Method and plant for producing a continuously cast metal product

Abstract

The invention relates to a method for producing a continuously cast metal product, the method involving: producing and providing molten metal in a sub-process; casting a metal strand in a subsequent sub-process; and subjecting the cast metal strand to a further process in a further subsequent sub-process, the sub-processes being computed and / or closed-loop or open-loop controlled by respective sub-process models. In order to carry out such a method with improved product quality and production output, it is provided that the sub-process models are connected to a superordinate overall process optimization system, wherein the overall process optimization system is connected to the sub-process models and checks whether an optimized target variable for a production parameter can be used for the production; if so, the production of the continuously cast metal product is initiated with the optimized target variable. The invention further relates to a plant for producing a continuously cast metal product.

Description

[0001] Process and plant for the production of a continuously cast metallic product

[0002] The invention relates to a method for producing a continuously cast metallic product, in which, in one sub-process, molten metal is produced and provided, in a subsequent sub-process, a metal strand is cast, and in a further sub-process, the cast metal strand is subjected to a further process, wherein the sub-processes are calculated and / or controlled by respective sub-process models. The invention further relates to a plant for producing a continuously cast metallic product.

[0003] Steel production in a suitable plant (e.g., a Nexus or CSP plant), as well as hot processing in a slab or billet mill, consists of several process steps. Separate process models are provided for each individual process step, which handle the calculation, control, and / or regulation of the respective subprocess. The calculation results and measurement results are passed from one subprocess to the next and taken into account there. The overall process can comprise the following subprocesses:

[0004] a converter process, calculated and controlled or regulated by a converter model,

[0005] a pan process, calculated and controlled or regulated by a pan model,

[0006] a distribution process, calculated and controlled or regulated by a distribution model,

[0007] Page 1 a casting process, calculated and controlled or regulated by a casting model,

[0008] a furnace process, calculated and controlled or regulated by a furnace model,

[0009] a pre-road process, calculated and controlled or regulated by a pre-road model,

[0010] a pre-band cooling process, calculated and controlled or regulated by a pre-band cooling model,

[0011] a hot rolling mill process, calculated and controlled or regulated by a hot rolling mill model,

[0012] a cooling process, calculated and controlled or regulated by a cooling model,

[0013] a reeling process, calculated and controlled or regulated by a reeling model,

[0014] a cold rolling mill process, calculated and controlled or regulated by a cold rolling mill model,

[0015] a galvanizing process, calculated and controlled or regulated by a galvanizing model.

[0016] Each individual subprocess model controls or regulates its corresponding subprocess according to separate specifications and may only have knowledge of the preceding subprocess. No overarching optimization of the entire manufacturing process takes place.

[0017] Page 2 WO 2022 / 223297 A1 describes a process optimizer that provides only a connection between the casting machine and the rolling mill. This is intended to achieve an optimal strand thickness to maximize overall production.

[0018] It is known from EP 2346625 B2 to optimize the residual melt in the distributor.

[0019] The invention is based on the objective of designing a generic process and a corresponding system in such a way that it becomes possible to achieve an overall improvement in the product quality of the metallic product. Furthermore, the production quantity should be optimally adjusted. Finally, production costs should also be minimized.

[0020] The solution to this problem by the invention is characterized by the fact that, in the procedure described above, it is provided that the sub-process models are connected to a higher-level overall process optimization system, wherein

[0021] a) the overall process optimization system sends a request to a subprocess model (requested subprocess model) in order to obtain from it an optimized target value for a production parameter for the subprocess controlled or regulated by this subprocess model,

[0022] b) the requested subprocess model determines a possible optimized target variable for the subprocess it monitors and reports this back to the overall process optimization system,

[0023] c) the overall process optimization system, with the reported optimized target variable, directs the inquiry to at least one first subprocess model upstream or downstream of the requested subprocess model, as to whether the operation of the subprocess controlled or regulated by this subprocess model is possible with the optimized target variable,

[0024] Page 3 and the overall process optimization system with the reported optimized target variable to at least one second subprocess model upstream or downstream of the requested subprocess model, the request is made as to whether the operation of the subprocess controlled or regulated by this subprocess model is possible with the optimized target variable,

[0025] d) In the event that the requests according to step c) are answered positively by the first and the second upstream or downstream sub-process model: Initiate the production of the continuously cast metallic product with the optimized target size.

[0026] The first sub-process model is preferably a sub-process model downstream of the requested sub-process model, the second sub-process model is preferably a sub-process model upstream of the requested sub-process model.

[0027] The initiation of production according to step d) above preferably comprises the following steps:

[0028] Confirmation of the optimized target variable for the requested sub-process model,

[0029] Modification of the sub-process based on the optimized target variable and confirmation of this to the overall process optimization system,

[0030] The overall process optimization system specifies the optimized target variable for the first and second sub-process models.

[0031] Each sub-process model preferably includes:

[0032] an online model that calculates and / or controls or regulates the subprocess,

[0033] Page 4, an offline model for calculating at least one production parameter and

[0034] an optimizer that determines optimal values ​​for production parameters calculated by the offline model.

[0035] In this case, it is preferably provided that the offline model computes the subprocess at a speed that is at least 10 times that of real time, preferably at least 50 times that of real time.

[0036] A material model is preferably assigned to the sub-process model, or preferably includes such a model. It is preferably provided that a single material model is assigned to all sub-process models.

