Converter deoxidation method based on sublance oxygen determination

Abstract

The invention relates to a converter deoxidation method based on sublance oxygen determination, and belongs to the technical field of metallurgical methods. According to the technical scheme, the method comprises the following steps: analyzing historical deoxidation data to obtain a general relationship between the converter terminal sublance oxygen determination amount and the deoxidizing agent adding amount; establishing a mathematical model by using a grey prediction model theory; using a statistical test method to test whether the model is good or not; and storing the data of the heat which is predicted hit by the model into a deoxidation database as the data of the heat after prediction, and eliminating the heat which is not predicted by the model so as to achieve the purpose of self-learning. The method has the beneficial effects that the waste of a deoxidizing agent is reduced, the smelting cost is reduced, the argon station entering oxygen activity is effectively controlled, the refining deoxidizing burden is reduced, the refining treatment time is shortened, and the gas content of molten steel is reduced; molten steel with good fluidity is provided for continuous casting, and the casting blank quality is improved.

Description

Technical Field

[0001] This invention relates to a converter deoxidation method based on secondary lance oxygen control, belonging to the field of metallurgical method technology. Background Technology

[0002] Currently, the main strong deoxidizer used in Chengde Steel's converter steelmaking is an aluminum-based deoxidizer. The primary method for determining the dosage is for workers to assess the oxidizing properties of the molten steel at the blowing endpoint and then, based on experience, estimate the required amount of deoxidizer. This results in a lack of strict standards for the amount of deoxidizer added, relying solely on practical experience. However, this experience often leads to either too much or too little deoxidizer being added, resulting in waste, increased steelmaking costs, and frequent process accidents. Adding too much deoxidizer prevents proper flow from the ladle, while adding too little increases the deoxidation burden on the LF (forging ladle) and increases inclusions in the molten steel. Summary of the Invention

[0003] The purpose of this invention is to provide a converter deoxidation method based on auxiliary lance oxygen control. By applying a grey prediction model and a self-learning model, it reduces deoxidizer waste and lowers smelting costs. Under the same endpoint conditions, the oxygen activity at the argon station can be controlled within the key process parameters, exceeding 95%. It provides high-quality molten steel for the refining process, reduces the deoxidation burden, facilitates rapid white slag formation and desulfurization in refining, reduces refining time, and consequently lowers the gas content in the molten steel. It also provides molten steel with good fluidity for continuous casting, reduces the phenomenon of ladle not flowing due to high aluminum content in the molten steel, improves billet quality, and effectively solves the aforementioned problems existing in the background technology.

[0004] The technical solution of this invention is: a converter deoxidation method based on auxiliary lance oxygen control, comprising the following steps: (1) The approximate relationship between the final oxygen content of the auxiliary lance and the amount of deoxidizer added was obtained by analyzing historical deoxidation data; (2) A mathematical model was established using the grey prediction model theory; (3) Use statistical testing methods to test the goodness of the model; (4) In actual production, the data of the furnaces that the model predicts hits are saved to the deoxidation database as data for predicting future furnaces. The furnaces that the model predicts inaccurately are removed. In this way, the self-learning purpose is achieved, and finally, accurate deoxidation is realized.

[0005] In step (1), the historical deoxidation data includes steel grade, amount of iron added, aluminum content of deoxidizer, temperature drop, target oxygen activity and final temperature; the oxygen is determined by TSO in the auxiliary lance at the blowing end.

[0006] In step (2), the main method of grey prediction theory is the least squares method. The formula fitted by the mathematical model is: y=alnx-b, where y represents the activity of molten steel after deoxidation and x represents the oxygen activity of molten steel at the end point.

[0007] In step (3), the statistical tests are the T-test and the F-test.

[0008] In step (4), the deoxygenation database is a dynamic database, meaning that new data added to the database serves as the basis for future predictions. This deoxygenation database constitutes the core of self-learning.

[0009] The beneficial effects of this invention are as follows: by applying the grey prediction model and the self-learning model, deoxidizer waste is reduced and smelting costs are lowered; under the same endpoint conditions, the oxygen activity at the argon station can be controlled within the key process points of over 95%; high-quality molten steel is provided for the refining process, reducing the deoxidation burden of refining, providing conditions for rapid white slag formation and rapid desulfurization in refining, reducing refining processing time, and thus reducing the gas content of molten steel; and molten steel with good fluidity is provided for continuous casting, reducing the phenomenon of ladle not flowing due to high aluminum content in molten steel, and improving the quality of cast billets. Attached Figure Description

[0010] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0011] To make the purpose, technical solutions, and advantages of the invention's embodiments clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only a small part of the embodiments of the present invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the protection scope of the present invention.

[0012] A converter deoxidation method based on auxiliary lance oxygen control includes the following steps: (1) The approximate relationship between the final oxygen content of the auxiliary lance and the amount of deoxidizer added was obtained by analyzing historical deoxidation data; (2) A mathematical model was established using the grey prediction model theory; (3) Use statistical testing methods to test the goodness of the model; (4) In actual production, the data of the furnaces that the model predicts hits are saved to the deoxidation database as data for predicting future furnaces. The furnaces that the model predicts inaccurately are removed. In this way, the self-learning purpose is achieved, and finally, accurate deoxidation is realized.

[0013] In step (1), the historical deoxidation data includes steel grade, amount of iron added, aluminum content of deoxidizer, temperature drop, target oxygen activity and final temperature; the oxygen is determined by TSO in the auxiliary lance at the blowing end.

[0014] In step (2), the main method of grey prediction theory is the least squares method. The formula fitted by the mathematical model is: y=alnx-b, where y represents the activity of molten steel after deoxidation and x represents the oxygen activity of molten steel at the end point.

