockwise 5.0 No Clockwise 5.0 No Clockwise 10.0 No Clockwise 11.1 No Clockwise 12.5 No Clockwise 12.0 Fig. 2. The photos of flow field in slab mold (Front view) (Convex bottom: (a) NO swirling (b) swirling (c) Swirling + (TNO); Concave bottom: (d) NO swirling (e) swirling (f) Swirling + (TNO)). (Online version in color.) of water ( m3/h), the immersion depth (50, and 100 mm) and the angle of turning around nozzle outlet for and nozzle (ATNO α and β ) are considered. The flow outlet from the nozzle forms a strong jet, the nozzle outlet is parallel to the wide face (WF) and jet directly impacts the narrow face (NF) without swirling. When it impacts NF, it forms the two flow, one goes upwards which is upwards flow, and another goes downwards which is downwards flow. At the 0.3 s, the angle between the two downwards streams from the outlet is θ (shown in Fig. 2(a)). θ is smaller with swirling (value listed in Table 2). The probable reason: the jet out of nozzle becomes divergent as fan-shaped flow and direction of the flow points to the wide face, the velocity of jet along x and z direction vx and vz (vertical and horizontal) decreases with 1244

4 Fig. 3. The photos of flow field in slab mold (Top view) (Convex bottom: (a) NO swirling (b) swirling (c) Swirling + (TNO); Concave bottom: (d) NO swirling (e) swirling (f) Swirling + (TNO)). (Online version in color.) Fig. 4. The relation between ATNO and (a) Flow (b) immersion depth. (Online version in color.) swirling, and the decrement of v z is less than v x. The flow trace at the top view is the projection of the downward jet flow, along WF from the nozzle to NF without swirling ((a) and (d)); it impacts the quarter position at WF at the 0.35 s with swirling ((b) and (e)) and v x decreases; it is also along WF from the nozzle to NF and v x decreases with swirling + TNO in clockwise α and β ((c) and (f)). It is not good for flow control if it impacts WF with swirling. ATNO is obtained after turning around the nozzle outlet in clockwise direction until the jet flow parallel to WF again. (ATNO value α and β of Exp. No. 3, No. 6 to No. 17 listed in Table 2). The relation between ATNO and the flow rates of water (casting speed), the immersion depth is obtained (shown in Fig. 4). There is corresponding ATNO for the different flow rates of water and immersion depth. ATNO (α and β ) is proportion to the flow rates of water, and there may be a minimum value of α and β between the immersion depth 50 mm to 100 mm (in this paper it should be mm). The relation is given by using the least square regressed empirical correlation equation. That is, the adapt angle should be obtained and used in the practical industry F D (1) F D (2) α and β is the angle of turning around nozzle outlet for convex and respectively, F is the flow rates of water, D is the immersion depth. And the scale coefficient of velocity in model experiment for casting speed is 0.7. The flow field of three profiles parallel to WF is obtained, 2D average velocity vector in xz plane are shown in Figs The swirling direction is anti-clockwise, the profile 1(=25 mm) is at the upstream part, the profile 3( = 25 mm) is at the downstream part, the profile 2( =0 mm at the middle part) is between them, and the three profiles are parallel. The average velocity vectors in three profiles without swirling are similar shown in Fig. 5 (). The flow outlet from the nozzle impacts NF, then forms the two flows, one goes upwards and another goes downwards upwards and downwards flow, which veirfies the flow visualization result. In the case with anticlockwise swirling, the vectors in three profiles are not similar again, the profile 1 is the 1245