[0037] The requested sub-process model according to step a) above is preferably a continuous casting model for continuous casting.

[0038] However, the upstream sub-process model is preferably a converter model for the converter process and / or a ladle model for the provision of melt and / or a distributor model for the distribution of the melt.

[0039] The downstream sub-process model is preferably a furnace model for operating a furnace. It can also be a rolling mill model for rolling the metallic product, in particular a roughing mill model for rolling in a roughing mill and / or a hot rolling mill model for hot rolling the metallic product and / or a cold rolling mill model for cold rolling the metallic product.

[0040] Furthermore, the downstream sub-process model can also be a model for a pre-band cooler and / or a cooling model and / or a coiling model. Finally, it is also possible that the downstream sub-process model is a model for a galvanizing plant.

[0041] Page 5. By using the overall process optimization system, it is also possible to better address customer requirements for the metallic product and, in particular, the desired metal strip. Therefore, a further development of the invention provides that the query of the overall process optimization system to a sub-process model according to step a) above is made depending on a requirement for the metallic product to be manufactured, in particular for the metal strip, wherein the requirement relates in particular to the geometry of the metal strip and / or the quality of the metal strip.

[0042] The proposed method can also be carried out using copper-containing scrap, in which case it is preferably provided that the overall process optimization system optimizes the alloy to be produced and / or the cooling, in particular the intensive cooling, of the metal strip to be produced depending on the strip thickness to be produced.

[0043] The plant for the production of a continuously cast metallic product, preferably for carrying out the process described above, comprises a sub-process in which metallic molten metal is produced and supplied, a subsequent sub-process in which a metal strand is cast, and a further sub-process in which the cast metal strand is subjected to a further process, wherein the sub-processes have respective sub-process models designed for the calculation and / or control or regulation of the sub-process, and wherein a higher-level overall process optimization system is arranged which is connected to the sub-process models, is characterized in that

[0044] a) the overall process optimization system is designed to send a query to a subprocess model (queried subprocess model) in order to obtain from it an optimized target value for a production parameter for the subprocess controlled or regulated by this subprocess model,

[0045] Page 6b) wherein the requested subprocess model is designed to determine a possible optimized target variable for the subprocess it monitors and to report this back to the overall process optimization system,

[0046] c) wherein the overall process optimization system is configured to send the feedback of the optimized target variable to at least one first subprocess model upstream or downstream of the requested subprocess model, asking whether the operation of the subprocess controlled or regulated by this subprocess model is possible with the optimized target variable, and wherein the overall process optimization system is configured to send the feedback of the optimized target variable to at least one second subprocess model upstream or downstream of the requested subprocess model, asking whether the operation of the subprocess controlled or regulated by this subprocess model is possible with the optimized target variable.

[0047] d) wherein the overall process optimization system is designed, in the event that the queries according to step c) are answered positively by the upstream and downstream sub-process model, to cause the production of the continuously cast metallic product with the optimized target size to take place.

[0048] Each sub-process model preferably includes:

[0049] an online model that is trained to calculate and / or control or regulate the subprocess,

[0050] an offline model that is trained to calculate at least one production parameter and

[0051] Page 7 an optimizer who is trained to determine optimal values ​​for the production parameters calculated by the offline model.

[0052] A material model is preferably assigned to the sub-process model or it includes such a model. However, preferably only a single material model is assigned to all sub-process models.

[0053] The requested sub-process model is preferably a continuous casting model for continuous casting.

[0054] The upstream sub-process model is preferably a converter model for the converter process and / or a ladle model for the provision of melt and / or a distributor model for the distribution of the melt.

[0055] The downstream sub-process model is preferably a furnace model for operating a furnace, a rolling model for rolling the metallic product, in particular a roughing mill model for rolling in a roughing mill and / or a hot rolling mill model for hot rolling the metallic product and / or a cold rolling mill model for cold rolling the metallic product, a model for a strip cooler, a cooling model, a coiling model and / or a model for a galvanizing plant.

[0056] The invention also includes a computer program for carrying out the method described above.

[0057] The proposed concept therefore envisages that a higher-level process chain optimizer monitors, controls and optimizes the individual sub-processes in order to optimize product quality, production quantity and product costs.

[0058] Each subprocess has a separate model for calculating, regulating, and / or controlling the respective current process (online model). This is associated with a

[0059] Page 8: Material Model. This material model is preferably used for all sub-processes. In addition, each sub-process has an offline model. This allows calculations to determine, for example, whether a property desired by the overall optimizer (temperature, speed, strand geometry, or microstructure) can be achieved with the otherwise currently available process parameters.

[0060] Furthermore, each subprocess has its own optimizer, which tells the overall optimizer how far, for example, the process speed can be changed.

[0061] This results in several advantages, which arise from a comprehensive consideration of the process chain, particularly for the production of hot-rolled wide strip or cold-rolled strip:

[0062] Central segregation in the sheet metal of boron-alloyed steels can be reduced by casting near the liquidus.

[0063] In ferritic steels, tensile grooving on the cold-rolled strip surface can be avoided by a high globulitic content in the primary microstructure of the slabs. The globulitic microstructure content can be influenced by lower superheating during casting, which can be adjusted at the steelworks. The globulitic microstructure content can be increased by using agitators in the strand guide; conversely, an insufficient globulitic microstructure content can be increased by using an agitator, even at higher superheating levels.

