Technical Field
-
The present invention relates to a method for, when continuously casting steel, determining the quality of a product or a continuous cast slab from casting conditions and actual measured values, and also relates to a method for determining the destination of the continuous cast slab, and a method for determining a continuous casting condition. In the following description, the unit "t" of mass represents 103 kg, and the unit "L" of volume represents 10-3 m3. In addition, symbol "N" added to the unit of the volume of a gas represents the volume of the gas in the standard state, where the standard state is assumed to be a state with a temperature of 0°C and a pressure of 101325 Pa.
Background Art
-
During continuous casting of steel, gas bubbles blown into a nozzle as well as non-metallic inclusions, such as deoxidation products and sulfides, may be captured by a solidified shell, and thus may remain in the surface layer portion of a final product. Such gas bubbles and non-metallic inclusions degrade the quality of steel products, especially, thick steel plates. For example, hydrogen induced cracking occurs in line pipe materials for use in the transportation of oil or the transportation of a natural gas due to the action of a sour gas, with gas bubbles or non-metallic inclusions serving as the starting point. Similar problems also occur in offshore structures, storage tanks, and oil tanks, for example. Further, in recent years, there has been increasing demand for the use of steel materials under a severe environment, such as a lower-temperature environment or a more highly corrosive environment. Thus, it has become more important to reduce gas bubbles and non-metallic inclusions in cast slabs.
-
To this end, an HIC (Hydrogen Induced Cracking) test is performed on line pipe steel, which is referred to as a sour-resistant material, before shipment, so that only products with no HIC are shipped as sour-resistant materials. However, the HIC test requires several weeks for the availability of test results. When a product is found to have HIC, it cannot be shipped as a sour-resistant material, resulting in a significant reduction in yields. Therefore, evaluating the HIC performance at the stage of cast slab prior to plate rolling, without conducting an HIC test, could reduce production time and significantly improve yields.
-
Patent Literature 1 discloses a method of measuring the thickness of an opening resulting from horizontal cracking as well as the maximum segregated grain size along a cut plane of a slab, and determining a threshold from the measurement results as well as the test results of HIC measurement, and then changing the destination of the slab. Patent Literatures 2 and 3 each disclose a method for continuously casting steel involving compensating for HIC by satisfying a given Ca/S ratio as well as a relational expression of Ca, S, and O, and by further setting the amount of decrease in Ca to a threshold or less. Patent Literature 4 discloses an evaluation method that can accurately detect the center segregation by binarizing an etched print image of a cross-section of a steel material.
Citation List
Patent Literature
-
- Patent Literature 1: JP-2015-58473A
- Patent Literature 2: JP-2016-125137A
- Patent Literature 3: JP-2016-125140A
- Patent Literature 4: JP-2017-181030A
Summary of Invention
Technical Problem
-
However, the foregoing conventional technologies have the following problems.
-
The technology disclosed in Patent Literature 1 involves measuring segregated grains along a cut plane of a slab, which is problematic in reducing the production period. While each of the technologies disclosed in Patent Literatures 2 and 3 can prevent HIC cracks originating from Ca-based inclusions, neither can address those originating from other non-metallic inclusions. Further, the technology disclosed in Patent Literature 4 only evaluates the center segregation, and has not clarified the correlation between gas bubbles and non-metallic inclusions and HIC cracks.
-
A test for HIC cracks resulting from non-metallic inclusions requires a period of seven days even after the production of a steel plate. The quality of a product is deemed defective only after a large quantity has already been produced, sometimes resulting in a significant number of defective products. Meanwhile, if defects could be predicted after casting, measures could be taken, such as remelting, without proceeding to the subsequent processes. However, since it is currently impossible to determine the presence or absence of defects in the casting stage, the process needs to be continued until a final product is produced. If the final product is evaluated defective after its production, it cannot be shipped as a non-defective product, which results in increased costs.
