Dynamic Water Modeling and Application of Billet Continuous Casting

Available online at www.sciencedirect.comScienceDirectJOURNAL OF IRON AND STEEL RESEARCH, INTERNATIONAL.2008, 15(2); 14-17Dynamic Water Modeling and Application of Billet Continuous CastingLIU Wen-hong'，XIE Zhi'，J1 Zhen-ping' ,WANG Biao'，LAI Zhao-yi", JIA Guanglin'(1. School of Information Science and Engineering, Norheastern University, Shenyang 10004, Liaoning, China;2. Sanming Iron and Steel Corp, 365000 Fujian, China; 3. School of Materials and Metallurgy,Northeastern University, Shenyang 110004, Liaoning, China)of secondary cooling is extremely important for output of the casting machine and bllet quality. A dynamie watermodel was introduced, including solidification model in the secondary cooling, feedforward control strategy based oncontinuous temperature measurement in tundish, and feedback control strategy based on surface temperature meas-urement. The mathematical model of solidification process was developed according to the principle of solidification,and the solidification model was validated by measuring bllet shell thickness through shooting nail and sulfur print.Primary water distribution was calculated by the solidification model according to procedure parameters, and it wasadjusted by the other two control strategics online. The model was applied on some caster and bllt quality was obvi-ously improved, indicating that the dynamic water model is better than conventional ones,Key words: billet continuous casting; secondary cooling; water model; bllt qualityThe continuous casting process is the method In all the methods, the last one is more advancedfor the solidification of molten steel into semi-fin- than the others. However, it is necessary to im-ished shapes, such as blooms, bllets, and slabs1.prove the control technology.Fig.1 shows the bow bllet caster, found in manyDue to reheating and improper cooling profilessteel plants.in the secondary cooling zone (SCZ), surface and in-The control technology of the secondary cooling ternal cracks form, affecting the billet qualty3.J.water is extremely important for output and billetLittle research has been done on the application ofquality. To provide the background to this study, a control theory to the SCZ. A probable reason is thatbrief description is given. There are mainly threea closed loop control structure is difficult to imple-control methods of water distribution in internal ment since reliable continuous measurements of tem-casterse?. The first method, manual control, is de-perature profiles along the strand are not feasible.fined that the operator determines water distribur The amount of steam in the SCZ causes inaccuratetion, according to the experience and online condi-optical pyrometer measurements. However, surfacetions. The second method, water flowrate control or temperature measurement at the exit from the sCzproportion control, is a method, where dfferent wa- has been studied and applied in recent years, and theter flowrate and water dstribution are determined measured surface temperature can relet the tem-for dfferent steel grade, and then a table or water perature trend in the SCZ on the whole, which ofersflowrate-casting speed simple equation by data re the psbilitit to design a new water model system.gression is obtained. The third method, water flow-The superheat of liquid steel is one of the keytate-casting speed conic control or parameter con- factors afcine the bllet quality and the output. Alongtrol, is defined that water distribution is calculatedwith the superheat increasing, the shell thickness be-by sldification model and water flowratecasting comes thin, the surface temperature increases, andspeed conic equation is obtained by data regression.the li中国煤化工even if at the sameFoundation lem; Item Sponsored by National H-Tech Research and Development F;1HCNMH GBiogrnphy: LIU Wen-hong(1978-), Female, Doctor Candidate; E- mil; wenhong 1u040410353 com, Kevisea Date: February 5, 2007Issue 2Dynamic Water Modeling and Application of Billet Continuous Casting●15●In the radiation zone-入工T=eo(T-I)(5). Ladlewhere, T。is casting temperature; a, b are con-Continuous tenperaturestants; h is heat transfer coefficient; Tw is tempera-measurement in tundishture of cooling water; Tb. is surface temperature ofTundish, Stopper rodbillet; ε is emissivity; σ is Stefan-Boltzman con-Mould、stant; and n is exterior normal of cooling boundary.MeniscusLiquid phaseTo eliminate the effect of superheat variation,- Rollersthe feedforward control strategy is proposed basedon continuous temperature measurement in tundish.MetallurgicalCutterIt is assumed that the other procedure parametersSecondarycoolingStrandare invariable, water compensation is calculated byoodmodel and the coefficient values of equations with000 Cone variable are obtained through linear regression.Fig1 Bow billet casterThe integrated equation is given as:sQ,(T)=a;(T。