News

  • Home
  • Technical
  • Automation, Instrumentation and Modelling of Continuous Casting of Steel

Automation, Instrumentation and Modelling of Continuous Casting of Steel


Automation, Instrumentation and Modelling of Continuous Casting of Steel

Continuous casting process for liquid steel is a process whereby the liquid steel is solidified into a semi-finished steel product (billet, bloom, beam- blank, round, or slab) for subsequent rolling in the rolling mills. The basic operation of a continuous casting machine is to convert liquid steel of a given composition into a strand of desired shape and size through a group of operations like primary cooling zone, spray cooling zone, and straightener etc.

The process of continuous casting basically comprised of (i) a tundish, located above the mould, which receive the liquid steel from steel teeming ladle and feed it to the mould at a regulated rate, (ii) a primary cooling zone consisting of water cooled copper mould through which the liquid steel is fed from the tundish for generating a solidified outer steel shell sufficiently strong enough to maintain the strand shape as it passes into the secondary cooling zone, (iii) a secondary cooling zone in association with a containment section positioned below the mould, through which the steel strand (still mostly liquid) passes and is sprayed with water or a mix of water and air (air mist) for further solidifying of the steel strand, (iv) a section for the unbending and straightening of steel strand, (v) a cutting section consisting of cutting torches or mechanical shears for the cutting of the solidified steel strands into desired lengths for removal, and (vi) a run out roller table to cooling beds or directly to a product transfer area.

The continuous casting process of steel is a complex technological process which includes issues related to heat transfer, the solidification process of liquid steel, flow of liquid steel, and phase transitions from liquid to the solid state. This involves considerable difficulty in creating the optimal process control system, which is to include the influence of all the physico‐chemical phenomena which can occur during the continuous casting process of liquid steel. Because of this, the control of the continuous casting process is one of the most difficult tasks in the process of steelmaking.



The continuous casting process due to its complexity is accompanied by several physical phenomena. The liquid steel solidification process within the mould and after leaving the mould in the secondary cooling zone has the most of these important phenomena. In the primary cooling zone, the partial processes which are  taking place are (i) turbulent flow of liquid steel through a complex geometry area  and a submerged entry nozzle or a shroud which is caused by convection, (ii) heat transfer within the liquid steel area, (iii) heat transfer in the mould between the forming shell and the mould wall, (iv) heat flow through the layer of solid and liquid slag, (v) formation of thermal stress, (vi) shrinkage of the solidifying shell related to transitions occurring during the steel solidification process, (vii) thermal effect accompanying the solidification phenomenon, (viii) mechanical impact of the mould walls on the solidifying strand, (ix) the process of an air gap formation between the mould wall and the solidifying strand, and (x) the formation of crystals within the solidification zone accompanied by element segregation effects.

The formation of surface defects takes place in the secondary cooling zone. The processes which are taking place in this zone are (i) heat transfer within the liquid core area (conduction and convection), (ii) heat conduction in the solidified shell layer, (iii) thermal effect accompanying the solidification phenomenon, (iv) multi‐stage heat transfer resulting from the strand cooling by the nozzle system, related to the number of spray zones and the applied cooling type, (v) shrinkage of the solidifying strand, related to transitions occurring during the steel solidification process, (vi) formation of individual solidification zones (zone of dendritic crystals and zone of equi-axed crystals), and (vii) formation of stress related to the contact of rolls with the strand, and the possibility of bulging between the continuous casting machine rolls.

There are several drivers for the automation, instrumentation, and modelling of continuous casting of steel. These drivers include increased customer demands for quality, increased competition, more stringent environmental regulations, and increased safety requirements. Further, the overall production system at the continuous casting machine is to ensure process consistency with previous and subsequent units. Further, the continuous casting process automation system is also required to perform the essential tasks which include production planning and scheduling, quality assurance, and the more conventional supervisory control functions.

Process control of the continuous casting process needs elaborate instrumentations for the full control of the strand solidification process. The measurement system of a continuous casting machine provides a lot of process information. However, key information, such as changes in the shell thickness at individual points of the machine and the metallurgical length (the length of the liquid core), is missing. Hence, the mathematical models are extremely important for the control system of the continuous casting process. The accuracy of these mathematical models allows them to be used for taking technological decisions during the process.

Automation and instrumentation system along with the mathematical models improves and asserts the quality of the continuous cast products and reduces the machine downtime in various ways. Expert systems incorporating the mathematical models have been developed. New challenges which stem from ever increasing quality demands as well as new ideas for the tackling of various well known issues of the continuous casting process have led to several advances in the automation and control of the process.

