A converter smelting method based on a big data steelmaking model under a low iron water ratio condition

Abstract

The application discloses a kind of low iron water ratio conditions based on big data steelmaking model converter high-efficiency smelting method.Acquire converter smelting historical data;First model is obtained by training using historical data;Material data is obtained when actual converter smelting;Material data is input into first model, and first feeding data is obtained;According to first feeding data, feeding is carried out, and TSC measurement data is obtained after smelting stage is completed;Material data, first feeding data and TSC measurement data are input into first model, and second feeding data is obtained;After feeding according to second feeding data, continue to blow and refine, and TSO measurement data is obtained by inputting material data, first feeding data, TSC measurement data and second feeding data into first model;Whether TSO measurement data meets target requirement is judged;If it meets target requirement, then execute unequal sample tapping.The application scheme utilizes the advantage of big model to help steel smelting process, and production efficiency and product quality can be improved.

Description

Technical Field

[0001] This invention relates to the field of iron and steel smelting technology, and in particular to a converter smelting method based on a big data steelmaking model under low iron-to-metal ratio conditions. Background Technology

[0002] In the traditional heavy industry of steel smelting, the introduction of technologies such as network models and large-scale artificial intelligence models is driving a profound intelligent transformation. Steel production involves complex processes such as blast furnace ironmaking, converter steelmaking, and continuous casting and rolling, and its process optimization, energy consumption control, and quality management are extremely complex.

[0003] Currently, existing methods rely on empirical formulas and manual adjustments. Large-scale models, however, can integrate massive amounts of production data (such as temperature, pressure, and composition parameters) to construct high-precision digital twin systems, enabling dynamic simulation and prediction of the entire process. Therefore, how to leverage the advantages of large-scale models to assist the steel smelting process and improve production efficiency and product quality is a pressing technical problem that needs to be solved.

[0004] Currently, with the development of the steel industry, low-carbon production methods have emerged. Adding scrap steel during the smelting process helps reduce dependence on iron ore and lower carbon emissions. However, as the amount of scrap steel increases during smelting, the content of unknown elements in the steel increases, prolonging the smelting time. Furthermore, it makes existing steelmaking models unsuitable, causing significant challenges to current smelting processes. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a converter smelting method based on a big data steelmaking model under low iron-to-metal ratio conditions.

[0006] The technical solution of this invention is implemented as follows:

[0007] This invention provides a converter smelting method based on a big data steelmaking model under low iron-to-metal ratio conditions. The method includes:

[0008] Acquire historical data of converter smelting; the historical data includes material data, first charging data, TSC measurement data, second charging data, and TSO measurement data; wherein, the material data includes molten iron weight, molten iron temperature, molten iron Si content, molten iron Mn content, molten iron P content, molten iron S content, converter age, heavy scrap weight, light scrap weight, pig iron weight, ladle scrap weight, and ladle scrap temperature; the first charging data includes lime addition amount, lightly calcined magnesia ball addition amount, exothermic ferrosilicon addition amount, coal-based carbonizer addition amount, and oxygen supply before TSC measurement; the TSC measurement data includes TSC measurement temperature, TSC measured C content, and TSC measured P content; the second charging data includes oxygen supply after TSC measurement; the TSO measurement data includes TSO measurement temperature, TSO measured C content, TSO measured O content, and TSO measured P content;

[0009] An initial model is obtained, and the initial model is trained using the historical data to obtain a first model. The first model includes a first sub-model, a second sub-model, and a third sub-model. The input of the first sub-model is material data; the output of the first sub-model is first feeding data. The input of the second sub-model is material data, first feeding data, and TSC measurement data; the output of the second sub-model is second feeding data. The input of the third sub-model is material data, first feeding data, TSC measurement data, and second feeding data; the output of the third sub-model is TSO measurement data.

[0010] Obtain material data during actual converter smelting; input the material data into the first sub-model to obtain the first feeding data output by the first sub-model;

[0011] The material is added according to the first feeding data output by the first sub-model, and TSC measurement is performed after the smelting stage is completed to obtain TSC measurement data.