-
The present invention has been made in view of the foregoing circumstances, and an object of the present invention is to propose a method for determining the quality of a product that includes predicting the quality of a product produced by rolling a cast slab during the cast slab phase. It is another object of the present invention to propose a method that can determine the quality of a cast slab produced by a continuous casting machine, particularly the HIC properties resulting from gas bubbles and non-metallic inclusions, during or after casting. Furthermore, the present invention proposes a method for determining the destination of a continuous cast slab, a method for determining a continuous casting condition, and a method for continuously casting steel.
Solution to Problem
-
The inventors have completed the present invention by discovering that it is possible to predict the area ratio of hydrogen induced cracking (HIC) derived from gas bubbles and non-metallic inclusions, based on the casting track record data on a casting process, including parameters such as the cross-sectional size of a cast slab; the component composition; casting rate; the electromagnetic agitation conditions; the lead time taken from the secondary refining to the start of casting; the amounts of auxiliary raw materials added; the inert gas flow rate blown into a nozzle; the immersion depth of an immersion nozzle.
-
That is, it has been found that the following invention can advantageously solves the foregoing problems.
- [1] A method for determining a quality of a product, when determining a quality of a product produced by rolling a cast slab cast by a continuous casting machine, including utilizing a prediction model for hydrogen induced cracking in a surface layer portion of the product and predicting hydrogen induced cracking in the surface layer portion of the product by employing one or more actual measured values of casting track record data obtained during casting as an input variable for the prediction model.
- [2] A method for determining a quality of a continuous cast slab, when determining a quality of a cast slab cast by a continuous casting machine using the method for determining a quality of a product according to [1], in which
- the prediction model is obtained by associating casting track record data with a crack area ratio of hydrogen induced cracking that has occurred in a surface layer portion of a product, and
- the method comprises inputting one or more actual measured values of the casting track record data obtained during casting into the prediction model, thereby predicting a crack area ratio of hydrogen induced cracking that will occur in a surface layer portion of a product obtained from the cast slab during or after casting.
- [3] The method for determining a quality of a continuous cast slab according to [2], in which the casting track record data includes at least one of the following: the cross-sectional size of the cast slab, the component composition, the casting rate, the electromagnetic agitation condition, the lead time taken from secondary refining to start of casting, the amount of an auxiliary raw material added, the inert gas flow rate blown into a nozzle, and the immersion depth of an immersion nozzle.
- [4] The method for determining a quality of a continuous cast slab according to [3], in which the component composition includes at least one selected from the group consisting of a C concentration, a Mn concentration, an S concentration, and a C equivalent calculated as Ceq (mass%) using the expression below: Ceq = [C]-0.0616[Al] + 2.5275[S]-0.2652[P] + 0.0023[Si] + 0.0344[Mn] - 1.525[S][Mn] + 0.021[Si][Mn] + 0.02[Cu] - 0.02[Mo] + 0.06[Ni] + 0.02[Cr] - 0.04[V] - 0.04[Nb],
where [M] in the expression represents a content of an element M indicated by mass percentage.
- [5] The method for determining a quality of a continuous cast slab according to any one of [2] to [4], in which the prediction model employs principal component analysis and regression based on Random Forests, and optionally, the prediction model is trained through machine learning using an actual measured value of a crack area ratio of hydrogen induced cracking in the surface layer portion of the product.
- [6] A method for determining a destination of a continuous cast slab, including, based on quality prediction of a cast slab determined by the method for determining a quality of a continuous cast slab according to any one of [2] to [5], determining whether the cast slab is applicable as sour-resistant line pipe steel.
- [7] A method for determining a continuous casting condition, including, based on a quality prediction of a cast slab determined by the method for determining a quality of a continuous cast slab according to any one of [2] to [5], determining the casting condition through inverse analysis based on the casting track record data and the prediction model to allow a predicted value of a crack area ratio of hydrogen induced cracking in a surface layer portion of a product to asymptotically approach a predetermined value.
- [8] The method for determining a continuous casting condition according to [7], in which the predetermined value is 2% or less.
- [9] A method for continuously casting steel, including producing a cast slab based on the casting condition determined by the method according to [7] or [8].