[0064] It is possible to increase the casting and rolling speed. A higher casting and rolling speed allows for a reduction in the ordering temperature of the melts in the steelworks. Excessively high melt temperatures, due to the risk of strand shell perforations, necessitate a reduction in the casting speed.

[0065] Page 9 Accordingly, the proposed overarching optimization offers significant advantages.

[0066] Unlike previously known solutions, the proposed approach allows for a connection to all involved subprocesses, thus ensuring an overall optimized production. Action is not only taken when a predetermined actual value deviates from the target value; rather, an optimized new production plan can be created and implemented by querying several, preferably all, other subprocesses and incorporating the results of their responses (difference from the aforementioned EP 2 346 625 B2).

[0067] In addition, the expected properties of the slabs are passed on to the subsequent sub-processes (units) in advance. This allows, for example, the expected furnace inlet temperature to be used for the early control of the furnace temperatures.

[0068] The drawing shows exemplary embodiments of the invention. It shows:

[0069] Fig. 1 schematically shows a process chain for the production of a steel strip,

[0070] Fig. 2 schematically shows the process chain, with details given for one sub-process.

[0071] Fig. 3 schematically shows the flowchart for processing a request by a process chain optimizer,

[0072] Fig. 4a schematically shows the procedure in a current production of a metal strip,

[0073] Fig. 4b schematically shows the procedure for optimizing the final product,

[0074] Page 10, Fig. 5a shows an optimization process concerning the material composition of the metal strip to be produced, where the copper equivalent is plotted against the strip thickness in the diagram shown.

[0075] Fig. 5b shows, in the representation according to Figure 5a, the influence of cooling on the permissible copper equivalent, and

[0076] Fig. 6 schematically shows a flowchart for the optimization of the process in the casting machine.

[0077] Figure 1 illustrates the process chain for the production of a stall strip in a casting and rolling mill. Liquid steel is first supplied and distributed in a ladle using a ladle model and a distributor model. The liquid steel is then poured into a continuous casting plant, a process monitored by a casting model.

[0078] The cast slab is then heated in a first furnace and rolled in a pre-furnace, followed by heating of the metal in a second furnace. For each of these processes, specific models are used to monitor the respective operation.

[0079] Then comes hot rolling, cooling and coiling followed by cold rolling and galvanizing, which in turn is monitored by appropriate models.

[0080] Figure 1 shows that all models for monitoring and controlling the individual process steps are connected to a central optimization system (shown below the plant in Figure 1), in which the optimization of the process chain takes place with early forwarding of the expected process values ​​to other models in accordance with the above procedure.

[0081] Page 11 Figure 2 schematically shows, for one of the sub-processes of the process chain, namely for a “process i”, which is preceded and followed by other sub-processes, which individual elements are used for which purposes.

[0082] Each process step (1-n) is assigned an online process model for calculating and controlling the current process i. Calculated values ​​(e.g., calculated temperatures) are sent back to the process chain optimizer. Furthermore, a material model is assigned for calculating the material properties for the current analysis.

[0083] For a future process or one to be optimized, a process step optimizer, an offline process model, and a material model are assigned. The process step optimizer creates the data set for the offline process model and can optimize a target variable. The offline process model can be called up multiple times for this purpose. The material model is then called up for the current data set of the process step optimizer, along with the data set for the future analysis. Using this material data and the process values ​​from the process step optimizer, a rapid calculation (e.g., of temperatures) is performed. The material model is used to calculate the material values ​​for the future analysis.

[0084] Figure 3 schematically outlines the flowchart for processing a request by a process chain optimizer, as it corresponds to an embodiment of the procedure described above. The communication process of a change request by the process chain optimizer (PCO) is illustrated here using the example of a request to the continuous casting process.

[0085] The process starts (at number "1") with a request from the process chain optimizer PKO to the continuous casting model SGM regarding a new desired target value (for example, a higher casting speed is desired).

[0086] Page 12 At item “2”, the continuous casting model SGM queries the continuous casting computer / database SGR for limit values ​​for the target size and for possible side conditions (for example, it can be requested what maximum casting speed is permissible for the current material at the current strand thickness).

[0087] The return of the current limit values ​​from the continuous casting computer / database SGR to the continuous casting model SGM for the current process conditions takes place at point “3” (for example, a currently maximum possible casting speed is transmitted).

[0088] The continuous casting model SGM optimizes the target parameter (for example, by calculating the maximum possible casting speed so that solidification remains within the supported strand guide under the current process conditions such as material and casting temperature), as indicated by the number "4".

[0089] At point “5”, the response of the continuous casting model SGM to the process chain optimizer PKO is given regarding the target variable.

[0090] At point "6", the process chain optimizer PKO queries the subsequent processes to determine whether the identified target variable is advantageous or feasible (for example, whether the furnace can heat the produced slabs quickly enough). If so, at point "7", the process chain optimizer PKO queries the preceding processes to determine whether they can meet the new requirements (for example, whether the steelworks can install the ladle connection).

[0091] If this is also the case, the process chain-optimized PKO confirms the new target value for the continuous casting model SGM at item "8".

[0092] At point "9", the new target value is sent from the continuous casting model SGM to the continuous casting computer SGR for the purpose of changing the plant control (for example).

[0093] In this context, page 13 shows the display of the new casting speed in the “Level 1 HMI” for the operator of the plant).