- T)+b;, i=l,2,*,m (6)casting speed5. Thus, the superheat is taken intowhere 0Q,(T,) is the compensative water flowrateaccount for the water model system,of segment i; T. is casting temperature; T is liqui-In this study, the water distribution aspects ofdus temperature; a; and b; are coefficients, and i isthe SCZ was investigated, and the dynamic waterthe number of segments in the SCZ.model system was built based on solidification modelFor calculating the water compensation, a pre-in the SCZ, feedforward control strategy, and feed-cise soldification model is first needed. The shoot-back control strategy. It is assumed that sufficienting nail and sulfur print are adopted here to measurecooling has been realized in the mold.the actual shell thickness and validate the model for1 Dynamic Water Modelthe first time in China.One of the metallurgical criteria is that the surfaceThe soldification model is developed accordingtemperature of the strand before straightener rollsto the metallurgical criteria for billet and target tem-should be controlled out of embrittlement tempera-perature controlling principle1-.o]. To model theture region to prevent cracking. The relevant equa-mathematical treatment, it is assumed that the finitetions are given as;volume of the strand moves with the strand at a ve-T,≥T。(7)locity equal to the casting speed. The 2D heat con-T,≤Ta(8)duction equation for heat transfer and the boundarywhere T, is the surface temperature of the strand be-conditions are shown as follows:fore straightener rolls; Tm is the upper limit temper-Governing equationature; and T is the lower limit temperature accord-pxpT_寻| +s(1)ing to the metallurgical criteria.However, the conventional system is an openwherep,cp, T, t,入,工, y, and S represent densi-loop and the surface temperature of the strand at thety, specific heat capacity, temperature, time, con-exit of the SCZ is uncontrollable-7,8]. Thus, thductivity, coordinate of the thickness and width di-feedback control strategy is proposed to solve therection, and source item, respectively.problem. The controlled segments are the last two(1) Initial conditionTbeim= T。(2)ones of the SCZ, because they are closer from theexit of the SCZ and easier to be controlled. Th(2) Boundary conditionstrategy is not applied at the time of unstable opera-In the moldtions t→n=a-bri(3)中国煤化工tem is shown inIn the secondary cooling zoneFig.2.TYHC N M H Gameters, the so--λ=h(T-r.)+eo(T-TY)(4)lidification model' calculates primary water distribu-ntion, which is adjusted by the other two control stra16●Journal of Iron and Steel Research, InternationalVol. 15of the billet continuous casting. Table 1 gives theI Water Dlowrate Correlaliveshooting position and corresponding shell thickness.compensation 厂1 coeticientThe actual error is less than 2 mm, and the solidif-Waler Dowrale |Contol WaterSurfacecation model is validated.calculated byDH valvefowrate| temperature_modelTable 2 gives the water flowrate of three seg-ments in the SCZ when superheat varies and castingWater flowratespeed is 2 m/ min. For the same segment, the watermeasurementflowrate increases along with increasing superheat,Fig 2 Flowchart of dynamic water model systemand the increased water flowrate will eliminate thesurface temperature variation, because of superheattegies online. The solidification model is the base inincreasing.this control model, and it is important to validateFig. 4 shows the surface temperature curves be-the model, which is explained in detail in Section 2.fore and after applying the dynamic water model sys-tem. The curve after applying the system is relative-2 Application and Discussionly smooth and is out of embrittlement temperatureThe dynamic water model system has been ap-region, which means that the surface temperature isplied in some steel plant. Steel grade is L1008, andwell controlled by the new system.billet size is 150 mm X 150 mm. The radius is 10 mHowever, the temperature curve slightly fluc-The casting speed is 2. 32 m/ min. The casting tem-tuates at certain times. The probable factors are:perature is 1550 C, and the liquidus temperature is(1) The gathered temperature signal is dis-1 524 C. The water flowrate in the SCZ is 10. 