Advantageous Effects of Invention
-
According to the present invention, the actual measured value of casting track record data is input into a prediction model prepared in advance, enabling the quality of a product or a cast slab, particularly the crack area ratio of hydrogen induced cracking in the surface layer portion of the product to be predicted, during or after casting. This achieves accurate prediction of whether the cast slab is applicable as a predetermined product, thus ensuring high production yields. Further, it becomes possible to produce products in higher yields by determining the casting conditions such that the obtained predicted value asymptotically approaches a predetermined value and by producing cast slabs under the determined casting conditions. This increases productivity, which is industrially advantageous. It is also possible to determine if a product is applicable as, for example, sour-resistant line pipe steel, from the HIC prediction value of the product, without performing a time-consuming HIC test. This allows steel products to be produced promptly in response to demand, with various specifications, which is industrially advantageous.
Brief Description of Drawings
-
- [Fig. 1] is a schematic side view schematically illustrating a slab continuous casting machine suitable for the implementation of the present invention.
- [Fig. 2] is a graph showing the relationship between the actual measured value and the predicted value of the crack area ratio (CAR) of HIC in a surface layer portion.
- [Fig. 3] is a flowchart showing an exemplary method for predicting the quality of a continuous cast slab.
- [Fig. 4] is a schematic flowchart from a continuous casting process to shipment.
- [Fig. 5] is a graph representing the magnitude of the influence of each variable on a principal component 1 and a principal component 2 in an Example.
Description of Embodiments
-
Hereinafter, an embodiment of the present invention will be specifically described. The following embodiment only illustrates examples of a facility and a method for embodying the technical idea of the present invention. Thus, the configuration of the present invention is not limited thereto. That is, the technical idea of the present invention may be changed in various ways within the technical scope recited in the claims.
-
Fig. 1 is a schematic side view schematically illustrating a slab continuous casting machine that can be suitably used for a method for continuously casting steel according to an embodiment of the present invention. As illustrated in Fig. 1, a slab continuous casting machine 1 is provided with a casting mold 5 for injecting and solidifying molten steel 9 to form the outer shape of a cast slab 10. A tundish 2 is provided at a predetermined position above the casting mold 5 to relay the supply of the molten steel 9 from a ladle (not illustrated) to the casting mold 5. A sliding nozzle 3 is provided at the bottom of the tundish 2 to control the flow rate of the molten steel 9, and an immersion nozzle 4 is provided on the lower surface of the sliding nozzle 3. Meanwhile, a plurality of pairs of cast-slab support rolls 6 including support rolls, guide rolls, and pinch rolls are disposed below the casting mold 5. A secondary cooling zone with spray nozzles (not illustrated) such as water spray nozzles or air mist spray nozzles arranged therein is provided in the gap between the cast-slab support rolls 6 that are adjacent to each other in the casting direction FD. The secondary cooling zone is designed to cool the cast slab 10 while pulling it with cooling water (also referred to as "secondary cooling water") sprayed from the spray nozzles. A plurality of conveyance rolls 7 for conveying the cast slab 10, formed by casting, are provided on the downstream side of the final cast-slab support rolls 6 in the casting direction. A cast-slab cutter 8 is disposed above the conveyance rolls 7 to cut the cast slab 10, formed by casting, into a cast slab 10a with a predetermined length. The cross-sectional size of the cast slab is represented by the cast slab width Lw (mm) and the cast slab thickness Lt (mm).
-
A soft reduction zone 14, configured with a plurality of pairs of cast-slab support rolls as illustrated in Fig. 1, is provided on the upstream side and the downstream side in the casting direction with a solidification completion position (final solidification position; crater end: CE) 13 of the cast slab 10 interposed therebetween. The soft reduction zone 14 is designed such that the gap (referred to as a "roll gap") between the cast-slab support rolls facing each other with the cast slab 10 interposed becomes narrower sequentially toward the downstream side in the casting direction. That is, a reduction gradient (i.e., the state of the roll gap set such that it becomes narrower sequentially toward the downstream side in the casting direction) is provided. The cast slab 10 can be subjected to soft reduction in the entire region or a partial selected region of the soft reduction zone 14. A spray nozzle for cooling the cast slab 10 is also provided between the adjacent cast-slab support rolls in the soft reduction zone 14. The cast-slab support rolls 6 arranged in the soft reduction zone 14 are also referred to as reduction rolls.
