Water Modeling of Twin-Roll Strip Casting

Available online at www.sciencedirect.comBCIENCE@.)DIn.CT.JOURNAL OF IRON AND STEEL RESEARCH, INTERNATIONAL. 2006, 13(1): 14-17Water Modeling of Twin-Roll Strip CastingWANG Bo'.2，ZHANG Jie-yu'，FAN Jun-feit，ZHAO Shun-li2，FANG Yuan' ，AN Shenglil2(1. School of Metallurgical and Ecological Engineering, University of Science and Technology Beiing, Beijing 100083, China;2. Material and Metallurgical School, Inner Mongolia University of Science and Technology, Baotou 014010, InnerMongolia，China; 3. College of Materials Science and Engineering, Shanghai University, Shanghai 20002，China;4. Research Institute, Baoshan Iron and Steel Co Ltd, Shanghai 201900, China)Abstract: Twin roll strip casting is regarded as a prospective technology of near net shape continuous casting. Thefluid flow field and level fluctuation in the pool have a strong influence not only on composition and temperaturehomogeneity of pool, but also on the strip quality. A1 : 1 water model of a twin- roll strip caster was set up basedon the criteria of Froude number and Reynold number similarity. The level fluctuation was measured. The influ-ence of pool depth, casting speed and feeding system configuration on level fluctuation in the pool was studied.The experimental results provided a basis for the optimization of feeding system and process parameters.Key words: twin roll strip casting ; level fluctuation; water modelingTwin- roll strip casting is regarded as a prospec-field pattern influences the solidification negatively.tive technology of near net shape continuous cast-The flow field pattern is affected by the geometry ofing. The thickness of strip produced is approximate-feeding system as well, and the heat transfer in thely 1-6 mm. Although thin strip casting has advan-mold formed by two roll surfaces is affected by thetages of low energy consumption，low investmentflow field pattern.and reduced emission， it has some problems OThe flow field study of twin- roll caster was con-smooth operation and product qualityt1.ducted with a water model to optimize feeding sys-During strip casting, the flow field pattern and lev-tem.' The effects of casting conditions and geometryel fluctuation in the pool affect the strip quality-2- 4.of feeding system on movement and distribution oWhen the level fluctuation exceeds土(1.6- 2) mm, thefluid in the pool were investigated. A water modellongitudinal cracking will be observed. On the con-of 1 : 1 was built based on a twin- roll strip caster attrary，if the meniscus is too calm， the turbulenceBaosteel. The level fluctuation was measured with awould be strengthened in lower part of the pool andlevel senor. The effect of roll rotation on the levelthe surface layer would solidify. Besides， the con-fluctuation was ignored. This resulted in the differ-figuration of feeding system, casting speed and poolence of boundary conditions and flow field compareddepth also affect flow field. So it is very difficult to .with practical process. Since only the effect of nozzlecontrol flow field in the pool.configuration on level fluctuation was considered ，At present, the key factors influencing the stripthe experimental results may provide a basis for thequality are dam edge, feeding system and character-optimihe. feedino. eretem.istics of rolls. The configuration of the feeding1v中国煤化工system is of decisive importance to flow field in theMYHCNMHG.pool. The molten steel must be supplied homogene-1.1 Experimental set-upously over the pool width, because asymmetric flowThe water model is depicted in Fig. 1. Two rollsFoundation Item: Item Sponsored by Provincial Natural Science Foundation of Inner Mongolia of China (200408020715)Biography : WANG Bo( 1974-)，Male, Doctor Candidate, Lectureship; E-mail: wangbo197 4@ 263. net;Revised Date :December 2, 2004No. 1W ater Modeling of Twin Roll Strip Castingsurface tension. According to similarity theory,Froude number (Fr) and Reynolds number (Re) areconsidered to ensure the similarity of fluid flow.Re=ul, Fr=glV2-来where p is fluid density; v is fluid velocity; l is .43length; g is acceleration of gravity; and μ is kineticviscosity.The Re and Fr are approximately the same forwater at 20 C and steel at 1 600 C when the scale ofthe water model to practical caster is unity because thekinetic viscosity and flow rate are similar accordingly,i.e Remodel = Rerotype and Frmodel = Frpotrype.1.3 MethodThe feeding device was 760 mm in width，less4- Twin roll; 5- Reservoir; 6- Level senor; 7一 Flow meterthan that of the rolls， and had two rows of outletsFig. 1 Schematic of water modelfacing the rolls with seven outlets in each row. Irthis experiment， four kinds of side outlets wereare 1 000 mm in width and 700 mm in diameter. Thstudied, as shown in Fig. 2. Type A is identical toscale is 1 : 1. Two rolls, tundish, feeding systemType B，but its outlet is located higher and its twoand side dams are made of plexiglass. A new feedingoutlet are closer to side dams than Type B. Type Csystem including distribution device and feeding de-and Type D have different side outlet angle. Outletsvice was designed specially for this experiment. Theof Type C are bifurcated at a 0° angle and those ofwater in the tundish is distributed by the distribu-Type D are at 30° downward.tion device, and then flows into feeding device. TheThe flow rate was measured by a flow meterfeeding device feeds the water into mold consistingand the level was controlled by a valve. A KGY-2of two rolls and side dams. The water flows fromlevel detector was used to measure and record thethe pool to the reservoir through gap between twolevel fluctuation. The measurement under each con-rolls. In this experiment, the gap is 3 mm. Underdition was repeated