[0094] If the operator agrees to the new target variable, he confirms the transfer from the continuous casting computer SGR to the continuous casting model SGM in accordance with item “10”.

[0095] At point “11”, the new target value is finally confirmed by the continuous casting model SGM to the process chain optimizer PKO.

[0096] Figures 4a / 4b show an example of how the optimization of the final product can be carried out:

[0097] The end product of a combined casting and rolling mill is the finished coil. Customer specifications for this include, for example, the strand thickness, strand width, material, and mechanical properties (such as tensile strength, yield strength, and hardness). Consequently, the strand geometry, for instance, can only be modified to the extent necessary to produce a specified order; similarly, strand width and strand thickness cannot be arbitrarily changed.

[0098] To meet the requirements, the following process data must be coordinated: the mass flow from the steelworks, the casting speed, the average temperature after the continuous casting plant (caster), the average temperature after the furnace or at the rolling mill, and the rolling force.

[0099] Figure 4a shows, as an example, the process of a current production.

[0100] New ladles are constantly being produced in the steelworks, so that the casting machine (caster) can cast a strand width of 2,000 mm at a casting speed of 5 m / min (and, for example, a strand thickness of 120 mm) without interruption.

[0101] Page 14The Caster online model calculates sump length and average temperature at the furnace inlet and sends these values ​​to the process chain optimizer.

[0102] The process chain optimizer sends the slab geometry and the furnace inlet temperature to the furnace online model.

[0103] The furnace online model controls the furnace so that, with the current values ​​of the caster, the furnace outlet temperature is greater than or equal to a minimum temperature for the rolling mill and sends the furnace outlet temperature to the process chain optimizer.

[0104] The process chain optimizer sends the slab geometry and the furnace outlet temperature to the rolling mill online model.

[0105] Using the data from the furnace, the rolling mill online model calculates the required rolling forces and sets up the stands to produce a 2 mm thick strip.

[0106] If, as shown in Figure 4b, problems in the steelworks result in less crude steel being produced, the casting speed in the caster must be reduced to ensure proper ladle connection. This lowers the average temperature of the slabs after the caster and, depending on the length and capacity of the furnace, also the average temperature at the entry into the rolling mill. To achieve the same number of cuts at the same width, a higher rolling force is required at a lower average temperature. If the maximum permissible rolling force is exceeded, the number of cuts must be reduced in the rolling mill, or the strand width must be reduced in the caster. Therefore, production planning must include a check to see if there are any orders for the current material with a smaller strand thickness or width.

[0107] Figure 4b shows numbers (1 to 10) for this purpose, to which the following is noted:

[0108] Page 15 According to item 1, a message is sent from the steel plant optimizer to the process chain optimizer: 20% less steel available.

[0109] According to section 2, the process chain optimizer requests the following from the caster optimizer: What casting speed is still possible for the smaller quantity of steel? The caster optimizer calculates: With the same strand geometry, reducing the casting speed from 5 to 4 m / min ensures the ladle connection. The caster offline model calculates the new furnace inlet temperature. The caster optimizer sends these values ​​to the process chain optimizer.

[0110] According to section 3, the process chain optimizer sends the new furnace inlet temperature to the furnace optimizer. The furnace optimizer then activates all burners to achieve a sufficiently high furnace outlet temperature. The offline furnace model calculates the new furnace outlet temperature and sends this information back to the process chain optimizer.

[0111] According to section 4, the process chain optimizer sends the new furnace outlet temperature to the rolling mill optimizer: The rolling force is too high. The production of 2.5 mm thick strips would be possible. Feedback is then sent back to the process chain optimizer.

[0112] According to point 5, the process chain optimizer asks during production planning: Are there any orders for a strip thickness of 2.5 mm? Answer: No, but there are orders for a smaller strip width (1,500 instead of 2,000 mm).

[0113] According to section 6, the process chain optimizer calculates that the mass flow rate of a strand width of 2,000 mm at a casting speed of 4 m / min corresponds to the same mass flow rate as 1,500 mm at 5.33 m / min.

[0114] According to section 7, the process chain optimizer requests the following from the caster optimizer: If the sump tip is still at a casting speed of 5.33 m / min

[0115] Page 16 in the system? Answer: No, but still at 5.2 m / min. The data is returned to the process chain optimizer along with the new furnace inlet temperature.

[0116] According to section 8, the process chain optimizer queries the furnace optimizer: Is the furnace inlet temperature high enough? Answer: Yes; feedback is provided.

[0117] According to section 9, the process chain optimizer requests the rolling mill optimizer to ask: Can the required rolling force be applied? Answer: Yes; feedback was provided.

[0118] According to section 10, once all sub-processes have completed their sub-task, the process chain optimizer sends the message to the Caster Online model: Reduce the strand width to 1,500 mm and increase the casting speed to 5.2 m / min.

[0119] The described sequence can be fully automatic or, in another embodiment, must be set by the caster operator.

[0120] Figures 5a and 5b illustrate further possible embodiments of the method.

[0121] The composition of the analysis can then be optimized in the steel plant. In the so-called conditioned stage, the final opportunity to modify the analysis, some alloying elements can be added or removed as needed. Every material has a possible minimum and a possible maximum proportion for each alloying element. Changes can be made within these limits.