-
An electromagnetic agitation device (not illustrated) is provided to the casting mold 5 or between the cast-slab support rolls 6. The electromagnetic agitation device has the effect of flowing molten steel 12 in an unsolidified solid phase and thus cleaning the inner surface of the solidified shell 11. In addition, an inert gas for preventing nozzle clogging is blown into the molten steel 9 from an upper nozzle (not illustrated) or the sliding nozzle 3 provided on the tundish 2.
-
As the component composition of the cast slab, the analysis value of a sample obtained from molten steel in the ladle or the tundish can be used. For example, C and Mn are known as the components having influence on the toughness of a product. It is also known that the degree of decrease in toughness increases as the C equivalent Ceq (mass%) represented by the following expression increases. The toughness of steel has influence on the HIC characteristics. Further, Mn and S, which form MnS-based non-metallic inclusions, have influence on the HIC characteristics of a surface layer portion. Ceq = [C] - 0.0616[Al] + 2.5275[S] - 0.2652[P] + 0.0023[Si] + 0.0344[Mn] - 1.525[S][Mn] + 0.021[Si][Mn] + 0.02[Cu] - 0.02[Mo] + 0.06[Ni] + 0.02[Cr] - 0.04[V] - 0.04[Nb],
where [M] in the expression represents the content of an element M indicated by mass%.
-
The casting rate Vc (m/min) has influence on the discharge rate of molten steel from the immersion nozzle 4, and serves as the index of the depth to which gas bubbles and non-metallic inclusions are penetrated into the solidified pool. The value I (A) of current applied during electromagnetic agitation has influence on the capture of gas bubbles and non-metallic inclusions as the cleaning power for the inner surface of the solidified shell 11. The lead time "time" required from the secondary refining to the start of the casting process has influence on the flotation of deoxidized inclusions. The amount of auxiliary raw materials added, for example, the unit consumption (kg/t-molten steel) of CaSi and the unit consumption (kg/t-molten steel) of FeSi has influence on the control of the morphology of S-based non-metallic inclusions and the flotation of inclusions. The flow rate QAr (NL/min) of an inert gas, for example, an Ar gas blown into the upper nozzle or the sliding nozzle serves as the index of gas bubbles to be captured by the cast slab. The immersion depth Ld (mm) of the immersion nozzle 4 has influence on the flotation separation of gas bubbles and non-metallic inclusions or on their depth of entry into the solidified pool, in relation to the direction of the flow of the discharged molten steel.
-
The foregoing casting track record data, which has influence on the amount of gas bubbles and non-metallic inclusions to be captured by the solidified shell, is used as input variables for a prediction model for the crack area ratio CAR of hydrogen induced cracking in the surface layer portion of a product. Herein, the surface layer portion of a product refers to a range extending up to 0.2 times the thickness of the product from its surface in the thickness direction. For example, the prediction model can accurately predict the crack area ratio CAR of HIC in the surface layer portion, by compressing the dimension through principal component analysis to reduce variables and then performing regression with Random Forests.
-
The principal component analysis is a method of compressing data on correlated variables without reducing the information contained in the data, thus facilitating analysis by reducing the number of variables in complex data. In the present embodiment, variables, such as "the slab width Lw, the slab thickness Lt, the C concentration [C], the Mn concentration [Mn], the S concentration [S], the C equivalent Ceq, the casting rate Vc, the value I of current applied during electromagnetic agitation, the lead time "time" required from the secondary refining to the start of casting, the amount of auxiliary raw materials added (the amount of FeSi added and the amount of CaSi added), the flow rate QAr of an Ar gas blown into the nozzle, and the immersion depth Ld of the immersion nozzle during casting" are compressed into five variables, for example. When the variables are compressed into five variables, the resulting variables can be represented as principal components 1 to 5. Thus, it is possible to represent data, which is represented by many variables, as data with fewer variables without reducing the volume of information on the data as much as possible. If the variables in the data are to be narrowed down without the use of the principal component analysis, some of the variables need to be discarded. In such a case, there may be a case where important variables also need to be discarded. The principal component analysis generates principal components in order from a first principal component so as to contain as much information as possible on the respective variables, thereby reducing the number of variables more efficiently than ordinary methods.