5 ISIJ International, Vol. 58 (2018), No. 7 Fig. 7. Fig. 5. Velocity vectors in the mold (no swirling, ). (Online version in color.) Velocity vectors in the mold (swirling +TNO, ). (Online version in color.) Fig. 6. Velocity vectors in the mold (swirling, ). (Online version in color.) Fig. 8. Velocity vectors in the mold (no swirling, ). (Online version in color.) upstream part and velocity decreases, the profile 3 is the downstream part and velocity increases, velocity value in the profile 2 is between profile 1 and 3 shown in Fig. 6 (). The vector at upstream part in the profile 1 changes obviously, there is no classic upwards flow and downwards flow, the vortices center of the upwards flow 1246

6 Fig. 9. Fig. 10. Velocity vectors in the mold (swirling, ). (Online version in color.) Velocity vectors in the mold (swirling +TNO, concave bottom). (Online version in color.) disappears because the major flow is not out from the nozzle but the flow impacts the mold wall then turn to the nozzle. The case with swirling +TNO in clockwise α is shown in Fig. 7. The vectors in three profiles are similar again. It means that turning around the nozzle makes flow parallel to WF again, which avoids the flow impacting WF directly. It is different from the, the upward flow decreases and fluctuation becomes bigger near the meniscus, the center of upward flow vortex is higher in case shown in Fig. 8. It is similar to the nozzle case, there is little effect of the nozzle bottom with swirling, that is, the swirling is the main reason for flow behavior shown in Fig. 9. There is an intresting phenomenon especially around the meniscus with swirling +TNO case shown in Fig. 10. It is very different from the case. The velocity near the mensicus increases and is larger than that in no swirling case and only swirling case. But no matter convex bottom and, both swirling and swirling +TNO case, the velocity of jet along x and z direction v x and v z (vertical and horizontal) decreases with swirling, and the decrement of v z is less than v x, which verifies the flow visualization result. For better description and discussion of the flow behavior, the velocity curve at different position along Z axis is given in profile 2. Velocity x curve in the mold from Z =10 mm to 130 mm below the meniscus with the convex and nozzle in profile 2 is shown in Fig. 11. (1) No swirling case (a) (d), 10 mm mm below the meniscus, v x is obviously influenced by the nozzle bottom shape. v x is usually symmetry at 1/4 WF and v xmax occurs around 1/4 WF in the nozzle case, and v x is very different in the nozzle case, it is small at 10 mm below the meniscus and becomes minus at 40 and mm below the meniscus from 1/4 WF and nozzle to NF, respectively; 87 mm 130 mm below the meniscus, the effect of the nozzle bottom shape becomes small, especially at 130 mm below the meniscus, v x is symmetry at 1/4 WF and both v xmax occurs at 1/4 WF. (2) Swirling case (b) (e), there is almost no effect of nozzle bottom shape on the velocity and the leading effect of swirling on the velocity, which is also verified by okoya. v xmax shifts toward NF at 10 mm 40 mm below the meniscus and toward the nozzle at 104 mm 130 mm below the meniscus respectively, v x is almost equal to mm and 87 mm below the meniscus; v x is smaller than that in no swirling case. (3) Swirling +TNO case, v x is obviously different between the concave and convex nozzle bottom in upward flow area; case, v x =0 at 10 mm below the meniscus, at 1/4 WF, v x 200 at 40 mm below the meniscus and no obvious maximum value of v x, v x =0 at mm below the meniscus; case, v x increases than that of no swirling case and swirling case, at mm below the meniscus, v x =0 at 87 mm below the meniscus, v xmax shifts toward the nozzle at 104 mm 130 mm below the meniscus, that is, upward flow increases, vortex center go down. And the effect of the nozzle bottom shape becomes small and the effect of swirling becomes large at 130 mm below the meniscus, both v xmax shift toward the nozzle. In sum, swirling decreases v x, swirling + TNO decreases v x in near the meniscus, but increases v x in concave bottom, and the effect of bottom shape gradually disappears and the effect of swirling appears at 130 mm below 1247

