MINIMIZATION OF SURFACE DEFECTS BY …

1 J. STETINA et al.: MINIMIZATION OF SURFACE DEFECTS BY increasing THE SURFACE TEMPERATURE .. MINIMIZATION OF SURFACE DEFECTS BYINCREASING THE SURFACE TEMPERATURE DURINGTHE STRAIGHTENING OF A CONTINUOUSLY CASTSLABZMANJ[EVANJE POVR[INSKIH NAPAK Z ZVI[ANJEMTEMPERATURE POVR[INE KONTINUIRNO ULITEGA SLABAMED RAVNANJEMJ osef Stetina, Tom { Mauder, Lubomir Klimes, Frantisek KavickaBrno University of Technology, Technicka 2, 616 69 Brno, Czech rokopisa received: 2012-08-31; sprejem za objavo accepted for publication: 2012-10-23 SURFACE temperatures of cast slabs on small-radius segments as well as on the unbent areas belong to the parameters that affectthe SURFACE quality of continuously cast slabs. Older machines for continuous casting were designed with regard to the quantity(the amount of cast slabs) rather than the quality of the production. Therefore, an adaptation of the secondary cooling is requiredin order to obtain the desired SURFACE temperatures .]]]]}

2 The modification consists of a dynamic control of the secondary cooling, SURFACE -temperature monitoring by means of a numerical model of the temperature field as well as a prospective replacement ofthe cooling nozzles. In order to optimize and control the secondary cooling, characteristics of the nozzles, especially theinfluences of the water-flow rate, air pressure, casting speed, SURFACE temperatures and heat-transfer coefficient under thenozzles have to be known. Moreover, the heat-transfer coefficient can also be influenced by the age of the nozzles. The paperdeals with the relationships between these influences and their impacts on the temperature field of a cast slab. The results arepresented for the 1530 mm 250 mm slabs that are cast in Evraz V tkovice Steel where the main author s dynamic, 3 Dsolidification model is used, in its off-line version, to control the production interface.

3 The results can be used for thepreparation of a real casting : optimization of the temperature field, SURFACE temperature of a slab, characteristics of nozzles, continuous castingTemperatura povr{ine ulitega slaba pri segmentih z majhnim radiusom, kot tudi na neukrivljeni povr{ini, spada k parametrom, kivplivajo na kvaliteto povr{ine kontinuirno ulitega slaba. Starej{e naprave za kontinuirno ulivanje slabov so bile pripravljene boljza ve~jo zmogljivost kot pa za kvaliteto. Zato je potrebna prilagoditev sekundarnega ohlajanja, da se zagotovi doseganje `elenetemperature povr{ine. Prilagoditev sestoji iz dinami~ne kontrole sekundarnega hlajenja, kontrole temperature povr{ine znumeri~nim modelom temperaturnega polja, kot tudi morebitna zamenjava hladilnih {ob. Optimiranje in kontrola sekundarnegahlajenja je mogo~a s poznanjem zna~ilnosti {ob in {e posebno vpliva hitrosti pretoka vode, tlaka zraka, hitrosti ulivanja,temperature povr{ine in koeficienta prenosa toplote pod {obami.}}}}}}}}}}}

4 Poleg tega na koeficient prenosa toplote lahko vpliva tudistarost {ob. Ta ~lanek obravnava odnos med na{tetimi vplivi in njihov u~inek na temperaturno polje ulitega slaba. Predstavljeniso rezultati za ulit slab 1530 mm 250 mm. Avtorjev dinami~ni 3D-model strjevanja se uporablja za kontrolo vmesnika priproizvodnji in te~e v off-line-verziji. Rezultati se lahko uporabijo kot pripravljalno orodje za realni postopek ~ne besede: optimiranje temperaturnega polja, temperatura povr{ine slaba, zna~ilnosti {ob, kontinuirno ulivanje1 INTRODUCTIONThe presented in-house model of the transient-temperature field of the blank from a slab caster (Figure1) is unique as, in addition to being entirely 3D, it canwork in real time. The numerical model covers thetemperature field of the complete length of the blank( , from the meniscus inside the mould all the waydown to the cutting torch) with up to one million concasting machine (caster) for the casting ofslabs (Figure 1) has the secondary-cooling zone sub-divided into thirteen sections due to the convection of agreater amount of heat from the voluminous slab first section engages the water nozzles from all sidesof a slab.}}}}

5 The remaining twelve sections engage air-mistcooling nozzles, positioned only on the upper and lowersides of the concasting. It is therefore very important todetermine the correct boundary conditions for a numeri-Materiali in tehnologije / Materials and technology 47 (2013) 3, 311 316311 UDK 1580-2949 Original scientific article/Izvirni znanstveni ~lanekMTAEC9, 47(3)311(2013)Figure 1:Radial caster and positions of the nozzles along the slabcaster in 13 individual zonesSlika 1:Livni stroj z radijem in pozicija {ob vzdol` naprave za uliva-nje slabov v 13 posameznih podro~jihcal model of the temperature field2taking into account areal caster that has many types of nozzles with varioussettings positioned inside a closed cage. A real castercontains a total of 8 nozzle types and geometrical lay-outs. The aim is to modify the secondary cooling zones6, 8, and 10 so as to increase the SURFACE temperature of aslab in a small radius at the point of the , the Lechler air-mist nozzles areinstalled in the cooling zones 6, 8, and 10 (Figure 2).}