Modern automation systems for the continuous casting process use several mathematical models to simulate the different phases of the casting process. The input data for these computations is obtained in real time by the specific transport adapter from the Level-1 automation. The target parameter of the casting process is to be specified by the given production program or by the operator. An expert system is used to calculate the optimal values for the process-parameter, to control the production quality, to run the on-fly modelling of the different states of the technological process, and to check the rollers and segments setup. The expert system is part of the Level-2 automation. A clear software-architecture and stable middleware platform for the data-transfer plays the important role for a successful interaction between different automation systems, expert system, and the operators.

The HMI (human machine interface) guides operators through the production process. The operating personnel interaction is limited to quality-related and safety-related activities. An overview of important information is presented on the main display and details can be easily accessed through an extensive set of dedicated screens. Operator screens are displayed in the language and units which are understood by the operating personnel. The entire system comprises a configurable set of applications and the user can select predefined texts instead of inputting them. Fig  shows typical automation system of continuous casting process.

Fig 1 Typical automation system of continuous casting process

Hierarchy of the automation system

The following classification of the levels of responsibility, requirements, and the response of the automated system is defined when designing and developing automation and control systems for the technological processes in a continuous casting machine.

Level 0 automation – It consists of control of individual units. Individual units within an automated area are controlled using connected sensors, transducers, rotary transducers, drives, controls, and control circuits. Direct manual control is performed through the unit, the drives, and interlocks which are handled in turn by the Level 0 automation system. Most safety mechanisms are also stored at this level.

Level 1 automation – It control of groups of units through PLC (programmable logic controller) control. The tasks handled by the Level 1 automation system include the control systems of several devices within one automated area. Control tasks are normally performed by the PLC modules and micro-controllers in real time, with an ensured system response time interval of 20 milliseconds (ms) to 150 ms for PLC control, and between 10 ms and 20 ms for micro-controllers (for example, motion controllers). Due to the strict limitation of the response time for these systems, complex production process models cannot be implemented, for example, the tasks relating to material tracking and coverage planning using mobile units are delegated to other automation levels.

The Level 1 automation functions for the continuous casting machine normally include (i) control of the turret, dummy bar car, and the tundish, (ii) determination of the position of the dummy bar, (iii) adjustment of the drive rollers, (iv) width adjustment, mould taper, and mould level control, and (iv) regulation of the air and water according to the selected set points in the primary and secondary cooling systems

Level 2 automation – Level 2 automation is for the process control. The Level 2 automation system plays a crucial role in determining the efficiency and quality assurance of the production process. The Level 2 automation system manages and monitors the casting process using instructions and settings pre­-defined by the operational engineer or the relevant stan­dards. In addition, each casting instruction includes a set of quality assessment parameters which indicate the best conditions for producing the target product quality. The use of a set of metallurgical models means the casting process can be fully automated, minimizing the need for operator input or interven­tion. The process model connects all faculties to achieve optimize overall performance.

The Level 2 automation system is responsible for the areas such as (i) production quality assurance, (ii) process control and transmission of commands and parameters to the Level 1 automation system, (iii) automated production data acquisition, (iv) simulation and forecasting of the system’s condition using integrated mathematical models of the technological process, (v) material tracking, (vi) optimization of the material handling and coverage planning systems using mobile equipments, and (vii) warning and fault indication system, including evaluation of production fault and setting times.

The Level 2 automation system requirements for the continuous casting process include (i) collection and display of process parameters during casting, (ii) calculation of the 3D temperature distribution in the strand, on the strand surface, and at the edges, (iii) calculation of the strand shell growth, solidification length, edge shrinkage, scale, and other casting characteristics, (iv) dynamic positioning of the strand secondary cooling system, (v) performing of the dynamic adjustments of the segments (soft reduction), (vi) tracking of the material changes and the solidification position, and (vii) accepting and forwarding operator interventions.

The programmes and models which are normally included in the Level 2 automation include (i) liquid steel flow control with stopper mechanism, (ii) mould level control, (iii) mould powder level control, (iv) automatic start of the casting, (v) mould breakout prevention system, (vi) mould, strand, and final stirring, (vii) hydraulic mould oscillation, (viii) heat-tracking model, (ix) real-time quality evaluation, (x) cut optimization model, (xi) real-time strand solidification model, (xii) on-line / off-line solidification curves calculator, (xiii) dynamic mechanical soft reduction, (xiv) dynamic secondary cooling control, (xv) cast product marking machine, (xvi) optical product recognition system, (xvii) process analysis and simulation,  (xviii) metallurgical data management, (xix) production delays detection, (xx) equipment life tracking, and (xxi) cast product handling logistics, including the tertiary cooling.