-
Random Forests is one of the machine learning algorithms. It is an ensemble learning algorithm that improves generalization ability by combining a plurality of weak learners of decision trees and performing cross-validation. In the regression of the present embodiment, roughly several hundred decision trees were calculated, and the results were integrated based on their mean value. That is, numerous explanatory variables were compressed into approximately five variables through principal component analysis, and regression was performed using Random Forests, enabling highly accurate regression with even a small number of pieces of data.
-
Although the present embodiment illustrates an example in which the crack area ratio CAR of hydrogen induced cracking in the surface layer portion of a product is predicted using all pieces of the foregoing casting track record data, it is also possible to, even when some of the pieces of the casting track record data are used, increase the accuracy of regression by using the compression of variables through principal component analysis.
-
Fig. 3 shows a flowchart illustrating an example of a method for predicting the crack area ratio CAR of HIC in the surface layer portion of a product. The casting conditions and online measured values are input into a prediction model (S1), and variables are compressed through principal component analysis (S2). Regression is performed with the compressed variables using Random Forests (S3), so that the crack area ratio CAR of HIC in the surface layer portion of a product resulting from gas bubbles and non-metallic inclusions is predicted (S4). In addition, the actual measured value of the crack area ratio CAR of HIC in the surface layer portion of the product is used as the training data for the principal component analysis (S5) to predict the crack area ratio CAR of HIC with higher accuracy. The obtained predicted value of CAR can be used to determine whether to proceed to the next rolling step. In addition, the obtained predicted value of CAR can be used to improve the quality of a cast slab to be formed by, during casting, adjusting the immersion depth Ld of the immersion nozzle or adjusting the value I of current applied during electromagnetic agitation to allow the obtained predicted value of CAR to asymptotically approach a predetermined value (S6).
-
Fig. 4 is a flowchart from a continuous casting process S11, rolling S12, and shipment S13. Usually, to determine the quality of a slab cast by a continuous casting machine, a distribution of gas bubbles and non-metallic inclusions is analyzed with an EPMA (S14). This analysis takes a period of one to two weeks. An HIC test is also performed to determine whether a product after rolling is shippable (S15). The HIC test involves immersing a specimen in hydrogen sulfide and then evaluating the crack area ratio (CAR) when hydrogen induced cracking occurs in the center in the thickness direction of a sheet (product) or in the surface layer thereof. This test takes a period of at least about one week. In order to determine the shippability of a product, the CAR of the product needs to meet or fall below a specified threshold. In cases where a product is deemed to be defective in this test, a significant number of products have already been produced. This means that a substantial quantity of defective products have been produced. In the present embodiment, however, product quality can be predicted during or immediately after casting, without performing an HIC test. This can significantly reduce the lead time and thus prevent the production of non-conforming products produced during the saved time.
Examples
<Example 1>
-
Hereinafter, the present invention will be described in further detail based on Example.
-
The continuous casting machine used for the following test is similar to the continuous casting machine 1 shown in Fig. 1. Low-carbon aluminumkilled steel was cast using this continuous casting machine. Tables 1 to 3 show the casting track record data such as the casting conditions, and the actual measured value and the predicted value of the crack area ratio CAR of HIC in the surface layer portion of each product, when the continuous casting method according to the foregoing embodiment was performed. Herein, the surface layer portion of each product refers to a range extending up to 0.2 times the thickness of the product from its surface in the thickness direction. The casting track record data shown in Tables 1 and 2 were input into the prediction model for the crack area ratio CAR of HIC in the surface layer portion of a product, and principal component analysis as well as regression based on Random Forests was performed. Fig. 2 illustrates a graph of the relationship between the actual measured value and the predicted value of the crack area ratio CAR of HIC in the surface layer portion of each product. With the prediction model, explanatory variables were compressed into five variables through principal component analysis, and regression was performed with Random Forests. Fig. 5 illustrates the relationship between the principal component 1 and the principal component 2 of the principal component analysis and the correlation coefficients of various operating conditions. An operating condition having a large sum of correlation coefficients with respect to the principal component 1 and the principal component 2 was determined to be a variable having large influence on the crack area ratio CAR of HIC in the surface layer portion of the product. For example, from Fig. 5, the "lead time" (i.e., the lead time "time" required from the secondary refining to the start of the casting process), "the amounts of auxiliary raw materials added" (the amount of FeSi added and the amount of CaSi added), "the immersion depth Ld of the immersion nozzle," "the casting rate Vc," and "the current I applied during electromagnetic agitation" were extracted as variables (i.e., operating conditions) having large influence. According to such a method, it is found that the actual measured value and the predicted value of the crack area ratio CAR of HIC in the surface layer portion of each product match with a high degree. Thus, such a method has made it possible to predict HIC cracks in the surface layer portion of each product during or immediately after casting.