three times and good reproduc-steady flow conditions， the height difference be-ibility was obtained. The level was measured intween the levels of pool and reservoir serves as pres-three positions, as shown in Fig. 3.sure drop for flow rate adjustment. Two rolls were2 Results and Discussionstationary in this experiment. The water is recycled.2.1Effect of feeding device configuration on level1.2 TheoryfluctuationTo water modeling，the geometric similarityThe effect of feeding device on level fluctuationand fluid dynamic similarity must be pursued. Inis illustrated in Table 1. In order to prevent solidifi-twin-roll strip casting process，fluid flow is mainlycation of liquid steel, Type B was used to increaseinfluenced by inertia force, gravity, viscous force andthe turbulent flow near side dams. Therefore the level(a)Type A(bType B(cType C(d)Type Ddevice中国煤化工TYHCNMHG入30°Feedingdericedeutece ”Feeding .Fig. 2 Schematic of four kinds of side outlet configurationsJournal of Iron and Steel Research, InternationalVol. 13, Feeding device4.25 |i 4.00 I3.75323.5040555Outlet diarneter/mm1Fig.4 Effect of outlet diameter on level fluctuationSide damFig. 3 Position of lever sensor6Table 1 Level fluctuation for four kinds of side outletTestingLevel fluctuation/ mnpositionType AType BTypeC Type D4.084.716.706.474. 944.834.364.773. 52fluctuation with type B was bigger than that withOutlet angle/(°)Type A.The level fluctuation is bigger for Type A be-Fig.5 Effect of outlet angle on level fluctuationcause it was closer to the surface. On the contrary,the level fluctuation is calm for Type B because itfluctuation measured in position 2. For same outletwas lower and fluid flowed downwards.diameter, there were two type outlets. One is theThe level fluctuation with Type D is smallerordinary and the other is with bell mouthed nozzle.than that with Type C. The reason is that when theFor the outlet with bell- mouthed nozzle, the outletoutlet angle is downwards，the fluid flows down-area is increased, leading to the decrease of flow ve-wards and the level fluctuation is reduced. But thislocity. So the level fluctuation should be smaller.leads to the increase of turbulent flow near solidifi-But in this case， the air can be entrained and bubb-cation region and prevents liquid steel from completeing pool was observed, especially at low level in pool.solidification near the roll gap. .2.3 Effect of process parameters on level fluctuation2.2 Effect of distribution device configuration onThe pool depth was changed by varying thelevel fluctuationcontact angle. When the contact angles are 40°，45°For the same feeding device, the effect of outletand 50° ，the level was low，normal and high，re-diameter and angle of distribution device on levelspectively. Fig. 6 shows the level fluctuation meas-fluctuation was studied. Fig. 4 shows the effect ofured in position 1. It can be seen that the level fluc-outlet diameter on the level fluctuation. It can betuation is reduced with the increase of pool depth.easily found that the bigger the outlet diameter, theThe re中国煤化工of pool depth increa-smaller the level fluctuation is. The reason is thatses thYHCN MHGin the pool, makingunder same flow rate，when outlet diameter is rethe ratio between the liquid mass in the pool and theduced, the fluid flow velocity is increased， and thefluid mass poured in unit time increase. F urther-turbulent kinetic energy is increased corresponding-more，when pool depth is increased, the fluid flowsly, leading to the increase of level fluctuation.down to the bottom of pool directly and only smallFig.5 shows the effect of outlet angle on levelproportion flows back to the surface of pool. So theNo.1W ater Modeling of Twin Roll Strip Casting17●level fluctuation is reduced. But the flow near theel fluctuation is increased with the flow rate becausebottom of the pool is enhanced, and the heat centerthe flow rate is increased, leading to more turbu-is probably descended.lence in the pool. Thus the level fluctuation can beFig. 7 shows the effect of flow rate on levelreduced by increasing pool depth or adopting suit-fluctuation. From Fig. 7, it can be seen that the levable outlet angle.3 Conclusions(1) The configuration of feeding system is ver-E5.8y important to reduce the level fluctuation and im-prove strip quality.(2) The side outlet angle and position of feed-营5.4ing device affect the level fluctuation.(3) The level fluctuation is reduced with the”5.0increase of outlet diameter of distribution device.4.8(4) The casting parameters also affect the lev-4(50el fluctuation. The level fluctuation is reduced withPool dept/(°)increasing pool depth or decreasing flow rate.Fig. 6 Level fluctuation at various pool depthsReferences:9.5[1] Cook R, Grocock P G, Thomas P M, et al. Development of theTwin Roll Casting Process [J]. Journal of Materials ProcessingMeasured in position 2Technology，1995. 55(2): 76-84.8.[2] Bouchard D, BusqueF, Hamel F G. Fluid Flow Study of aTwin Roll Strip Caster Mold Using a Water Model [A]. Steel-7.5making Conference Proeeding [C]. Chicago: IsS, 1999. 427-436.[3] Miyake S, Yamane H，Yukumoto M. Strip Quality of Highly吕6.5Alloyed Metals by Twin Roll Casting [J]. ISU International,Measured in position 31991, 31(7): 689-695. .[4] Yasunaka H, Taniguchi K, Kokita M, et al. Surface Quality of5.5Stainless Steel Type 304 Cast by Twin Roll Type Strip Caster5.0[J]. ISI International, 1995, 35(6): 784-789.14.015.016.017.0[5] Hee Kyung Moon, Dong Kyun Choo, Taewook Kang, et al.Flow rate/(m2.h)The Status of Twin Roll Strip Casting Process for Steel Strip inPOSCO/RIST [A]. Proceedings of the 60th Electric FurnaceFig. 7 Level fluctuation at various flow ratesProceedings [C]. Warrendale, PA: IsS, 2002. 499- 508.中国煤化工MHCNMHG.