When implementing Level 2 automation control functions for the continuous casting process, complex data structures are used to model the various technical terms for a domain. Among other things, the mapping of the problem domain using spray plans, air plans, reference temperature curves, casting powders, casting parameter data sets, steel grades, chemical reference analyses, steel grade groups, cracks, and sample cuts is examined from different angles with regard to typical metallurgical problems. Moreover, mathematical models, control mechanisms, and user interfaces require a simplified, standardized description of the continuous casting process in which the properties of the real objects which are relevant for the calculation are clearly defined such as strand guidance, mould, segment, roll, spray nozzle, secondary cooling control circuit, and the cooling segments. The terms mentioned can be easily described by introducing concise, domain-specific languages.

Tasks of Level 1 or Level 2 automation systems in case of continuous casting cannot always be clearly classified. Conclusions as to the placing and distribution of each individual automation task are determined by the localization of input parameters and process data, possible response times of the embedded model, the storage space needed, and by the degree of autonomy. Requirements are frequently distributed between the two systems such as material tracking and some are even duplicated such as safety interlocks. Both systems have their own user interfaces which are normally designed for the respective automation level tasks.

Level 3 – Level 3 automation is for production planning. It deals with the generation of production plans, e.g., the casting program, work scheduling and preparations, or store management, and with maintenance planning, shutdown times and maintenance tasks.

Software architecture of the distributed Level 2 automation system –  The important requirement which is required to be met by the Level 2 automation system is the stable and secure communication of the integrated mathematical models with the Level 1 control system, Level 3 planning system, connected databases, and with the continuous casting machine operating personnel. During the designing and configuring of the Level 2 automation, the variety of different components, their interfaces and connected data sources are important, and the behavioural logic of every single component is not to be underestimated.

The mathematical production process models which are integrated in the Level 2 automation system form the core of the process control system. By monitoring the actual status of the process, the models can be continuously supplied with actual values from the Level 1 automation system. On the other hand, the data are also supplemented with short-term planning results, and with the material and order data from the Level 3 automation system. The calculation results and an overall view of the production process are displayed at the continuous casting machine for the operating personnel on a user interface. Any necessary process control interventions by the operating personnel can be carried out by using both the Level 1 automation system input masks such as speed settings, and the Level 2 automation system user interface such as changing the control regime (reference temperature and spray plan control).

A variety of software architectural patterns are used to implement such systems, and these frequently include models (multi-agent architecture), events and messaging dispatchers (event-driven architecture), and distributed services (service-oriented architecture).

Instrumentation

Instruments have been used on continuous casting machines since the early days of continuous casting. Instruments are being used on every major component of the continuous casting machine between the turret, or ladle car, and the run-out roller table for the cast product. Instruments are used extensively in continuous casting machines for monitoring variables in the ladle, tundish, mould, secondary cooling zone, radiation zone, and run-out roller table, as shown in Fig 2. In fact, the increased use of instruments to measure and control casting parameters has been credited as one of the major contributors to the large gains achieved in the continuous casting machine productivity and quality.

Fig 2 Parameters measured on continuous casting machines

Instruments are crucial to any control and automation system and their contribution to achieving modern productivity and quality standards cannot be overestimated. Instruments are the ‘eyes’ of the control and automation system, and with present technology permanent instrumentation is available for the process and quality control system to ‘see’ the most important ladle, tundish, and mould variables.

The main functions of instruments in the continuous casting process are to (i) measure the parameters which are utilized for controlling the performance of the mechanical and metallurgical functions of the continuous casting, (ii) assign a quality rating for each cast section, (iii) diagnose operating and machine problems, (iv) develop knowledge which correlates product quality and productivity to the casting machine design and operation.

The number and sophistication of instruments used on the continuous casting machines has been growing rapidly. The main reasons for the rapid growth are the ever-increasing demands for higher productivity and as cast product quality, and the availability of the modern on-line digital computer. This is particularly true for slab casting machines, where quality and productivity demands are the most stringent. Earlier, the emphasis was placed on the mould instrumentation since the mould practices and parameters have the most impact on product quality and productivity. However, lately, significant progress has been made in developing and applying instrumentation on the ladle, tundish, containment, and casting machine run out roller table.

Given the hazardous environment found within the spray chamber of the secondary cooling zone of the continuous casting machine, it is not surprising that casting machine control system is frequently ‘blind’ to changes in critical process variables, such as the strand surface temperature, in this zone. Instrumentation used here is normally of a temporary nature and is hence employed on an experimental basis. Other important instrumentations include those used to measure inter-roll bulging, thickness of the solidified shell, and mould / strand friction.

Modelling of continuous casting process

Modelling the process of steel continuous casting is a very complex task, and can be accomplished using various types of mathematical models.  At present, it is not possible to simultaneously capture all of the effects occurring during the complete process of continuous casting, and to present them in the form of a single comprehensive numerical model. The natural division applied in the continuous casting process modelling is related to an attempt at identifying the ensuing problem during actual casting of liquid steel, or focusing on a selected section of the process in order to improve the existing technology.