[0012] The material data, the first feeding data, and the TSC measurement data are input into the second sub-model to obtain the second feeding data output by the second sub-model;

[0013] After feeding according to the second feeding data output by the second sub-model, the blowing process continues, and the material data, the first feeding data, the TSC measurement data and the second feeding data are input into the third sub-model to obtain the TSO measurement data output by the third sub-model;

[0014] Determine whether the TSO measurement data output by the third sub-model meets the target requirements; if it does, then execute unequal sample steel output.

[0015] In one embodiment, if the TSO measurement data output by the third sub-model does not meet the target requirements, then the following is executed:

[0016] The TSO measurement temperature in the TSO measurement data is taken as the latest TSC measurement temperature, the C content measured by TSO is taken as the latest TSC measurement C content, the P content measured by TSO is taken as the latest TSC measurement P content, and the current oxygen supply is taken as the oxygen supply before the latest TSC measurement.

[0017] The updated data is input into the second sub-model to obtain the second feeding data output by the second sub-model again; after feeding according to the second feeding data output by the second sub-model, the blowing process continues, and the updated material data, the first feeding data, the TSC measurement data and the second feeding data are input into the third sub-model again to obtain the TSO measurement data output by the third sub-model;

[0018] The system then determines whether the TSO measurement data output by the third sub-model meets the target requirements. If it does, it performs unequal sample steel output. If it does not meet the target requirements, it repeats the above operation until the TSO measurement data output by the third sub-model meets the target requirements.

[0019] In one embodiment, the initial model is a gradient boosting classifier algorithm model.

[0020] In one embodiment, the parameters in the gradient boosting classifier algorithm model are in the following ranges: the number of weak classifiers is in the range of 800~1200, the learning rate is in the range of 0.01~0.02, the maximum depth of the decision tree is in the range of 4, the minimum number of samples for a decision tree split is in the range of 2, and the minimum number of samples for a decision tree leaf node is in the range of 1.

[0021] In one embodiment, the process of determining the parameter range in the gradient boosting classifier algorithm model is as follows:

[0022] Obtain the initial value range of each parameter of the gradient boosting classifier algorithm model;

[0023] The initial value range of each parameter is traversed using grid search or random search to form multiple parameter combinations;

[0024] Cross-validation is performed on each parameter combination to obtain model performance evaluation results;

[0025] The optimal parameter combination is determined based on the model performance evaluation results; the optimal parameter combination is then used as the final range of values ​​for the gradient boosting classifier algorithm model parameters.

[0026] In one embodiment, after smelting is completed, the slag composition is detected to determine whether the data for this furnace run has been updated to the converter smelting history data:

[0027] To determine whether the data for this heat run meets the basic requirements, the following basic requirements are defined: 120 t ≤ molten iron weight ≤ 200 t, 1000 ℃ ≤ molten iron temperature ≤ 1550 ℃, 0.1% ≤ molten iron Si content ≤ 0.9%, 0.05% ≤ molten iron Mn content ≤ 0.6%, 0.06% ≤ molten iron P content ≤ 0.13%, molten iron S content ≤ 0.1%, converter age ≤ 10000 heats, heavy scrap weight ≤ 60 t, light scrap weight ≤ 10 t, pig iron weight ≤ 70 t, 5 t ≤ ladle scrap weight ≤ 25 t, 0.5 t ≤ lime addition ≤ 13 t, lightly calcined magnesia ball addition ≤ 3 t, exothermic ferrosilicon addition ≤ 3 t, coal-based carbonizer addition ≤ 3 t, 6000 m³ ≤ oxygen supply before TSC measurement ≤ 12000 m³, 1520 ℃≤TSC measurement temperature≤1700℃, 0.005%≤C content measured by TSC≤0.5%, 0.002%≤P content measured by TSC≤0.1%, oxygen supply after TSC measurement≤6000 m³, 1540 ℃≤TSO measurement temperature≤1700 ℃, 0.02%≤C content measured by TSO≤0.08%, 0.03%≤O content measured by TSO≤0.13%, 0.002%≤P content measured by TSO≤0.08%;

[0028] If the data for this furnace batch meets the second requirement, provided that the basic requirements are met, then it is determined that the data for this furnace batch can be updated to the historical data of converter smelting.