[0122] In this context, the copper content in the material is also important:

[0123] Page 17. Especially when recycling secondary scrap, the copper content of the resulting melt is high. Depending on the amount of material removed during hot rolling, there are upper and lower limits for the copper equivalent. With larger removal rates, the elongation increases, which renders the "hot shortness cracks" that occur after casting harmless. However, with smaller removal rates (corresponding to larger strip thicknesses), this can lead to a reduction in quality.

[0124] The copper equivalent can be determined, for example, by:

[0125] Copper equivalent = % Cu + 10 (% Sn + % Sb) - % Ni

[0126] (Cu: copper; Sn: tin; Sb: antimony; Ni: nickel)

[0127] Accordingly, the copper equivalent can be reduced by adding nickel. If the steel mill knows the strip thickness to be rolled in the near future, the strip quality can be improved directly at the mill by adding nickel. Conversely, the expensive alloying element nickel can be saved if thinner strips are to be produced in the future.

[0128] Figure 5a shows the curve of the copper equivalence over the strip thickness of the metal strip to be produced, illustrating the influence of the alloying element nickel.

[0129] Since the process chain optimizer knows the strip thickness to be produced in the future, the nickel alloy can be optimized in the steelworks.

[0130] Intensive cooling in the casting machine results in grain refinement, which allows for a higher copper equivalent.

[0131] Figure 5b schematically illustrates the influence of intensive cooling on the permitted copper equivalent.

[0132] Page 18. On the other hand, active intensive cooling reduces the strand temperature in the caster. The process chain optimizer can now decide in advance whether it is more advantageous for a given strip thickness to alloy more nickel or to activate intensive cooling and increase the burner output in the furnace. With a ladle that has already been produced and has a high copper content, the only options are to activate existing intensive cooling or to produce an order with a thinner strip (corresponding to a higher uptake in the rolling mill).

[0133] Figure 6 schematically illustrates an example of optimizing a strand thickness. The figures in the figure are numbered 1 to 14, which will be referenced below.

[0134] According to item 1, the caster is currently producing a strand thickness H = 65 mm at a casting speed of 4.5 m / min.

[0135] According to point 2, a larger take-off in the rolling mill would be advantageous, which is why the rolling mill optimizer asks the process chain optimizer whether a strand thickness of 70 mm can also be produced in the caster under the current other process values ​​(casting speed, casting temperature, water quantities, water temperature ...).

[0136] According to point 3, the process chain optimizer asks the caster optimizer whether H = 70 mm is also possible under the current process conditions.

[0137] According to section 4, the caster optimizer queries the caster database for the maximum possible swamp length.

[0138] According to section 5, the Caster database specifies a maximum sump length of 8,100 mm for the current material. Additionally, the Caster optimizer receives information indicating that a sump length between 8,050 and 8,100 mm is also considered optimal. This range is intended to ensure that an optimal strand thickness can be achieved with two

[0139] Page 19: Decimal places, preferably with only one decimal place, are displayed as the optimal value.

[0140] According to section 6, the caster optimizer calls the fast caster offline model for the first time to calculate the sump length for a strand thickness of 70 mm using otherwise current process data.

[0141] According to section 7, the fast caster offline model calculates the developing sump length in a few seconds and sends it back to the caster optimizer, e.g.: 8,050 mm <= first sump length <= 8,100 mm.

[0142] The caster optimizer sends a signal to the process chain optimizer indicating that a strand thickness of H = 70 mm can be cast.

[0143] The process chain optimizer sends a message to the Caster Online model indicating that a strand thickness of 70 mm should be used in the current casting operation.

[0144] According to section 8, if the first calculated sump length is too long, the caster optimizer optimizes the strand thickness. For example, it might proceed as follows: the current thickness of 65 mm is possible, 70 mm is not.

[0145] According to item 9, a new second invoice is issued with a thickness of 67.5 mm.

[0146] According to item 10, it is determined that a second calculated sump length of 8,030 mm is too short. Further strand thicknesses can be determined by the caster optimizer using the interval box method or better mathematical procedures (secant method, Müller method).

[0147] According to clause 11: If the caster optimizer has found a sump length between 8,050 and 8,100 mm (e.g. for a thickness of 69.2 mm), it sends back to the process chain optimizer that a strand thickness of 70 mm is not possible, but 69.2 mm is.

[0148] Page 20 According to item 12, the process chain optimizer asks the rolling mill optimizer whether a strand thickness of 69.2 mm would also be advantageous and should be produced.

[0149] According to section 13, the rolling mill optimizer checks whether this thickness is advantageous and sends the result back to the process chain optimizer.

[0150] According to section 14, it is possible that no benefit can initially be achieved, and therefore the process chain optimizer takes no further action and waits for the next request. However, it is also possible that a strand thickness of 69.2 mm offers advantages: in this case, the process chain optimizer sends a message to the Caster Online model indicating that the strand thickness should be increased to 69.2 mm.

[0151] The new strand thickness can now be changed fully automatically, or the caster operator is shown that a larger strand thickness offers advantages in the rolling mill; he then changes the thickness manually.

[0152] The new strand thickness can be achieved by dynamically changing the LCR and / or soft reduction setting.