At the early stage of problem solving, the correct selection of the model type and related possibilities for its adaptation to the class of the solved problem has been a difficult challenge. Theoretically, a more complex model (which is to say more ‘intelligent’) can easily answer questions about the primary technological parameters of the casting process. Yet in practice a number of limitations are encountered. Assuming hypothetically that a complex model has been verified as correct, in the best case an unnecessary extension of the computing time is needed. It results from the fact that the model calculates much more parameters than is needed to solve the defined problem. The second danger caused by non‐synchronising the complexity of the set problem with the ‘intelligence’ of the used tool is the issue of the verification of model parameters and their correlation with the process data. The more theoretical is the elaboration of the model, the more parameters and the higher risk of occurrence of unmeasurable parameters is there. The last comment concerns the problem of the strategy of acquiring knowledge of the value of the required model parameters. Several years of experience in modelling the continuous casting process show that the best choice is an experimental measurement of all measurable model parameters. It can be illustrated by parameters in the form of the specific heat of the steel cast as a function of temperature, heat conductivity for the steel, and viscosity etc.

Physical modelling – Physical modelling of the continuous casting process such as using water to simulate liquid steel enables considerable insight into the flow behaviour of the liquid steel during the continuous casting process. Previous understanding of fluid flow in the continuous casting process has come about mainly through experiments using physical water models. This technique is a useful way to test and understand the effects of new configurations before implementing them in the process. A full scale model has the important additional benefit of providing operator training and understanding.

Construction of a physical model is based on satisfying certain similitude criteria between the model and actual process by matching both the geometry and the force balances which govern the important phenomena of interest. For reproducing the liquid steel flow pattern with a water model, all of the ratios between the dominant forces are to be the same in both the systems. This ensures that velocity ratios between the model and the steel process are the same at every location. The size of a dimension-less group indicates the relative importance of two forces. Very small or very large groups can be ignored, but all dimensionless groups of intermediate size in the casting process are required to be matched in the physical model. An appropriate geometry scale and fluid are to be chosen to achieve these matches.

It is fortunate that water and steel have very similar kinematic viscosities. Hence, Reynolds and Froude numbers can be matched simultaneously by constructing a full-scale water model. Satisfying these two criteria is sufficient to achieve reasonable accuracy in modelling isothermal single-phase flow systems, such as the continuous casting nozzle and mould and this has been done with great success.

A full-scale model has the extra benefit of easy testing of machine components and operator training. Actually, a water model of any geometric scale produces reasonable results for most of the flow systems, so long as the velocities in both systems are high enough to produce fully turbulent flow and very high Reynolds numbers. Because flow through the tundish and mould nozzles are gravity driven, the Froude number is normally satisfied in any water model of these systems where the hydraulic heads and geometries are all scaled by the same amount.

Physical models sometimes are to satisfy heat similitude criteria. In physical flow models of steady flow in ladles and tundishes, for example, thermal buoyancy is large relative to the dominant inertial- driven flow, as indicated by the size of the modified Froude number, which therefore is to be kept the same in the model as in the liquid steel system. In ladles, where velocities are difficult to estimate, it is convenient to examine the square of the Reynolds number divided by the modified Froude number, which is called the Grashof number. Inertia is dominant in the mould, so thermal buoyancy can be ignored there. The relative magnitude of the thermal buoyancy forces can be matched in a full-scale hot water model. This is not easy, however, as the phenomena which govern heat losses depend on properties such as the fluid conductivity and specific heat and the vessel wall conductivity, which are different in the model and the steel vessel. In other systems, such as those involving low velocities, transients or solidification, simultaneously satisfying the several other similitude criteria important for heat transfer is virtually impossible.

The complexity of the continuous casting process and the phenomena which govern it as indicated in Fig 3 make it difficult to have physical model. However, with the increasing power of computer hardware and software, mathematical modelling has become an important tool for controlling all the aspects of the continuous casting process.

Fig 3 Schematics of the continuous casting and the phenomena in the mould region

Computational or mathematical modelling – In the present time, decreasing computational costs and the increasing power of commercial modelling packages have make it easier to apply mathematical models as an additional tool to understand complex materials process steps of the continuous casting process of liquid steel. Computational models have the advantage of easy extension to other phenomena such as heat transfer, particle motion and two phase flow, which is difficult with isothermal water models. The computational models are also capable of more faithful representation of the flow conditions experienced by the liquid steel. For example, there is no need for the physical bottom which interferes with the flow exiting a strand water model, and the presence of the moving solidifying shell can be taken into account.