[0029] The second requirement includes: steel tapping temperature -30℃ ≤ slag melting point ≤ steel tapping temperature -20℃, FeTOT ≤ 25%, 1% ≤ P2O5 ≤ 3%, CaO / SiO2 > CaO / SiO 2(中间值) ;

[0030] The process of determining whether the data for this batch meets the second requirement includes:

[0031] Based on the slag composition analysis results, the five-element phase diagram of CaO-SiO2-FeO-Fe2O3-MgO in this furnace was calculated using Factsage software. Here, MgO represents the slag composition analysis value, Fe2O3 is 0.25FeTOT, FeO is 0.75FeTOT, and FeTOT is the slag composition analysis value. The region in the five-element phase diagram where the slag melting point falls between -30℃ and -20℃ from the steel tapping temperature was identified. The range and median value of CaO / SiO2 within this region were calculated, and the median value was denoted as CaO / SiO2. 2(中间值) The melting point of the slag in this furnace is calculated using the aforementioned pentagonal phase diagram.

[0032] In one embodiment, obtaining an initial model and training the initial model using the historical data to obtain a first model includes:

[0033] Based on the steel grades in the historical data, the historical data is divided into multiple categories; the initial model is trained using the historical data of each category to obtain the first model corresponding to the historical data of each category.

[0034] The solution in this embodiment has the following beneficial effects:

[0035] This embodiment leverages the advantages of large-scale models to assist the steel smelting process, thereby improving production efficiency and product quality. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of the converter smelting method based on a big data steelmaking model under low iron ratio conditions according to an embodiment of the present invention.

[0037] Figure 2 This is a schematic diagram of multivariate linear programming in an embodiment of the present invention;

[0038] Figure 3 This is a schematic diagram of the CaO-SiO2-FeO-MgO-Fe2O3 pentagonal phase diagram obtained by Factsage calculation in an embodiment of the present invention;

[0039] Figure 4 This is a schematic diagram comparing the predicted and actual values ​​of metallurgical lime according to an embodiment of the present invention;

[0040] Figure 5 This is a schematic diagram comparing the predicted and actual values ​​of the lightly sintered magnesium balls in an embodiment of the present invention.

[0041] Figure 6 This is a schematic diagram comparing the predicted and actual values ​​of the coal-carbonizing agent in an embodiment of the present invention.

[0042] Figure 7 This is a schematic diagram comparing the predicted and actual values ​​of the TSC measured temperature in an embodiment of the present invention.

[0043] Figure 8 This is a schematic diagram comparing the predicted and actual values ​​of C content measured by TSC in an embodiment of the present invention;

[0044] Figure 9 This is a schematic diagram comparing the predicted and actual values ​​of P content measured by TSC in an embodiment of the present invention.

[0045] Figure 10 This is a schematic diagram comparing the predicted and actual oxygen supply values ​​after TSC measurement in an embodiment of the present invention.

[0046] Figure 11 This is a schematic diagram comparing the predicted and actual values ​​of the TSO measured temperature in an embodiment of the present invention.

[0047] Figure 12This is a schematic diagram comparing the predicted and actual values ​​of C content measured by TSO in an embodiment of the present invention;

[0048] Figure 13 This is a schematic diagram comparing the predicted and actual values ​​of O content measured by TSO in an embodiment of the present invention.