[0153] It is also possible to vary the pouring speed or water cooling to save energy:

[0154] By varying the cooling water volume and / or the casting speed, the casting machine optimizer can adjust the sump length to the permissible end of the supported strand guide. This is done depending on the other existing process parameters (superheating, strand thickness, water temperature, mold values, etc.). At maximum sump length, the strand energy increases; the reheating unit requires less energy and therefore also produces fewer CO2 emissions. A higher casting speed reduces the transport time of the strand head between a shear and the inlet to the reheating unit, thus minimizing energy loss through the

[0155] Page 21: Reduced heat radiation. Optimizing the casting speed can also be done as part of an overall optimization together with the strand thickness.

[0156] Since the process models upstream of the casting machine and any downstream furnace can exchange data via the higher-level process optimizer, the furnace inlet temperature optimized by the faster, second process model can be sent to the furnace early on. This enables targeted, predictive heating of the furnace zones. In this way, changes in the current casting process can be taken into account in other system components at an early stage.

[0157] In continuous casting, a casting interruption followed by a restart should be avoided. If problems arise with the ladle connection in the steel plant, the casting machine's optimizer can calculate the minimum possible mass flow rate using the minimum possible casting speed and transmit this information to the higher-level process optimizer. There, a decision is then made as to whether or not to interrupt the casting process. During optimization, the casting speed and strand thickness must be selected so that the calculated bottom tip lies behind the LCR zone in the casting direction and that the strand surface does not become too cold in the straightening area.

[0158] In the event of current problems in the rolling mill, the current production of the casting machine can be reduced so that no casting interruption is necessary and the furnace can temporarily store the produced slabs. It may also be possible to temporarily store more material in the furnace by using a greater strand thickness in the current casting process.

[0159] A uniform temperature distribution is also sought to improve quality:

[0160] Temperature differences can occur within a slab after continuous casting. By knowing the tolerances allowed in the rolling mill and the current temperature equalization possible in the furnace, the casting optimizer can be used to try to correct these differences.

[0161] Page 22 will be used to minimize these temperature differences by optimizing the pouring speed and the amount of splash water in the current pour.

[0162] The structure is then calculated and optimized as follows:

[0163] To calculate and optimize the mechanical properties, the microstructure properties present after the casting process (such as solidification structure, precipitates, for example the formation and dissolution of carbides, austenite-ferrite transformation or grain size) can be passed from the casting computer to the higher-level process model.

[0164] Additionally, the casting optimizer makes it possible to change the process parameters of the current casting process in such a way that the microstructure properties and thus also the mechanical properties of the currently produced strip are optimized.

[0165] Hot-casting of slabs is a well-known method for reducing energy consumption in steel production. The casting machine is typically pre-defined in a production plan, specifying the material to be cast, its strand width, thickness, casting speed, and cooling water volume. Control systems allow for adjustments to the water volume used for the current casting or to vary the average temperature at the furnace inlet.

[0166] These fixed specifications, such as the strand thickness, are pre-selected to ensure they can be processed in all subsequent process steps (reheating, hot rolling mill). Actual process conditions are not taken into account. For example, the current analysis (chemical composition) could differ from the reference analysis of the material to be cast, resulting in different material properties, or the current melt temperature could deviate from the planned target temperature, leading to changes in the casting speed.

[0167] Page 23 With real-time calculation, it is not possible to determine in advance the necessary changes to the current process parameters for the casting process and to ascertain the influence on the subsequent process steps.

[0168] Accordingly, in a preferred embodiment of the present invention, it is also provided that a second process model and an additional casting optimizer are integrated into the existing online process model for the control or regulation of the current casting process.

[0169] The casting optimizer is in constant data exchange with the higher-level process optimizer, which handles communication with production planning and the process models (steelworks, ladles, and distributors) of upstream process steps, as well as with downstream process steps (reheating unit, hot rolling mill, cooling sections, etc.). The target specifications required for the current casting process are sent—as explained above—from the process models to the higher-level process optimizer, which then forwards them to the casting machine's optimizer.

[0170] The casting optimizer also communicates with the real-time online process model and a second, faster-than-real-time process model. Additionally, it considers the casting specifications stored in the database, which are specific to each steel grade. The target parameters required for optimization (such as casting speed, temperature, exit thickness, microstructure, and precipitation state) are provided to the casting optimizer by the higher-level process optimizer from upstream and downstream processes, as well as from production planning.

[0171] The second process model, which calculates faster than real time, can predict how the influence of, for example, a different chemical composition or a change in the casting temperature or casting speed of the current casting process will affect the target specifications and forwards the results to the casting optimizer. The same applies to requirements from downstream systems.

[0172] Page 24 Process steps (such as higher rolling speed or higher forming temperature). The casting optimizer determines which casting parameters of the current casting must be changed to meet the target specification. The new casting parameters are checked by the second process model, which calculates faster than real time. If the target specification cannot be achieved, the best possible solution is determined and forwarded by the casting optimizer to the higher-level process optimizer. Only after acceptance by the higher-level process optimizer are the casting parameters passed on to the current casting process.

[0173] The second process model, which calculates faster than real time, is also capable of performing optimizations independently.

[0174] The optimizer contains models with optimization strategies that adjust the current casting process so that the required target specifications are met in the subsequent process steps.

[0175] Possible optimization strategies include, for example, root finding, gradient method, linear and integer optimization, discrete optimization, nonlinear optimization, and data mining.

[0176] The following should be noted regarding the possibilities for optimizing the casting process:

[0177] All optimizations must prioritize a reliable casting process: The sump tip must remain within the supported strand guide to prevent wriggling (see the tested sump length in Figure 8). The strand shell thicknesses in the upper part of the system must be sufficient to prevent fluctuations in the casting level. Surface temperatures in the bending and straightening area must not be too low to prevent surface cracking. Only when these conditions are met can the casting process be further optimized.