-
Table 3 illustrates an example in which during casting, the immersion depth Ld of the immersion nozzle as well as the value I of current applied during electromagnetic agitation was changed to control the obtained predicted value of CAR to asymptotically approach 0.00%. Such control has significantly decreased the crack area ratio of HIC in the surface layer portion of each product.
-
The threshold of the crack area ratio CAR of HIC in the surface layer portion of each product would differ depending on the quality required of the product. For example, regarding a steel material for which the target crack area ratio CAR of HIC to occur in the surface layer portion of the product to be produced is 2% or less, the prediction model for the crack area ratio CAR of HIC in the surface layer portion of a product according to the foregoing embodiment was used to change the destination of a slab that has a predicted CAR value of greater than 2%. Consequently, the effect of increasing yields by 7% was obtained.
[Table 1] | No. | [C] | [Mn] | [S] | Ceq | Lw | Lt | Vc | I | Time From Secondary Refining to Start of Casting | Amount of CaSi Added | Amount of FeSi Added | QAr | Ld | Actual Measured Value of CAR | Predicted Value of CAR |
| mass% | mass% | mass% | mass% | mm | min | m/min | A | min | kg/t | kg/t | NL/min | mm | % | % |
| 1 | 0.037 | 1.34 | 0.0006 | 0.093 | 2100 | 250 | 1.03 | 400 | 47 | 70.4 | 4.17 | 90 | 186 | 0.00 | 0.05 |
| 2 | 0.039 | 1.35 | 0.0004 | 0.096 | 2100 | 250 | 1.13 | 400 | 34 | 63.2 | 4.35 | 90 | 186 | 9.48 | 9.33 |
| 3 | 0.039 | 1.35 | 0.0005 | 0.096 | 2100 | 250 | 1.12 | 400 | 39 | 61.4 | 4.24 | 90 | 266 | 1.55 | 1.11 |
| 4 | 0.039 | 1.35 | 0.0005 | 0.096 | 2100 | 250 | 1.03 | 400 | 39 | 61.4 | 4.24 | 89 | 246 | 1.38 | 1.73 |
| 5 | 0.039 | 1.34 | 0.0004 | 0.096 | 2100 | 250 | 1.13 | 500 | 57 | 60.9 | 4.21 | 90 | 209 | 2.95 | 2.02 |
| 6 | 0.039 | 1.35 | 0.0004 | 0.096 | 2100 | 250 | 1.13 | 500 | 34 | 63.2 | 4.35 | 90 | 186 | 1.04 | 1.17 |
| 7 | 0.040 | 1.53 | 0.0005 | 0.105 | 2100 | 250 | 1.13 | 500 | 43 | 61.2 | 4.01 | 89 | 226 | 0.40 | 0.88 |
| 8 | 0.040 | 1.26 | 0.0006 | 0.092 | 2100 | 250 | 1.13 | 700 | 51 | 108.4 | 3.93 | 90 | 185 | 0.77 | 0.39 |
| 9 | 0.040 | 1.35 | 0.0004 | 0.096 | 2100 | 250 | 1.03 | 700 | 45 | 62.1 | 4.21 | 90 | 247 | 0.95 | 0.57 |
| 10 | 0.040 | 1.36 | 0.0004 | 0.098 | 2100 | 250 | 1.03 | 700 | 45 | 69.5 | 4.26 | 90 | 246 | 1.35 | 1.13 |
| 11 | 0.040 | 1.35 | 0.0005 | 0.097 | 2100 | 250 | 1.03 | 700 | 46 | 63.4 | 4.23 | 90 | 227 | 7.24 | 7.27 |