[0049] Figure 14 This is a schematic diagram comparing the predicted and actual values ​​of P content measured by TSO in an embodiment of the present invention. Detailed Implementation

[0050] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0051] This invention provides a converter smelting method based on a big data steelmaking model under low iron-to-metal ratio conditions, such as... Figure 1 As shown, the method includes:

[0052] Step 101: Obtain historical data of converter smelting; the historical data includes material data, first charging data, TSC measurement data, second charging data, and TSO measurement data; wherein, the material data includes molten iron weight, molten iron temperature, molten iron Si content, molten iron Mn content, molten iron P content, molten iron S content, converter age, heavy scrap weight, light scrap weight, pig iron weight, ladle scrap weight, and ladle scrap temperature; the first charging data includes lime addition amount, lightly calcined magnesia ball addition amount, exothermic ferrosilicon addition amount, coal-based carbonizer addition amount, and oxygen supply before TSC measurement; the TSC measurement data includes TSC measurement temperature, TSC measured C content, and TSC measured P content; the second charging data includes oxygen supply after TSC measurement; the TSO measurement data includes TSO measurement temperature, TSO measured C content, TSO measured O content, and TSO measured P content;

[0053] Step 102: Obtain an initial model, and train the initial model using the historical data to obtain a first model; the first model includes a first sub-model, a second sub-model, and a third sub-model; the input of the first sub-model is material data; the output of the first sub-model is first feeding data; the input of the second sub-model is material data, first feeding data, and TSC measurement data; the output of the second sub-model is second feeding data; the input of the third sub-model is material data, first feeding data, TSC measurement data, and second feeding data; the output of the third sub-model is TSO measurement data;

[0054] Step 103: Obtain material data during actual converter smelting; input the material data into the first sub-model to obtain the first feeding data output by the first sub-model;

[0055] Step 104: Add materials according to the first feeding data output by the first sub-model, and perform TSC measurement after the smelting stage is completed to obtain TSC measurement data;

[0056] Step 105: Input the material data, the first feeding data, and the TSC measurement data into the second sub-model to obtain the second feeding data output by the second sub-model;

[0057] Step 106: After feeding according to the second feeding data output by the second sub-model, continue blowing, and input the material data, the first feeding data, the TSC measurement data and the second feeding data into the third sub-model to obtain the TSO measurement data output by the third sub-model;

[0058] Step 107: Determine whether the TSO measurement data output by the third sub-model meets the target requirements; if it meets the target requirements, then execute unequal sample steel output.

[0059] In practical applications, if the TSO measurement data output by the third sub-model does not meet the target requirements, then the following steps will be executed:

[0060] The TSO measurement temperature in the TSO measurement data is taken as the latest TSC measurement temperature, the C content measured by TSO is taken as the latest TSC measurement C content, the P content measured by TSO is taken as the latest TSC measurement P content, and the current oxygen supply is taken as the oxygen supply before the latest TSC measurement.

[0061] The updated data is input into the second sub-model to obtain the second feeding data output by the second sub-model again; after feeding according to the second feeding data output by the second sub-model, the blowing process continues, and the updated material data, the first feeding data, the TSC measurement data and the second feeding data are input into the third sub-model again to obtain the TSO measurement data output by the third sub-model;

[0062] The system then determines whether the TSO measurement data output by the third sub-model meets the target requirements. If it does, it performs unequal sample steel output. If it does not meet the target requirements, it repeats the above operation until the TSO measurement data output by the third sub-model meets the target requirements.

[0063] TSC and TSO measurements are part of the auxiliary lance inspection process. Auxiliary lance inspection is an operational procedure to check whether the molten steel in the converter steelmaking process meets the required standards. TSC measurements performed by the auxiliary lance during converter steelmaking primarily measure the carbon content of the molten steel in the molten pool during the blowing process; TSO measurements performed by the auxiliary lance primarily measure the carbon and oxygen content of the molten pool at the end of the blowing process.

[0064] This embodiment leverages the advantages of large-scale models to assist the steel smelting process, thereby improving production efficiency and product quality.

[0065] The present solution will now be described using a specific embodiment.