[0015] In step (3), the statistical tests are the T-test and the F-test.

[0016] In step (4), the deoxygenation database is a dynamic database, meaning that new data added to the database serves as the basis for future predictions. This deoxygenation database constitutes the core of self-learning.

[0017] In practical applications, this invention overcomes the adverse consequences of manually adding deoxidizer based on experience. Considering the significant fluctuations in molten steel conditions, this invention deeply analyzes the converter steelmaking metallurgical process and its thermodynamic and kinetic characteristics. Based on mass and heat transfer theories, and through analysis of extensive historical deoxidation data, the approximate relationship between the final oxygen content of the auxiliary lance and the amount of deoxidizer added can be derived. Then, a mathematical model combining theory and experience is established using grey prediction model theory. The model's performance is then tested using t-tests or F-tests based on statistical theory. In actual production, the model saves the predicted heats to a database for future predictions, while incorrect predictions are discarded. This process is repeated to achieve self-learning, ultimately realizing precise deoxidation.

[0018] The model consists of two parts: a prediction module and a self-learning module. The prediction module primarily inputs the oxygen content determined at the converter endpoint using a secondary lance TSO into the model, which then calculates the required amount of deoxidizer. The self-learning module automatically saves suitable deoxidation cycles to the database.

[0019] The application of this model aims to transform the manual, experience-based addition of deoxidizer into precise calculation of the amount to be added, thereby maximizing the automation of steelmaking deoxidation and achieving optimal deoxidation. The implementation of this model reduces the impact of human factors on deoxidation and promotes the standardization of steelmaking deoxidation and alloying processes.

[0020] The formula fitted by this model is: y=alnx-b (y represents the activity of molten steel after deoxidation, and x represents the final oxygen activity of molten steel). This model can learn on its own, mainly by adjusting the values ​​of parameters a and b based on historical heats.

[0021] Table 1 Deoxygenation data before model application

[0022] Deoxygenation was poor before using the model; the oxygen activity at the argon station ranged from a minimum of 5 ppm to a maximum of 110 ppm.

[0023] Table 2 Deoxygenation data after using the model

[0024] After using the model, deoxygenation was very effective, with the oxygen activity at the argon station ranging from a minimum of 6 ppm to a maximum of 36 ppm, which met production requirements.

[0025] The specific operating procedure of this invention is as follows: (1) First, manually input the parameters required by the model. For example: steel grade, amount of iron added, aluminum content of deoxidizer, temperature drop, target oxygen activity, endpoint temperature, etc.

[0026] (2) Use a secondary gun to determine oxygen at the end of the blowing process.

[0027] (3) Input the oxygen determination results into the “endpoint oxygen activity” column in the model, and the model will automatically calculate the amount of deoxygenation to be added.

[0028] (4) Vibrate the material from the silo according to the calculation results.

[0029] (5) During the tapping process, deoxidizer is added to the ladle in accordance with the process requirements.

[0030] (6) After the steel is tapped, the molten steel is blown into the argon blowing station to blow argon gas. After the molten steel is uniform, the oxygen is determined by the oxygen lance.

[0031] (7) If the oxygen determination result matches the prediction result, the data of this furnace will be stored in the database; otherwise, it will be removed.

[0032] Compared with manually adding deoxidizer based on experience, the beneficial effects of this invention are: (1) Reduce deoxidizer waste and lower smelting costs.

[0033] (2) Under the same endpoint conditions, the oxygen activity at the argon station can be controlled within the key process points of more than 95%.

[0034] (3) Provide high-quality molten steel for the refining process, reduce the deoxidation burden of refining, provide conditions for rapid white slag formation and rapid desulfurization in refining, reduce refining processing time, and thus reduce the gas content of molten steel.

[0035] (4) It provides molten steel with good fluidity for continuous casting, reduces the phenomenon of ladle not flowing down due to high aluminum content in molten steel, and improves the quality of cast billets.

Claims

1. A converter deoxidation method based on secondary lance oxygen control, characterized in that... Includes the following steps: (1) The approximate relationship between the final oxygen content of the auxiliary lance and the amount of deoxidizer added was obtained by analyzing historical deoxidation data; (2) A mathematical model was established using the grey prediction model theory; (3) Use statistical testing methods to test the goodness of the model; (4) In actual production, the data of the furnaces that the model predicts hits are saved to the deoxidation database as data for predicting future furnaces. The furnaces that the model predicts inaccurately are removed. In this way, the self-learning purpose is achieved, and finally, accurate deoxidation is realized.

2. The converter deoxidation method based on auxiliary lance oxygen control according to claim 1, characterized in that: In step (1), the historical deoxidation data includes steel grade, amount of iron added, aluminum content of deoxidizer, temperature drop, target oxygen activity and final temperature; the oxygen is determined by TSO in the auxiliary lance at the blowing end.

3. The converter deoxidation method based on auxiliary lance oxygen control according to claim 1, characterized in that: In step (2), the main method of grey prediction theory is the least squares method. The formula fitted by the mathematical model is: y=alnx-b, where y represents the activity of molten steel after deoxidation and x represents the oxygen activity of molten steel at the end point.

4. The converter deoxidation method based on secondary lance oxygen control according to claim 1, characterized in that: In step (3), the statistical tests are the T-test and the F-test.

5. The converter deoxidation method based on auxiliary lance oxygen control according to claim 1, characterized in that: In step (4), the deoxygenation database is a dynamic database, meaning that new data added to the database serves as the basis for future predictions. This deoxygenation database constitutes the core of self-learning.

Images