01 t/h,torted, because of the scale and the cooling water on12. 23 t/h, and 2. 96 t/h, respectively.the surface of the strand, A filter program is proposedThe solidification model calculates surface tem-perature curve and shell thickness by inputting sev-Table 1 Shooting position and shell thicknesseral offline procedure parameters. However, theShootingDistance from MeasuredCalculatedsurface temperature in the SCZ cannot be measured.positionmeniscus/m thickness/mm thickness /mmHence, the method of measuring billet shell thick-3.0721Position 28.714847ness through shooting nail and sulfur print is pres-Position 312. 4464ented to validate the model. Fig. 3 shows the sche-matic of the solidification model.Three positions in the SCZ are selected forTable 2 Water flowrate with dfferent superheatshooting nail at the same time, and the shell thick-Superheat/Cness is separately measured after the sulfur print.Secondery cooling 2zone3542The heat transfer coefficients of the solidificationWater flowrate of segment 08. 078.358.63model are adjusted until the error between the meas-Water flowrate of segment 17.627. 90ured shell thickness and the calculated one is lessWater flowrate of segment 22. 142.212. 28than 3 mm, which is allowable for the theoretical modelInput procedureLay a course1100间paranetersShoot nail1 050Solidincation model1000I Adjust heat| Sulfur print and shell| Calculate temperaturel transferthickness100向)fheld and shell thicknessl coeffcientTNo巨1050wmmMrrEror≤3 mmYes中国煤化工4050Calculate water flowrate bythe validated solidification modelDHCN MH G_rg+ suaur enuperaturc cur& before (a) andFig 3 Schematic of solification model validationafter (b) applying control systemIssue 2Dynamic Water Modeling and Application of Bilet Continuous Casting，17●to solve the problem. It may be programmed in pro-All the work is to improve billet quality. Fig. 5gram logic controller, which makes feedback control shows the micrographs before and after applying thestrategy realized more conveniently.water model system, Central crack is severe for im-(2) The control system of continuous casting proper cooling water before the dynamic water mod-is a typical time delay system. The strand passes the el system is applied. And billet quality is obviouslySCZ in a few minutes, and the controlled segment improved after applying the proposed system, whichmay be different from the target. Smith predictor indicates that the new system is effective for billetcan deal well with the problem, though further continuous casting.study is needed.For the previous factors and proposed solu-3 Conclusionstions, some further research is necessary.The dynamic water model system for the SCZ ofaFig.5 Macrographs before (a) and ater (b) applying new systembillet continuous casting was proposed. The super-[2] ZHANG Keqiang, LIN Xian, GAO Hai, et al. Application ofLevel 2 Distribution of Spraying Water in Continuous Castingheat and billet surface temperature were taken into[J]. Iron and Steel, 2006, 41(1); 39 (in Chinese).account besides the conventional solidification mod-[3] Machara Y. Surface Cracking Mechanism of Continuously Castel, and thus, the water distribution was more appro-Low Alloy Steel Slabs [J]. Materials Science and Technology,priate than the conventional open-loop one.1990, 6(9); 793 (in Chinese).The proposed system has been applied on some4] Brimacombe J. K, Sorimachi K. Crack Formation in the Con-tinuous Casting of Steel []. Metallurgical Transactions, 1977,caster and the controlled temperature curve was bet-8B(3): 489.ter than earlier one. Billet quality was obviously im- [s] Camisani P R. Crig I K, Pstorius P c. Speed Ditubenceproved and monthly retirement amount was largelyCompensation in the Secondary Cooling Zone in Continuousreduced in the rolling mill from 80 t to 20 t. It canCasting [J]. ISU International, 2000, 40(5); 469.be found that the dynamic water model is practical[6] Louhenkilpi Seepo. Real-Time Simuation of Heat Transfer inContinuous Casting [J]. Metallurgical Transactions, 1993,and has a favorable foreground.24B(3): 685.Euntai Kim. A New Approach to Numerical Stability AnalysisReferences:of Fuzry Control Systems [J]. IEEE Trans on Systems, 2001,31(1); 107.[1] SHI Chen: xing. Practical Casting Metallurgy Technology [M].[8] Camisani F R, Craig I K, Pistorius P C. Specificationo FrameBejing, China; Metallurgical Industry Press, 1998 (in Chi-work for Control of the Secondary Cooling Zone in Continuousnese),中国煤化工38(5); 447.YHCNMHG