[0066] Specifically, this solution includes the following steps:

[0067] S1: Collect valid historical data and establish a database. The data includes molten iron weight, molten iron temperature, molten iron Si content, molten iron Mn content, molten iron P content, molten iron S content, converter age, heavy scrap weight, light scrap weight, pig iron weight, ladle scrap weight, ladle scrap temperature, lime addition amount, light calcined magnesia ball addition amount, exothermic ferrosilicon addition amount, coalification carburizing agent addition amount, oxygen supply before TSC (i.e., oxygen supply before TSC measurement), TSC temperature (i.e., TSC measurement temperature), TSC C (i.e., C content measured by TSC), TSC P (i.e., P content measured by TSC), oxygen supply after TSC (i.e., oxygen supply after TSC measurement), TSO temperature (i.e., TSO measurement temperature), TSO C (i.e., C content measured by TSO), TSO O (i.e., O content measured by TSO), and TSO P (i.e., P content measured by TSO).

[0068] S2: Training database historical data. Using parameters such as molten iron weight, molten iron temperature, molten iron Si content, molten iron Mn content, molten iron P content, molten iron S content, converter age, heavy scrap weight, light scrap weight, pig iron weight, ladle scrap weight, ladle scrap temperature, and multivariate linear programming parameters such as lime addition, lightly calcined magnesia ball addition, exothermic ferrosilicon addition, coalification recarburizing agent addition, and TSC pre-oxygen supply, and then using parameters such as molten iron weight, molten iron temperature, molten iron Si content, molten iron Mn content, molten iron P content, molten iron S content, converter age, heavy scrap weight, light scrap weight, pig iron weight, ladle scrap weight, ladle scrap temperature, lime addition, lightly calcined magnesia ball addition, exothermic ferrosilicon addition, coalification recarburizing agent addition, and TSC pre-oxygen supply, multivariate linear programming TSC temperature, TSC C, and TSC. P, then using parameters such as molten iron weight, molten iron temperature, molten iron Si content, molten iron Mn content, molten iron P content, molten iron S content, converter age, heavy scrap weight, light scrap weight, pig iron weight, ladle scrap weight, ladle scrap temperature, lime addition, lightly calcined magnesia ball addition, exothermic ferrosilicon addition, coalification recarburizing agent addition, oxygen supply before TSC, TSC temperature, TSC C, TSC P, multivariate linear programming for oxygen supply after TSC. Finally, using parameters such as molten iron weight, molten iron temperature, molten iron Si content, molten iron Mn content, molten iron P content, molten iron S content, converter age, heavy scrap weight, light scrap weight, pig iron weight, ladle scrap weight, ladle scrap temperature, lime addition, lightly calcined magnesia ball addition, exothermic ferrosilicon addition, coalification recarburizing agent addition, oxygen supply before TSC, TSC temperature, TSC C, TSC P, oxygen supply after TSC, multivariate linear programming for TSO temperature, TSO C, TSO O, TSO P. The model is obtained after the data training is completed;

[0069] S3: Production Applications. See also... Figure 2The first step involves inputting real-time data on molten iron weight, temperature, Si, Mn, P, and S content into the model. A multivariate linear programming approach is then used to determine the amounts of lime, lightly calcined magnesia balls, exothermic ferrosilicon, coal-based carbonizer, and oxygen supply before TSC (Total Sterile Charge). The second step involves feeding materials according to the multivariate linear programming results. After this smelting stage is completed, temperature and composition measurements (TSC measurement) are performed to obtain the TSC temperature, TSC C, and TSC P. The third step involves inputting the following parameters into the model: molten iron weight, molten iron temperature, molten iron Si, molten iron Mn, molten iron P, molten iron S, converter age, heavy scrap weight, light scrap weight, pig iron weight, ladle scrap weight, ladle scrap temperature, lime addition amount, lightly calcined magnesia ball addition amount, exothermic ferrosilicon addition amount, coalification carburizing agent addition amount, oxygen supply before TSC, TSC temperature, TSC C, and TSC P. A multivariate linear programming method is then used to calculate the oxygen supply after TSC. The fourth step is to supply oxygen according to the calculation results and continue the blowing process. The fifth step involves using a multivariate linear programming approach to calculate the following parameters: molten iron weight, molten iron temperature, molten iron Si content, molten iron Mn content, molten iron P content, molten iron S content, converter age, heavy scrap weight, light scrap weight, pig iron weight, ladle scrap weight, ladle scrap temperature, lime, lightly calcined magnesia balls, exothermic ferrosilicon, coal-based carbon raiser, oxygen supply before TSC, TSC temperature, TSC C, TSC P, and oxygen supply after TSC. Based on the calculated results, steel is tapped only if the temperature and composition are within acceptable limits. If not, the third step is repeated. The oxygen supply before TSC is the cumulative oxygen supply value. After this stage of smelting, temperature, carbon, and oxygen (TSO measurement) are measured. To save smelting time, an unequal sample tapping method is adopted, allowing for direct tapping of the steel.