7 Fig. 11. Velocity x curve in the mold from Z =10 mm to 130 mm below the meniscus. (Online version in color.) Fig. 12. Velocity z curve in the mold from Z =10 mm to 130 mm below the meniscus. (Online version in color.) the meniscus. That is, swirling+tno case is influenced by both the swirling and the nozzle bottom shape. Velocity z curve in the mold from Z=10 mm to 130 mm below the meniscus with the convex and nozzle in profile 2 is shown in Fig. 12. It is shown that v z changes in a small range, in swirling case (b) (e) and swirling+tno case (c) (f) is smaller than that in no swirling case (a) (d). And v z becomes positive at 1/4 WF, at 130 mm below the meniscus. Swirling effect on slag entrapment is also analyzed. Slag entrapment occurs if surface velocity is larger than the critical velocity m/s for different steel grades. 21) Swirling reduces the meniscus fluctuation and homogenizing the distributions of flow, which inhibits slag entrapment. The velocity v z 0 at the position v xmax exists near the meniscus, that is, v max v xmax (shown in Figs ). Therefore, Fr xmax (=v 2 2 /(gl) v xmax /(gl)) near the meniscus is calculated by v xmax at 10 mm below the meniscus, then v xmax in steel is obtained to estimate if the value is smaller than the critical velocity (shown in Table 3). It is shown that v xmax is nearly smaller than 0.5 m/s when casting speed is m/min; but at high casting speed m/min, v xmax is larger than m/s for different steel grades without swirling; and with swirling the velocity is almost less than 0.5 m/s. That is, slag entrapment does not occur with swirling technique, and some steel grades should be combined with other flow control technique to enhance surface velocity to avoid stagnation on the meniscus. And the casting speed is even 7.5 m/min in some condition, 19) only one technique cannot control the flow good enough. Besides, electromagnetic 1248

8 Table 3. Fr xmax and v xmax at Z = 820 mm of Liquid Steel and Water. Z = 820 mm No. 1 No. 2 No. 3 No. 4 No. 5 No. 6 casting speed: m/min (Flow in water: m 3 /h), l = m, g = 9.80 m/s 2 v xmax in water m/s Fr xmax v xmax in steel m/s casting speed: 1.80 m/min (Flow in water: 3.82 m 3 /h), l = m, g = 9.80 m/s 2 v xmax in steel m/s casting speed: 2.10 m/min (Flow in water: 5.00 m 3 /h), l = m, g = 9.80 m/s 2 v xmax in steel m/s swirling technology in SEN does not only optimize the whole velocity vectors in the mold but also solves scabbing in the nozzle. Swirling in SEN is a potential method and combining other technologies is the further work in the continuous casting. 4. Conclusions In this paper, the effect of swirling and turning around the nozzle outlet and nozzle bottom shape on flow behavior was analyzed and main conclusions are as follows: (1) The relation between the angle of turning around the nozzle and the casting speed, the immersion depth for both convex and case is obtained by ink tracer visualization experiment. The relation is given by using the least square regressed empirical correlation equation F D (1) F D (2) α and β is the angle of turning around nozzle outlet for convex and respectively, F is the flow rates of water, D is the immersion depth. (2) The angle θ between flows discharged from the nozzle outlet decreases by swirling flow. It is because the jet out of nozzle becomes divergent as fan-shaped flow and direction of the flow points to the wide face, then v decreases (component of v, Δv z <Δv