Mathematical models can now simulate most of the phenomena important to continuous casting process. These include (i) fully-turbulent, transient fluid motion in a complex geometry (inlet nozzle and strand liquid pool), affected by argon gas bubbles, thermal and solutal buoyancies, (ii) thermodynamic reactions within and between the powder and steel phases, (iii) flow and heat transport within the liquid and solid flux layers, which float on the top surface of the steel, (iv) dynamic motion of the free liquid surfaces and interfaces, including the effects of surface tension, oscillation and gravity-induced waves, and flow in several phases, (v) transport of superheat through the turbulent liquid steel, (vi) transport of solute (including intermixing during a grade change), (vii) transport of complex-geometry inclusions through the liquid, including the effects of buoyancy, turbulent interactions, and possible entrapment of the inclusions on nozzle walls, gas bubbles, solidifying steel walls, and the top surface, (viii) thermal, fluid, and mechanical interactions in the meniscus region between the solidifying meniscus, solid slag rim, infiltrating liquid flux, liquid steel, powder layers, and inclusion particles, (ix) heat transport through the solidifying steel shell, the interface between shell and mould (which contains powder layers and growing air gaps), and the copper mould, (x) mass transport of powder down the gap between shell and mould, (xi) distortion and wear of the mould walls and support rolls, (xii) nucleation of solid crystals, both in the melt and against mould walls, (xiii) solidification of the steel shell, including the growth of dendrites, grains and microstructures, phase transformations, precipitate formation, and micro-segregation, (xiv) shrinkage of the solidifying steel shell due to thermal contraction, phase transformations and internal stresses, (xv) stress generation within the solidifying steel shell due to external forces (mould friction, bulging between the support rolls, withdrawal, and gravity), (xvi) thermal strains, creep, and plasticity (which varies with temperature, grade and cooling rate), (xvii) crack formation, and (xviii) coupled segregation, on both microscopic and macroscopic scales.

The staggering complexity of the continuous casting process makes it impossible to model all of these phenomena together at once. Hence, it is necessary to make reasonable assumptions and to uncouple or neglect the less-important phenomena. Quantitative modelling needs incorporation of all of the phenomena which affect the specific issue of interest. Hence each model needs a specific purpose. Once the governing equations have been chosen, they are normally discretized and solved using finite-difference or finite-element methods. It is important that adequate numerical validation be conducted.

Numerical errors normally arise from too coarse a computational domain or incomplete convergence when solving the nonlinear equations. Solving a known test problem and conducting mesh refinement studies to achieve grid independent solutions are important ways to help validate the model. Finally, a model is required to be checked against experimental measurements on both the laboratory and plant scales before it can be trusted to make quantitative predictions of the real process for a parametric study.

The final test of a model is if the results can be implemented and improvements can be achieved, such as the avoidance of defects in the steel product. Plant trials are ultimately needed for this implementation. Trials are to be conducted on the basis of insights supplied from all available sources, including physical models, mathematical models, literature and previous experience. As increasing computational power continues to advance the capabilities of numerical simulation tools, the modelling plays an increasing role in future advances to high-technology continuous casting process. Modelling can augment traditional research methods in generating and quantifying the understanding needed to improve any aspect of the process. Areas where advanced computational modelling plays a crucial role in future improvements include (i) transient flow simulation, (ii) mould flux behaviour, (iii) taper design, (iv) on-line quality prediction and control, especially for new problems and processes such as high-speed billet casting, thin slab casting, and strip casting.

Future advances in the continuous casting process are not going to come from models, experiments, or plant trials. They are going to come from ideas generated by people who understand the process and the problems. This understanding is rooted in knowledge, which can be confirmed, deepened, and quantified by tools which include computational models. As the computational tools continue to improve, their importance is increasing in fulfilling this important role, leading to future process advances.

The assumed computing objective and the required accuracy are to be the key in selecting the model. In several cases, the desired information is knowledge of the metallurgical length of the strand and the dynamics of changes in the shell thickness. This is the case when determining a place for carrying out the so‐called soft reduction operation. As can happen when the strand casting speed needs to be changed, a procedure allowing a new cooling intensity to be determined is needed. A problem like this does not need answering a series of questions related to stress occurring in the strand, the structure formed, or potential cast strand defects. Thus, it is understandable that the model is naturally simplified to a form, which still provides a credible answer to the questions which needs solution.

Some of the important models used in the automation and control of the continuous casting process are given below. The models are normally incorporated at the Level 2 automation level.

Dynamic secondary cooling control – There is a third dimension in the dynamic secondary cooling control. The model set up takes the precision and control possibilities to the next dimension allowing completely new philosophies for secondary cooling and soft reduction. When setting up the secondary cooling system, it is prerequisite to consider all known parameters which have a known influence to the calculation of 3-dimensional temperature profile of the strand. All different nozzle types are measured at the nozzle test stand to evaluate the spray water distribution. This derived information is input to the maintenance and setup system (MSS) of the 3D model. The visualized spray distribution can be seen in the maintenance system.