[0070] S4: If there are no abnormalities in the smelting and the data meets the preset requirements, the smelting data of this furnace will be imported into the database according to the category and used as historical data for periodic training.

[0071] The converter smelting data includes material data, first charge data, TSC measurement data, second charge data, TSO measurement data, and slag testing data; the slag testing data was obtained in the following way:

[0072] After smelting, the slag composition was analyzed, and the CaO-SiO2-FeO pentagonal phase diagram was calculated using Factsage software (MgO was the slag analysis value, and Fe2O3 was taken as 0.25FeTOT). (The CaO-SiO2-FeO pentagonal phase diagram can be found here.) Figure 3 As shown in the diagram, taking FeO = 0.75FeTOT, identify the region in the phase diagram where the slag melting point is between -30℃ and -20℃ from the steel tapping temperature. Calculate the range of CaO / SiO2 values ​​in this region and determine the median value, denoted as CaO / SiO2.2(中间值) The melting point of the slag in the current furnace batch is calculated using a phase diagram.

[0073] The preset requirements include basic requirements and expected transformation requirements.

[0074] The converter in this embodiment has a capacity of 210 t, and the basic requirements are as follows: 120 t ≤ molten iron weight ≤ 200 t, 1000 ℃ ≤ molten iron temperature ≤ 1550 ℃, molten iron 0.1% ≤ Si ≤ 0.9%, molten iron 0.05% ≤ Mn ≤ 0.6%, molten iron P ≤ 0.13%, molten iron S ≤ 0.1%, converter life ≤ 10000 heats, heavy scrap weight ≤ 60 t, light scrap weight ≤ 10 t, pig iron weight ≤ 70 t, 5 t ≤ ladle scrap weight ≤ 25 t, 0.5 t ≤ lime addition ≤ 13 t, lightly calcined magnesia ball addition ≤ 3 t, exothermic ferrosilicon addition ≤ 3 t, coal carbonizer addition ≤ 3 t, 6000 m³ ≤ oxygen supply before TSC ≤ 12000 m³, 1520 ℃≤TSC temperature≤1700℃, 0.005%≤TSC C≤0.5%, 0.002%≤TSC P≤0.1%, oxygen supply after TSC≤6000 m³, 1540 ℃≤TSO temperature≤1700 ℃, 0.02%≤TSO C≤0.08%, 0.03%≤TSO O≤0.13%, 0.002%≤TSO P≤0.08%.

[0075] The desired transformation requirements are: steel tapping temperature -30℃ ≤ slag melting point ≤ steel tapping temperature -20℃, FeTOT ≤ 25%, 1% ≤ P2O5 ≤ 3%, CaO / SiO2 > CaO / SiO 2(中间值) .

[0076] Furthermore, the function used in the multivariate linear programming is the Gradient Boosting Classifier (GBM) algorithm model. The range of values ​​for each parameter in the Gradient Boosting Classifier algorithm model can be optimized using the following steps;

[0077] First, determine the range of values ​​for each parameter based on experience and data characteristics; then, traverse the parameter space using either GridSearchCV or RandomizedSearchCV; next, perform cross-validation on each parameter combination to evaluate model performance; finally, select the optimal parameter combination based on the cross-validation results.