x ) and upward flow also decreases. (3) The flow in the slab is optimized with swirling in SEN as well as turning around the nozzle outlet in reverse, which avoids the impinging to both wide and narrow faces of the mold at the same time and maintains the advantage by swirling. The optimized method is also influenced by the nozzle bottom shape. Upward flow and surface velocity decreases with swirling+tno if the nozzle bottom is convex but it increases if the nozzle bottom is concave. And the effect of bottom shape gradually disappears and the effect of swirling appears at 130 mm below the meniscus. That is, swirling+tno case is influenced by both the swirling and the nozzle bottom shape. (4) It is shown that with swirling the surface velocity is almost less than critical value, slag entrapment does not occur. And some steel grades should be combined with other electromagnetic control technique to enhance the surface velocity to avoid the stagnation. In sum, swirling in SEN changes the flow behavior in the mold and slag entrapment never occurs near the meniscus. Optimized swirling method has better effect. It is very important and potential for practical continuous casting using the swirling technique. Acknowledgements This paper is financially supported by National Natural Science Foundation of China [ ] and [U ]. REFERENCES 1) S. okoya, R. Westoff,. Asako, S. Hara and J. Szekely: ISIJ Int., 34 (1994), ) S. okoya, S. Takagi, M. Iguchi,. Asako, R. Westoff and S. Hara: ISIJ Int., 38 (1998), ) S. okoya, S. Takagi, M. Iguchi, K. Marukawa, W. asugaira and S. Hara: ISIJ Int., 40 (2000), ) S. okoya, S. Takagi, S. Ootani, M. Iguchi, K. Marukawa and S. Hara: ISIJ Int., 41 (2001), ) S. okoya, P. G. Jönsson, K. Tada, S. Takagi and M. Iguchi: Scand. J. Metall., 33 (2004), 22. 6). Tsukaguchi, O. Nakamura, S. okoya and S. Takagi: Tetsu-to- Hagané, 91 (2005), ). Tsukaguchi, O. Nakamura, S. okoya and S. Takagi: Tetsu-to- Hagané, 91 (2005), ). Tsukaguchi, H. Hayashi, S. okoya, T. Tanaka and S. Hara: Tetsu-to-Hagané, 93 (2007), ). Tsukaguchi, H. Hayashi, H. Kurimoto, S. okoya, K. Marukawa and T. Tanaka: Tetsu-to-Hagané, 95 (2009), ). Tsukaguchi, H. Hayashi, H. Kurimoto, S. okoya, K. Marukawa and T. Tanaka: ISIJ Int., 50 (2010), ) J. C. He, K. Marukawa and Z. J. Su: Submerged Entry Nozzle with Swirling Flow by Electromagnetic Field, CN Patent, , (2005). 12) Z. J. Su, D. W. Li, L. W. Sun, K. Marukawa and J. C. He: Acta Metall. Sin., 46 (2010), ) H. Q. u and M.. Zhu: ISIJ Int., 48 (2008), ) K. Okazawa, T. Toh, J. Fukuda, T. Kawase and M. Toki: ISIJ Int., 41 (2001), ) S. Sarkar, V. Singh, S. K. Ajmani, R. Ranjan and K. Rajasekar: ISIJ Int., 56 (2016), ) B. J. Dong and M.. Zhu: Steel Res. Int., 86 (2015), ) Q. Q. Wang and L. F. Zhang: JOM, 68 (2016), ) H. amamura, K. Amada, K. Takase and T. Matsumiya: Tetsu-to- Hagané, 98 (2012), ) R. D. Morales,. Tang, G. Nitzl, C. Eglsäeer and G. Hackl: ISIJ Int., 52 (2012), ) X. Jin, D. F. Chen, X. Xie, J. L. Shen and M. J. Long: Steel Res. Int., 84 (2013), ) H. Lei, M.. Zhu and J. C. He: Acta Metall. Sin., 36 (2000), ) H. H. Jia, Z. u, Z. S. Lei, K. Dang, J. C. Chen, W. J. Hua and Z. M. Ren: Acta Metall. Sin., 44 (2008), ). H. Li, Z. J. Su and J. C. He: J. Iron Steel Res. Int., 17 (2010), ) K. Timmel, S. Eckert, G. Gerbeth, F. Stefani and T. Wondrak: ISIJ Int., 50 (2010), ) S. Franke, L. Büttner, J. Czarske, D. Räbiger and S. Eckert: Flow Meas. Instrum., 21 (2010), ) D. Brito, H. C. Nataf, C. Philippe, J. Aubert and J. P. Masson: Exp. Fluids, 31 (2001), ) S. Eckert, B. Willers and G. Gerbeth: Metall. Mater. Trans. A, 36 (2005), ) O. Andreev,. Kolesnikov and T. André: Exp. Fluids, 46 (2009), ) X. D. Wang,. Fautrelle and J. Etay: 18th French Cong. Mechanics, French Association of Mechanics, Courbevoie, France, (2007), 27.