| 12 | 0.041 | 1.37 | 0.0003 | 0.099 | 2100 | 250 | 1.03 | 700 | 50 | 61.2 | 4.11 | 90 | 226 | 0.39 | 0.87 |
| 13 | 0.041 | 1.37 | 0.0004 | 0.098 | 2100 | 250 | 1.03 | 700 | 37 | 61.2 | 4.37 | 90 | 265 | 0.67 | 0.54 |
| 14 | 0.041 | 1.37 | 0.0004 | 0.098 | 2100 | 250 | 1.03 | 700 | 37 | 61.2 | 4.37 | 90 | 265 | 1.70 | 1.18 |
| 15 | 0.041 | 1.35 | 0.0003 | 0.097 | 2100 | 250 | 1.13 | 500 | 47 | 64.3 | 4.21 | 90 | 238 | 1.92 | 1.46 |
| 16 | 0.041 | 1.36 | 0.0005 | 0.099 | 2100 | 250 | 1.03 | 500 | 32 | 68.1 | 3.20 | 90 | 266 | 1.98 | 2.50 |
| 17 | 0.041 | 1.33 | 0.0005 | 0.096 | 2100 | 250 | 1.13 | 400 | 45 | 60.9 | 4.15 | 90 | 204 | 2.02 | 1.82 |
| 18 | 0.041 | 1.32 | 0.0006 | 0.097 | 2100 | 250 | 1.13 | 400 | 30 | 73.3 | 3.21 | 90 | 205 | 3.40 | 3.80 |
| 19 | 0.041 | 1.32 | 0.0006 | 0.097 | 2100 | 250 | 1.13 | 400 | 30 | 73.3 | 3.21 | 90 | 186 | 6.30 | 6.46 |
| 20 | 0.041 | 1.36 | 0.0005 | 0.099 | 2100 | 250 | 1.03 | 400 | 32 | 68.1 | 3.20 | 90 | 266 | 1.98 | 2.50 |
[Table 2] | No. | [C] | [Mn] | [S] | Ceq | Lw | Lt | Vc | I | Time From Secondary Refining to Start of Casting | Amount of CaSi Added | Amount of FeSi Added | QAr | Ld | Actual Measured Value of CAR | Predicted Value of CAR |
| | mass% | mass% | mass% | mass% | mm | min | m/min | A | min | kg/t | kg/t | NL/min | mm | % | % |
| 21 | 0.041 | 1.32 | 0.0006 | 0.097 | 2100 | 250 | 1.13 | 400 | 30 | 73.3 | 3.21 | 90 | 186 | 4.60 | 5.46 |
| 22 | 0.042 | 1.33 | 0.0004 | 0.098 | 2100 | 250 | 1.13 | 400 | 61 | 106.8 | 3.21 | 89 | 186 | 0.33 | 0.48 |
| 23 | 0.042 | 1.33 | 0.0004 | 0.098 | 2100 | 250 | 1.13 | 400 | 61 | 106.8 | 3.21 | 89 | 186 | 0.88 | 0.77 |
| 24 | 0.042 | 1.34 | 0.0003 | 0.099 | 2100 | 250 | 1.13 | 400 | 35 | 61.2 | 4.26 | 90 | 226 | 0.99 | 0.65 |
| 25 | 0.043 | 1.35 | 0.0004 | 0.100 | 2100 | 250 | 1.13 | 400 | 41 | 64.3 | 3.24 | 90 | 247 | 5.50 | 5.00 |
| 26 | 0.043 | 1.34 | 0.0006 | 0.099 | 2100 | 250 | 1.03 | 600 | 38 | 69.3 | 3.20 | 90 | 246 | 2.42 | 2.18 |
| 27 | 0.043 | 1.35 | 0.0006 | 0.101 | 2100 | 250 | 1.13 | 600 | 25 | 74.2 | 3.26 | 90 | 205 | 4.14 | 3.53 |
| 28 | 0.043 | 1.35 | 0.0006 | 0.101 | 2100 | 250 | 1.13 | 600 | 25 | 74.2 | 3.26 | 90 | 205 | 3.17 | 2.51 |
| 29 | 0.043 | 1.35 | 0.0006 | 0.101 | 2100 | 250 | 1.13 | 700 | 25 | 74.2 | 3.26 | 82 | 205 | 7.65 | 6.48 |
| 30 | 0.044 | 1.34 | 0.0005 | 0.101 | 2100 | 250 | 1.13 | 700 | 80 | 108.9 | 4.27 | 90 | 205 | 0.34 | 0.24 |
| 31 | 0.044 | 1.35 | 0.0004 | 0.101 | 2100 | 250 | 1.13 | 700 | 26 | 62.7 | 3.82 | 90 | 266 | 0.43 | 0.32 |
| 32 | 0.044 | 1.35 | 0.0004 | 0.101 | 2100 | 250 | 1.13 | 700 | 26 | 62.7 | 3.82 | 90 | 266 | 0.65 | 0.43 |