[0078] Considering the small size of the model dataset, to prevent overfitting, cross-validation was performed using 2000 sets of historical smelting data. The validation results showed that:

[0079] To control the randomness parameter and ensure the reproducibility of the experiment, random_state=46 is set. To maintain model complexity and reduce the risk of overfitting, the minimum number of samples for decision tree splits (min_samples_split) is set to 2. Similar to min_samples_split, to maintain model complexity and reduce the risk of overfitting, the minimum number of samples for leaf nodes (min_samples_leaf) is set to 1. The sample ratio (subsample) is set to 1, the number of weak classifiers (n_estimators) can be set to 800~1200, the learning rate (learning_rate) can be set to 0.01~0.02, and the maximum depth of the decision tree (max_depth) can be set to 4.

[0080] Additionally, see Figures 4-14 The results show a comparison between the model's predicted values ​​and the actual values ​​using the aforementioned parameter range. The comparison indicates that, using the above parameter range, the model's prediction accuracy can reach over 90%, effectively predicting the converter smelting process under low iron-to-metal ratio conditions.

[0081] Furthermore, in this embodiment, steel grades are classified by grade, and the data for each category is stored separately and linearly programmed.

[0082] This embodiment leverages the advantages of large-scale models to facilitate the steel smelting process, thereby improving production efficiency and product quality, and enabling unequal steel output.

[0083] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0084] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A low iron water ratio based on big data steelmaking model of converter smelting method, characterized in that, The method comprises: obtaining converter smelting historical data; the historical data comprises material data, first charging data, TSC measurement data, second charging data and TSO measurement data; wherein the material data comprises molten iron weight, molten iron temperature, molten iron Si content, molten iron Mn content, molten iron P content, molten iron S content, converter campaign, heavy waste weight, light waste weight, pig iron weight, ladle waste steel weight and ladle waste steel temperature; the first charging data comprises lime addition amount, light-magnesia ball addition amount, hot-rolled ferrosilicon addition amount, coalification carbon additive addition amount and oxygen supply amount before TSC measurement; the TSC measurement data comprises TSC measurement temperature, TSC measured C content and TSC measured P content; the second charging data comprises oxygen supply amount after TSC measurement; and the TSO measurement data comprises TSO measurement temperature, TSO measured C content, TSO measured O content and TSO measured P content; obtaining an initial model, training the initial model by using the historical data to obtain a first model; the first model comprises a first sub-model, a second sub-model and a third sub-model; obtaining material data during actual converter smelting; inputting the material data into the first sub-model to obtain first charging data output by the first sub-model; charging according to the first charging data output by the first sub-model, and performing TSC measurement after the smelting stage is completed to obtain TSC measurement data; inputting the material data, the first charging data and the TSC measurement data into the second sub-model to obtain second charging data output by the second sub-model; continuing to blow after charging according to the second charging data output by the second sub-model, inputting the material data, the first charging data, the TSC measurement data and the second charging data into the third sub-model to obtain TSO measurement data output by the third sub-model; judging whether the TSO measurement data output by the third sub-model meets target requirements; if the TSO measurement data meets the target requirements, performing unequal sample tapping; wherein after smelting is completed, detecting slag composition, and judging whether the current data is updated into the converter smelting historical data: determine whether the current furnace data meets the basic requirements, the basic requirements are: 120 t≤ molten iron weight≤ 200 t, 1000℃≤ molten iron temperature≤ 1550℃, 0.1%≤ molten iron Si content≤ 0.9%, 0.05%≤ molten iron Mn content≤ 0.6%, 0.06%≤ molten iron P content≤ 0.13%, molten iron S content≤ 0.1%, converter campaign≤ 10000 furnace, heavy waste weight≤ 60 t, light waste weight≤ 10 t, pig iron weight≤ 70 t, 5 t≤ iron ladle scrap weight≤ 25 t, 0.5 t≤ lime addition≤ 13 t, light burned magnesium ball addition≤ 3 t, exothermic ferrosilicon addition≤ 3 t, coalification carbon additive addition≤ 3 t, 6000 m³≤ oxygen supply amount before TSC measurement≤ 12000 m³, 1520℃≤ TSC measurement temperature≤ 1700℃, 0.005%≤ TSC measured C content≤ 0.5%, 0.002%≤ TSC measured P content≤ 0.1%, oxygen supply amount after TSC measurement≤ 6000 m³, 1540℃≤ TSO measurement temperature≤ 1700℃, 0.02%≤ TSO measured C content≤ 0.08%, 0.03%≤ TSO measured O content≤ 0.13%, 0.002%≤ TSO measured P content≤ 0.08%; In the case of meeting the basic requirements, if the current furnace data meets the expected conversion requirements, it is determined that the current furnace data can be updated to the converter smelting historical data; Wherein, the desired transformation requirements include: steel tapping temperature - 30℃ ≤ Slag melting point ≤ steel tapping temperature - 20℃, FeTOT ≤ 25%, 1% ≤ P2O5 ≤ 3%, CaO / SiO2 > CaO / SiO2 2(中间值) ; The process of determining whether the current furnace data meets the expected conversion requirements includes: According to the slag composition detection result, the Factsage software is used to calculate the CaO-SiO2-FeO-Fe2O3-MgO quinary phase diagram of the present furnace slag; wherein, MgO is the slag composition detection value, Fe2O3 is 0.25FeTOT, FeO is 0.75FeTOT, FeTOT is the slag composition detection value, the region of the slag melting point in the quinary phase diagram between the steel tapping temperature-30℃ to the steel tapping temperature-20℃ is found out, the CaO / SiO2 value range and the intermediate value of the region are calculated, the intermediate value is recorded as CaO / SiO2 2(中间值) ; and the quinary phase diagram is used to calculate the present furnace slag melting point.