6 2 MODEL OF THE TEMPERATURE FIELD OF ASLABThe presented in-house model of the transient-tem-perature field of the blank from a slab caster (Figure 1)is unique as, in addition to being entirely 3D, it can workin real time. It is possible to adapt its universal code andapply it to any slab caster. The numerical model coversthe temperature field of the complete length of the blank( , from the meniscus inside the mould all the waydown to the cutting torch) with up to one million temperature field of the slab passing through aradial caster with a large radius can be simplified withthe Fourier-Kirchhoff equation, where only thevzcom-ponent of the velocity is = + + cTxkTxykTyzkTz ++r +cvTzQz &source(1)Equation (1) must cover the temperature field of theblank in all three stages: above the liquidus temperature( , the melt), the interval between the liquidus andsolidus temperatures ( , the so-called mushy zone) andbeneath the solidus temperature ( , the solid phase).

7 Itis therefore convenient to introduce the thermodynamicfunction of specific volume enthalpyHv=crT, which isdependent on the temperature and also includes thephase and structural heats (Figure 3).Heat conductivityk, specific heat capacitycanddensityrare thermophysical properties that are also thefunctions of temperature. Equation (1) therefore takesthe following form: HxkTxykTyzkTzvt= + + +vHzzv (2)The unknown enthalpy of the general node of theblank in the following instant (t+Dt) is given by theexplicit formula:HHQZQZ QYQY vijkvijkijijiji,,(),,(),, ,,(t+Dtt=+ +++11jQX QXxyz+++ 1)DtDDD(3)Figure 3indicates how the temperature model for thecalculated enthalpy in equation (3) determines the un-known temperature3. All the thermodynamic propertiesof cast steel, dependent on its chemical composition andthe cooling rate, enter the calculation as the functions is therefore a significantly non-lineartask because, even with the boundary conditions, theirdependence on the SURFACE temperature of the blank isconsidered boundary conditions are, therefore, as level of steel(4a)2.

8 =kTn 0the plane of symmetry(4b)3. = kTnhtc TT ()surfacemouldinside the mould(4c)J. STETINA et al.: MINIMIZATION OF SURFACE DEFECTS BY increasing THE SURFACE TEMPERATURE ..312 Materiali in tehnologije / Materials and technology 47 (2013) 3, 311 316 Figure 2:Diagram of the measurement configuration of the coolingeffects of a nozzleSlika 2:Prikaz konfiguracije meritve hladilnih u~inkov {obeFigure 3:Enthalpy function for steel showing the phase and structuralchangesSlika 3:Krivulja entalpije jekla, ki prikazuje fazne premene in spre-membe v mikrostrukturi4. = + kTnhtc TTTT ()()surfaceambsurface4amb4sewithin the secondary and tertiary zones(4d)5. =kTzq &beneath the rollers(4e)The boundary conditions are divided into the area ofthe mould, the area of the secondary cooling and the areaof the tertiary initial condition for the investigation is thesetting of the temperature in individual points of themesh.}

9 A suitable temperature is the highest possibletemperature, , the pouring temperature. The explicitdifference method is used for solving this problem. Thecharacteristic of this method is that the stability of thecalculation is dependent on the magnitude of the timestep. The model uses a method for adapting the timestep, , the time step entered by an operator is merely arecommendation and the software is modifying itthroughout the HEAT-TRANSFER COEFFICIENT OF THENOZZLEThe cooling by the air-mist water nozzles has themain influence and it is, therefore, necessary to establishthe relevant heat-transfer coefficient of the forced con-vection. Commercially sold models of the temperaturefield describe the heat-transfer coefficient beneath thenozzles as a function of the incident quantity of waterper unit area. They are based on various empiricalrelationships. However, this procedure is model discussed in this paper obtains its heat-trans-fer coefficients from the measurements of the sprayingcharacteristics of all the nozzles used by the caster on theso-called hot plate in an experimental laboratory and fora sufficient range of operational pressures of water and asufficient range of casting speeds of a slab.

10 Thisapproach represents a unique combination of an experi-mental measurement in a laboratory and a numericalmodel for the calculation of the non-linear boundaryconditions beneath the cooling laboratory device enables separate measurementsof individual nozzles. It includes a steel plate mountedwith 18 thermocouples, heated by an external electricsource. The steel plate is heated to the testing tempera-ture, than it is cooled by a cooling nozzle. On the returnmove the nozzle is covered with a deflector, whichenables the movement of the nozzle without cooling thesurface. This device measures the temperatures beneaththe SURFACE of the slab again by means of laboratory device allows the setting of: the nozzle type, the flow of water, the air pressure, the distance between the nozzle and the investigatedsurface, the SURFACE temperature, the shift the cooling nozzle for theminimum water flow appeared to be too intense, themeasurements were made for the smaller (Figure 4).