The exact positions of the nozzles in the cooling zones are entered and the spray distribution of one zone can be seen in the MSS. The heat removal of a cooling zone is calculated considering the heat removal of the spray water, rolls, and heat radiation. The MSS allows all cooling-relevant settings to be configured in such a way that the spray-water distribution in the cooling zones and the application of cooling practices are optimized for continuous casting machines. Metallurgical know-how can be easily incorporated into the automation setup. A built-in off-line simulation system enables comprehensive testing of new parameter settings prior to application in the production process.

The Level 2 automation system for secondary cooling provides a mathematical model for calculating the temperatures on the strand surface and inside the strand as a function of the spray plan to be used, the interpolation points in the reference temperature curve, or in relation to time changes for the spray water quantities across the complete machine. The dynamic secondary cooling control system can handle three control regimes namely (i) temperature control, (ii) strand age, and (iii) spray plan control.

In the temperature control regime the dynamic secondary cooling control system calculates the volumes of water for strand secondary cooling which are needed to maintain the specified reference temperature on the strand surface. The strand age regime is one way of controlling the secondary cooling process, taking into account the change in parameters over a given period of time. With the spray plan control regime a spray plan is produced for secondary cooling whereby preset spray water volumes correspond to a specific casting speed. As the change in casting speed takes effect, the spray water volumes are also immediately modified, and the resulting temperatures on the strand surface are displayed to the continuous casting machine operator. If temperature scanners are used in the plant, the dynamic secondary cooling control system is able to adapt to the values delivered by the scanners, so that the coefficients of the strand temperature field calculation can be adapted to the values measured.

Dynamic 3D secondary cooling system – The first-generation dynamic solution was characterized by a two-dimensional temperature calculation of the strand centre. The strand corners were largely neglected by the process model. Continuous improvements in computer performance have now made it possible to calculate the temperature at any point within the entire strand in real time, in a full three-dimensional mode and in a sufficiently fine discretization yielding very detailed temperature profiles as can be seen for strand surface and strand centre.

The model is based on an explicit finite-volume approximation which solves the heat transfer equation and takes into consideration temperature-dependent material properties such as density as well as the position-specific cast product thickness and width. Dynamic 3D secondary cooling system accurately assesses the heat transfer from the cast product surface resulting from radiation, heat transfer to the rolls, natural convection and spray water. Further, the dynamic 3D secondary cooling system can be applied for both spray cooling and air-mist cooling and takes into account the spray distribution pattern of the nozzles and the actual spray water temperature. This ensures an accurate spray-cooling heat transfer prediction to temperatures below 700 deg C when the Leidenfrost phenomenon disappears. The result is an even more precise determination of the strand surface-temperature profile and the final point of strand solidification.

Based on the precise temperature calculations the dynamic 3D secondary cooling system allows specifying the desired surface temperature not only along the strand length, but also across the strand width. Even individual control of the water flow and positioning of each cooling nozzle is possible. The control algorithms of the dynamic 3D secondary cooling system calculate the water flow set-points to achieve the target strand surface temperature values. Pyrometer measurement results show an excellent fit in between calculated and measured lateral strand temperature profile.

Application of the dynamic 3D secondary cooling system allows introducing completely new philosophies to set up cooling practices for upcoming challenges in continuous casting. The combination with moveable spray nozzles (3D sprays) yields unprecedented quality results.

The advanced secondary cooling dynamic 3D model derives correct water flow rates even in transient casting situations such as steel grade changes, casting speed variations, different tundish temperatures, tundish exchanges, and at the beginning and end of a casting sequence. The water flow rate for each cooling zone is calculated to maintain a defined surface temperature profile throughout the entire casting sequence. The maintenance system allows the process engineer to change cooling practices easily and introduce continuous casting shop specific cooling expertise. The off-line simulation system is used to test the effect of the new settings in various casting situations before utilization in the production process.

Dynamic phase calculation of material properties – In order to calculate a 3-dimensional temperature profile of the strand, material properties like enthalpy, solid fraction, density, and conductivity as a function of the temperature are to be known. In case, these properties are experimentally known for a given steel grade composition, these functions can be entered by the process engineers in the MSS, which is very time-consuming. Normally the process engineer does not know these thermo-physical properties. The software model calculates all the thermo-physical data used by 3D model. Dynamic phase calculation of material properties is available as an on-line tool to determine the material properties for the current steel grade analysis.