2. The low iron water ratio based on big data steelmaking model of the converter smelting method according to claim 1, characterized in that, If the TSO measurement data output by the third sub-model does not meet the target requirements, the following operations are performed: Take the TSO measurement temperature in the TSO measurement data as the latest TSC measurement temperature, the TSO measured C content as the latest TSC measured C content, and the TSO measured P content as the latest TSC measured P content, and take the current oxygen supply amount as the latest TSC measurement oxygen supply amount before measurement; Input the updated data into the second sub-model to obtain second charging data output by the second sub-model again; continue blowing after charging according to the second charging data output by the second sub-model again, and input the updated material data, the first charging data, the TSC measurement data and the second charging data into the third sub-model again to obtain the TSO measurement data output by the third sub-model; Determine again whether the TSO measurement data output by the third sub-model meets the target requirements; if it meets the target requirements, perform tapping without sampling; if it does not meet the target requirements, perform the above operations again until the TSO measurement data output by the third sub-model meets the target requirements.

3. The low iron water ratio based on big data steelmaking model for converter smelting method according to claim 1, characterized in that, The initial model is a gradient boosting classifier algorithm model.

4. The low iron water ratio based on big data steelmaking model of the converter smelting method according to claim 3, characterized in that, The range of parameters in the gradient boosting classifier algorithm model is: the range of the number of weak classifiers is 800-1200, the range of the learning rate is 0.01-0.02, the range of the maximum depth of the decision tree is 4, the range of the minimum number of samples for splitting the decision tree is 2, and the range of the minimum number of samples for the leaf node of the decision tree is 1.

5. The low iron water ratio based on big data steelmaking model for converter smelting method according to claim 4, characterized in that, The parameter range determination process in the gradient boosting classifier algorithm model is as follows: An initial value range of each parameter of the gradient boosting classifier algorithm model is obtained. A grid search or a random search is used to traverse the initial value range of each parameter to form a plurality of parameter combinations. Cross-validation is performed on each parameter combination to obtain a model performance evaluation result. An optimal parameter combination is determined according to the model performance evaluation result, and the optimal parameter combination is taken as a final value range of the parameters of the gradient boosting classifier algorithm model.

6. The low iron water ratio based on big data steelmaking model for converter smelting method according to claim 1, characterized in that, An initial model is obtained, and the initial model is trained using the historical data to obtain a first model, including: The historical data is divided into a plurality of categories according to the grade of the steel grade in the historical data; and the initial model is trained using the historical data of each category to obtain a first model corresponding to the historical data of each category.

Images