[0178] For this purpose, a variation in strand thickness is possible to increase production, as can be seen from the following explanations:

[0179] Page 25Since the casting machine is usually the bottleneck in the overall production during hot operation in a slab, CSP, Nexus or billet plant, the material flow of the casting machine plays a crucial role.

[0180] The mass flow rate (V * B * H * p) is limited by the maximum possible sump length. A constant sump length also corresponds to a constant K-factor, as defined below:

[0181] K = 0.5 * H / ^Swamp length / V

[0182] "Swamp length = V * H" 2 / (4*K 2 )

[0183] => V = 4*swamp length*K 2 / H 2

[0184] <

[0185]

[0186] >

[0187] Mass flow rate = 4 * B * p * sump length*K 2 / H

[0188] Here, H is the strand thickness, B is the strand width, V is the casting speed, and K is the K-factor.

[0189] A smaller strand thickness combined with a higher casting speed can lead to higher production with a constant sump length than a larger strand thickness.

[0190] Depending on the existing chemical composition and the resulting rolling forces required in the rolling mill for thickness reduction, the higher-level process optimizer can calculate a desired strand thickness and thus optimize overall production. This desired optimal strand thickness is then sent to the casting machine's optimizer, where it is determined whether it is suitable for the casting process.

[0191] In the positive case: Use of the optimized strand thickness for the current casting.

[0192] Page 26 In the negative case: Determination of the best possible strand thickness for the current casting process and return to the higher-level process optimizer.

[0193] The new strand thickness can be achieved by dynamically changing the LCR and / or soft reduction setting.

[0194] The following should be considered when calculating and optimizing the mass flow rate:

[0195] The mass flow rate is calculated (as explained above) from the product of strand width, strand height, casting speed, and density. Since density is a function of temperature, the mass flow rate at the end of the casting machine can change depending on the current temperature profile, even with the same strand cross-section and casting speed. The mass flow rate must then be optimized for the current casting to ensure the flow to the ladle, distributor, and dip tube. Likewise, the mass flow rate in the casting machine may need to be reduced to allow for a connection to the next ladle and prevent casting interruptions.

[0196] Due to problems in the subsequent rolling mill (for example, due to a temporarily reduced rolling force), there may be requirements to reduce the current strand cross-section.

[0197] The following applies to the calculation and optimization of the energy flow:

[0198] The energy flow is calculated by multiplying the mass flow rate by the enthalpy in J / s.

[0199] With the mass flow rate MP, the enthalpy of the caster Hcaster, the enthalpy of the furnace H F The new furnace load is calculated using the furnace efficiency r|Fumace and the furnace efficiency r|Fumace:

[0200] MP * (H Furnace-Hcaster) / hFurnace

[0201] Page 27. This means that, for example, with a higher casting speed or greater thickness and the same temperature, the mass flow rate (MP) and thus the load on the downstream furnace increases. Depending on the furnace's capacity, specifications for maintaining a maximum energy flow to the casting machine may be imposed, requiring adjustments to the current process parameters of the current casting.

[0202] Alternatively, the casting optimizer can also be used in production plants where the cast steel is not directly processed further in a rolling mill.

[0203] Depending on the current process parameters and the desired product quality, the produced slabs can pass through a cooling bed, a heating hood, and / or a pit, or be stacked afterward. The casting optimizer can optimize the process parameters of the current casting for the desired operating mode.

[0204] Page 28

Claims

Patent claims:

1. Method for producing a continuously cast metallic product, in which metallic molten metal is produced and provided in a sub-process, in which a metal strand is cast in a subsequent sub-process and in which, in a subsequent sub-process, the cast metal strand is subjected to a further process, where the subprocesses are calculated and / or controlled or regulated by respective subprocess models and where the sub-process models are linked to a higher-level overall process optimization system, where a) the overall process optimization system sends a request to a subprocess model (requested subprocess model) in order to obtain from it an optimized target value for a production parameter for the subprocess controlled or regulated by this subprocess model, b) the requested subprocess model determines a possible optimized target variable for the subprocess it monitors and reports this back to the overall process optimization system, Page 29c) the overall process optimization system, with the reported optimized target variable, sends a query to at least one first subprocess model upstream or downstream of the requested subprocess model, asking whether the operation of the subprocess controlled or regulated by this subprocess model is possible with the optimized target variable, and the overall process optimization system, with the reported optimized target variable, sends a query to at least one second subprocess model upstream or downstream of the requested subprocess model, asking whether the operation of the subprocess controlled or regulated by this subprocess model is possible with the optimized target variable. d) In the event that the requests according to step c) are answered positively by the first and the second upstream or downstream sub-process model: Initiate the production of the continuously cast metallic product with the optimized target size.

2. Method according to claim 1, characterized in that the first sub-process model is a sub-process model downstream of the requested sub-process model and the second sub-process model is a sub-process model upstream of the requested sub-process model.

3. Method according to claim 1 or 2, characterized in that the instigation of the manufacture according to step d) of claim 1 comprises the steps: Confirmation of the optimized target variable for the requested sub-process model, Modification of the sub-process based on the optimized target variable and confirmation of this to the overall process optimization system, Page 30: Specification of the optimized target variable to the first and second sub-process models by the overall process optimization system.