The traditional approach is to define the thermo-physical properties for grade groups with a pre-defined concentration range of the chemical analysis. Using the dynamic phase calculation, these data can be calculated for each individual steel grade. This makes the prediction of quantities such as the point of complete solidification on the strand and the temperature distribution of the strand during casting more accurate and hence allows for precise metallurgical treatments which can lead to an improved quality of the products. Further, the model indicates whether the current analysis of the steel is peritectic or not and alerts the operator in the event of an unexpected peritectic grade. This can reduce the risk of breakouts and improves quality.

The dynamic phase calculation is based on thermodynamical models. The liquid-solid phase transformation in the high temperature range is described by a Gibbs free energy model in combination with a micro-segregation model. For solid-solid phase transformations in the low temperature range an Avrami type model is employed. The free parameters of the models are determined with the help of experimentally measured quantities. Using off-line simulations of dynamic phase calculation together with the 3D model allows metallurgical development of new steel grades.

Traditionally the steel grades are grouped and a typical chemical analysis for the group is used to determine the material properties. With dynamic phase calculation, the material properties are derived from the actual steel analysis. Calculations can show that there can be a difference in the point of final solidification of half a meter or even more by comparing the results of the actual steel analysis versus the grade group analysis. This fact shows the importance of having an on-line calculation of the actual steel grade in order to improve the quality of the cast products.

Dynamic gap soft reduction – Dynamic gap soft reduction stands for dynamic roll-gap adjustment in the continuous casting process. This is made possible by specially designed strand-guide segments – known as ‘smart segments’ in which the roller gaps can be remotely adjusted for strand thickness changes and for improved internal strand quality. On the basis of the on-line information provided by the dynamic 3D thermal-tracking model, the dynamic gap soft reduction dynamically calculates the set points of the adjustable roll gap.

Supervision of the roll engagement, depending on the state of solidification (liquid, mushy, or solid) and the calculated strand-thickness profile, is a decisive factor for precise roll adjustments and thus improved product quality. An optimized roll engagement also reduces excessive forces on the strand and decreases roller wear. The more accurate control of the roller gaps allows additional casting strategies to be implemented such as liquid-core reduction and intentional bulging soft reduction. I.e. intentional dynamic gap increase before the soft reduction area allows for higher thickness reduction in this area. This further improves the casting flexibility and the product quality.

Dynamic gap soft reduction makes it possible to freely define scenarios for start-up, tundish change, and tail out strategies based on the strand thickness, steel grade, and casting status. In this way roll damage and production interruptions, which can arise from the different casting behaviour of the cold strand head or end, can be avoided.

Nozzle expert for early clogged nozzle detection – Cooling water is sprayed through nozzles onto the strand with the objective of achieving uniform cooling of the steel. However, if one or more of these nozzles are clogged, then a section of the strand cannot be uniformly cooled to the required temperature. This can lead to surface defects, and the cast product possibly has to be down-graded. The issue of changing segments in the continuous casting machine is also a source of difficulty. Hoses can easily be ruptured or jammed. Aware of the consequences of leakages or clogged nozzles, maintenance personnel spend a large number of working hours checking whether nozzles are operating properly.

The nozzle expert helps to detect clogged nozzles and broken hoses in the continuous casting machines and thus ensures that the strand is evenly cooled for high quality steel production. It automatically monitors the condition of the nozzles during the casting process. The model can also be manually activated during casting breaks. The advantage is that nozzle status can be checked following maintenance work or segment changes and immediately repaired before the casting process is restarted.

The model calculation considers parameters like nozzle type (measuring results from the nozzle test stand), height between pressure measuring device, water pressure, pipe lengths, pipe diameters, and nozzle positions. Any modifications to the secondary cooling system e.g. use of different nozzle types needs a change in the set-up of the nozzle expert in order to get correct computational results.

The nozzle expert is based on statistical models and indicates the clogging ratio in each zone (e.g., zone 2 nozzles clogged 10 % with a probability of 96 %). Operators need only to inspect zones for which an alarm is generated. Calculations begin automatically with the start cast signal, and the condition of the nozzles is monitored throughout the casting process. Several alarms help to detect leakages, clogged nozzles, and even falsely installed nozzles on a segment.

Inter-mix expert – It improves yield by prior simulations. During sequence casting, a mixing of steel grades takes place in the tundish and therefore in the strand with each ladle change. On the basis of the chemical composition of the steel, the inter-mix expert calculates whether the mixed steel zones can be used for the foreseen product application or if the steel has to be downgraded or even scrapped. Information acquired from tundish flow experiments combined with analysis results of steel samples taken from solidified products ensures a high degree of accuracy of Inter-mix predictions with respect to the actual composition of the mixed steel zones.