| 33 5 | 0.046 | 1.36 | 0.0005 | 0.104 | 2100 | 250 | 1.13 | 700 | 51 | 74.4 | 3.29 | 90 | 206 | 3.85 | 2.78 |
[Table 3] | No. | [C] | [Mn] | [S] | Ceq | Lw | Lt | Vc | I | Time From Secondary Refining to Start of Casting | Amount of CaSi Added | Amount of FeSi Added | QAr | Ld | Actual Measured Value of CAR | Predicted Value of CAR |
| mass% | mass% | mass% | mass% | mm | min | m/min | A | min | kg/t | kg/t | NL/min | mm | % | % |
| 34 | 0.042 | 1.34 | 0.0004 | 0.099 | 2100 | 250 | 1.13 | 700 | 61 | 64.3 | 4.21 | 90 | 186 | 0.00 | 0.00 |
| 35 | 0.044 | 1.35 | 0.0004 | 0.101 | 2100 | 250 | 1.03 | 700 | 35 | 69.3 | 3.20 | 90 | 185 | 0.00 | 0.00 |
| 36 | 0.046 | 1.35 | 0.0003 | 0.104 | 2100 | 250 | 1.13 | 700 | 41 | 74.2 | 4.15 | 90 | 205 | 0.17 | 0.00 |
<Example 2>
-
An example in which the inverse analysis of the operating conditions is performed will be described with reference to test No. 21 in Table 2. Under the initial operating conditions of test No. 21, the predicted value and the actual measured value of the crack area ratio CAR of HIC in the surface layer portion of the product were 5.46% and 4.60%, respectively. As a result of performing the principal component analysis of Example 1, the immersion depth Ld of the immersion nozzle and the current I applied during electromagnetic agitation, both of which significantly influence the Crack Area Ratio (CAR) and are adjustable during operation, were extracted as the variables to be changed. Then, inverse analysis was performed by changing these variables using the prediction model to achieve a predicted CAR value of 0.2% and thus search for adequate operating conditions. Then, the immersion depth Ld of the immersion nozzle was changed from 186 mm to 210 mm as the obtained condition, and also, the current I applied during electromagnetic agitation was changed from 400 A to 700 A as the obtained condition. Consequently, it was possible to allow the actual measured value of the crack area ratio CAR of HIC in the surface layer portion of the product to attain a target of 2% or less.
Reference Signs List
-
- 1
- continuous casting machine
- 2
- tundish
- 3
- sliding nozzle
- 4
- immersion nozzle
- 5
- casting mold
- 6
- cast-slab support roll
- 7
- conveyance roll
- 8
- cast-slab cutter
- 9
- molten steel
- 10
- cast slab
- 10a
- cast slab (obtained by cutting)
- 11
- solidified shell
- 12
- molten steel in unsolidified solid phase
- 13
- solidification completion position (i.e., crater end)
- 14
- soft reduction zone
- FD
- casting direction