4. Method according to any one of claims 1 to 3, characterized in that each sub-process model comprises: an online model that calculates and / or controls or regulates the subprocess, an offline model for calculating at least one production parameter and an optimizer that determines optimal values ​​for production parameters calculated by the offline model.

5. Method according to claim 4, characterized in that the offline model computes the subprocess at a speed that is at least 10 times that of real time, preferably at least 50 times that of real time.

6. Method according to one of claims 4 or 5, characterized in that a material model is assigned to the sub-process model or that the sub-process model comprises a material model.

7. Method according to claim 6, characterized in that a single material model is assigned to all sub-process models. Page 318. Method according to one of claims 1 to 7, characterized in that the requested sub-process model according to step a) of claim 1 is a continuous casting model for continuous casting.

9. Method according to one of claims 1 to 8, characterized in that the upstream sub-process model is a converter model for the converter process and / or a ladle model for the provision of melt and / or a distributor model for the distribution of the melt.

10. Method according to one of claims 1 to 9, characterized in that the downstream sub-process model is a furnace model for the operation of a furnace.

11. Method according to any one of claims 1 to 10, characterized in that the downstream sub-process model is a rolling model for rolling the metallic product, in particular a roughing mill model for rolling in a roughing mill and / or a hot rolling mill model for hot rolling the metallic product and / or a cold rolling mill model for cold rolling the metallic product.

12. Method according to one of claims 1 to 11, characterized in that the downstream sub-process model is a model for a pre-band cooler and / or a cooling model and / or a coiling model.

13. Method according to one of claims 1 to 12, characterized in that the downstream sub-process model is a model for a galvanizing plant. Page 3214. Method according to one of claims 1 to 13, characterized in that the query of the overall process optimization system to a sub-process model according to step a) of claim 1 is made depending on a requirement for the metallic product to be manufactured, in particular for the metal strip, wherein the requirement relates in particular to the geometry of the metal strip and / or the quality of the metal strip.

15. Method according to one of claims 1 to 14, characterized in that the process is carried out using copper-containing scrap, wherein the overall process optimization system optimizes the alloy to be produced and / or the cooling, in particular the intensive cooling, of the metal strip to be produced depending on the strip thickness to be produced.

16. Plant for the production of a continuously cast metallic product, in particular for carrying out the process according to any one of claims 1 to 15, comprising a sub-process in which metallic molten metal is produced and provided, a subsequent sub-process in which a metal strand is cast, and a further sub-process in which the cast metal strand is subjected to a further process, wherein the subprocesses have respective subprocess models which are designed for the calculation and / or control or regulation of the subprocess, wherein a higher-level overall process optimization system is arranged, which is linked to the sub-process models, where Page 33a) the overall process optimization system is designed to send a request to a sub-process model (requested sub-process model) in order to obtain from it an optimized target value for a production parameter for the sub-process controlled or regulated by this sub-process model, b) the requested sub-process model is designed to determine a possible optimized target variable for the sub-process it monitors and to report this back to the overall process optimization system, c) the overall process optimization system is configured to send the feedback of the optimized target variable to at least one first subprocess model upstream or downstream of the requested subprocess model, asking whether the operation of the subprocess controlled or regulated by this subprocess model is possible with the optimized target variable, and the overall process optimization system is configured to send the feedback of the optimized target variable to at least one second subprocess model upstream or downstream of the requested subprocess model, asking whether the operation of the subprocess controlled or regulated by this subprocess model is possible with the optimized target variable. d) the overall process optimization system is designed, in the event that the queries according to step c) are answered positively by the upstream and downstream sub-process model, to ensure that the production of the continuously cast metallic product with the optimized target size takes place. Page 3417. System according to claim 16, characterized in that each sub-process model comprises: an online model that is trained to calculate and / or control or regulate the subprocess, an offline model that is trained to calculate at least one production parameter and an optimizer trained to determine optimal values ​​for the production parameters calculated by the offline model.

18. System according to claim 17, characterized in that a material model is assigned to the sub-process model or that the sub-process model comprises a material model.

19. System according to claim 18, characterized in that a single material model is assigned to all sub-process models.

20. Plant according to one of claims 16 to 19, characterized in that the requested sub-process model is a continuous casting model for continuous casting.

21. Plant according to one of claims 16 to 20, characterized in that the upstream sub-process model is a converter model for the converter process and / or a ladle model for the provision of melt and / or a distributor model for the distribution of the melt. Page 3522. Plant according to one of claims 16 to 21, characterized in that the downstream sub-process model is a furnace model for the operation of a furnace.

23. Plant according to one of claims 16 to 22, characterized in that the downstream sub-process model is a rolling model for rolling the metallic product, in particular a roughing mill model for rolling in a roughing mill and / or a hot rolling mill model for hot rolling the metallic product and / or a cold rolling mill model for cold rolling the metallic product.

24. Plant according to one of claims 16 to 23, characterized in that the downstream sub-process model is a model for a pre-band cooler and / or a cooling model and / or a coiling model.

25. Plant according to one of claims 16 to 24, characterized in that the downstream sub-process model is a model for a galvanizing plant.

26. Computer program comprising program commands for executing the sub-process models and for communication with the higher-level overall process optimization system when carrying out the method according to any one of claims 1 to 15. Page 36

Images