The inter-mix expert determines traces of the previously cast heat present in the current heat. Steel mixing takes place not only in the tundish but also in the mould and upper parts of the strand. Mixing in these areas is evaluated by a mix-box-type sub-model of inter-mix which makes it possible to calculate the chemical composition of the steel at any position along the cast strand.

Tundish changes or the use of separator plates are treated individually. Inter-mix calculations are cyclically performed for selected chemical elements starting with the ‘ladle open event’ of a new heat. The final decision about the compatibility of heats cast in sequence is performed by the heat-assignment function of inter-mix. The concentration profiles of certain critical elements which have an impact on the final product disposition (prime, downgraded, or outright rejection) are determined. A deviation is detected if one of the critical elements does not match the steel-grade specification.

The full benefit from the inter-mix model is achieved by combining the model output with the yield expert model which assures maximum prime quality yield by applying cut-length optimization to incompatible steel areas along the strand which are designated as scrap.

Process engineers work with the powerful simulation environment, which makes it possible to simulate any combinations of different steel grades. Input parameters like analysis, tundish weight, and dimensions of the strand can be easily entered and modified and the computed results are made visible in the HMI. Graphs are displayed for single analysis elements or combinations of more elements. Valuable information like volume concentration, mixed steel length, scrap length, and heat ranges on the strand are shown on the bottom of the graphs.

Configuration of the model can be easily done in the MSS. The process engineer can choose which chemical elements are to be used to determine the inter-mix for any grade link. A powerful simulation environment allows simulating the mixing behaviour of two different grades and the computed volume concentration, calculated analysis along the strand, and heat ranges including possible scrap sections are displayed.

Speed expert – The speed expert model is for the optimum casting speed in any casting situation. Selection of a proper casting speed on the continuous casting machine is of high importance. Several aspects (e.g. quality, safety, machine limits, and production requirements) influence the choice of the casting speed. These different aspects are frequently contradictory such as increase in production calls for a high casting speed whereas the safety requirements limit the casting speed. Normally, several continuous casting shops have self-made software solutions to calculate the casting speed considering different aspects. The aim of the speed expert is to cover most aspects and to provide an easy maintenance tool which enables the process engineer to adjust the behaviour of the speed expert to the special needs.

The speed expert calculates cyclically the optimum casting speed. The calculation of the casting speed is based on various rules, which consider the different aspects and are specified in the speed expert practice. Each rule determines a speed range which satisfies its requirements. The speed expert first determines the inter-section of all these speed ranges. If the inter-section is not empty, then it selects a casting speed depending on the predefined strategy, which can be maximum speed, aim speed, or avoid speed changes as long as actual casting speed is in the valid range. If there is a conflict between the different rules, then the inter-section is empty. In this case the pre-selected conflict resolve strategy is applied which can be (i) priority (lower priority rules are to be neglected till a solution can be found), or (ii) minimum of maximum speed (the smallest of all maximum speeds is to be selected).

On the on-line HMI the casting machine operator can view the speed ranges of all rules and the derived optimum casting speed. The casting operator can change the priorities of the different rules, the strategy, and the conflict resolve strategy to fine adjust the calculation if necessary. Speed set-points are sent to Level 1 and can be executed automatically.

MSS is used to define the speed expert practice considering (i) quality related rules consisting of quality expert rule, minimum / aim / maximum speed for the steel grade, superheat, Mn/S ratio, low tundish weight, and optimum soft reduction, (ii) production related rules consisting of heat pacing, start cast, and clearing, and (iii) safety related rules consisting of machine protection, and forecast calculation for minimum and maximum speed.

Optimum soft reduction can be achieved if the final point of solidification is at the end of a strand segment. A pre-calculation assuming steady state conditions determines the required casting speeds for each strand segment.

Yield expert – The aim of the yield expert is to minimize scrap and to optimize the yield. It considers scrap portions, quality defects, weight restrictions, sample cuts, and width changes while producing the maximum number of scheduled products. The important features of the yield expert are (i) optimization of product length or product weight in the case of scrap sections or quality-related defects, (ii) scheduling of mould width adjustments, (iii) scrap section allocation algorithms, (iv) optimization steps can be switched on-line and off-line, and (v) replay of cut-to-length optimization steps, even in actual production situations

Quality expert – Quality expert determines the definitions necessary for the quality-related process parameters, tracks the actual data during production, predicts the quality of the cast products, and automatically determines the subsequent product disposition. It supports the continuous casting machine operators by on-line quality alerts and a preview of the quality of the cast strands in the casting machine. Quality expert is of two distinct types distinguished by basic or comprehensive product quality rating capability.

All the tracked information and calculation results can be transferred from the production module of the quality expert to the so-called discovery system. This system is dedicated to the long-term archiving and evaluation of the huge amount of information tracked.


Leave a Comment