An arc-based method and system for adjusting argon slag layer of molten steel

Abstract

The present application relates to the technical field of steel smelting, and particularly relates to a method and system for adjusting argon slag layer of molten steel based on electric arc heating, which comprises the following steps: constructing an error identification model based on a binary classification model, judging whether there is an accuracy error influence in the electrode voltage measurement process by using error source parameters of electrode voltage in the process of electric arc heating of molten steel, and then identifying influence factors based on error analysis; constructing three relationship models between electrode voltage, argon gas flow and slag layer state based on historical data; predicting the demand of argon gas flow and slag layer state by using the relationship models between three electrode voltages, argon gas flow and slag layer state, in combination with the identification condition of influence factors, according to the real-time monitored voltage data; generating and executing the adjustment strategy of argon gas flow and slag layer state according to the predicted demand of argon gas flow and slag layer state; and the present application realizes the accurate prediction and adjustment of argon gas flow and slag layer state, thereby improving the production efficiency and reducing the production cost.

Description

Technical Field

[0001] This invention relates to the field of refining furnace temperature control technology, and in particular to a method and system for adjusting the argon slag layer of molten steel based on electric arc heating. Background Technology

[0002] With the intensification of competition in the steel industry, enterprises have a strong demand to improve the production efficiency and quality of molten steel in the process of electric arc heating. Electrode voltage is an important parameter in the process of electric arc heating of molten steel. Its accurate measurement is crucial for understanding the arc state, energy input and controlling the heating process. Furthermore, the supply and control of argon gas and slag layer are particularly important for electric arc heating.

[0003] Existing technologies aim to improve steel quality and reduce the need for frequent electrode adjustments during refining due to unsuitable argon gas and slag layers in the ladle. This results in multiple energizations, severely impacting steel quality and contaminating the molten steel. Furthermore, these technologies fail to adequately consider the influence of factors such as electrode corrosion, environmental factors, and slag layer interference on electrode voltage measurement. Even when these factors are considered, a lack of systematic and precise error identification and quantification methods leads to inaccurate voltage measurements, affecting subsequent control decisions based on voltage data. Previously, adjusting the argon-slag layer required operators to visually assess the argon gas level before energizing, but the state of the argon gas inside the ladle could not be observed after energizing. Relying solely on simple experience or limited data lacks in-depth research and modeling of the complex relationship between electrode voltage, argon gas flow, and slag layer state. This makes it impossible to accurately predict the required argon gas flow and slag layer state based on real-time voltage data, resulting in highly arbitrary adjustment strategies and hindering efficient and precise argon-slag layer control.

[0004] Existing technologies require frequent electrode lifting for confirmation, resulting in multiple power applications; or they lack systematic and accurate error identification and quantification methods, affecting subsequent control decisions based on voltage data; when adjusting the argon slag layer, they rely only on simple experience or limited data, lacking in-depth research and modeling of the complex relationship between electrode voltage, argon gas flow, and slag layer state. Summary of the Invention

[0005] This invention provides a method and system for adjusting the argon gas and slag layer in molten steel based on electric arc heating. During the refining furnace power supply process, the changes in the three-phase electrode voltage are used to judge the argon gas and slag layer in the ladle, thereby scientifically and rationally adjusting the argon gas and slag layer. This eliminates the need for frequent electrode lifting and blind adjustment due to unsuitable slag layer and argon gas, and solves the problems of multiple electric heating of molten steel, which pollutes the molten steel and wastes refining power consumption.

[0006] To achieve the above objectives, the present invention employs the following technical solution:

[0007] A method for adjusting the argon slag layer in molten steel heated by electric arc includes the following steps:

[0008] S1. Identify the error source parameters of the electrode voltage during the electric arc heating of molten steel. The error source parameters include electrode corrosion parameters, environmental parameters, and slag layer interference parameters.

[0009] S2. Based on the binary classification model, an error identification model is constructed. The error source parameters of electrode voltage during the electric arc heating of molten steel are used to determine whether there is an accuracy error in the electrode voltage measurement process.

[0010] S3. When it is determined that there is an accuracy error in the electrode voltage measurement process, the influencing factors are identified based on error analysis. The influencing factors are used to quantify the degree of influence of the error on the electrode voltage measurement.

[0011] S4. Monitor the voltage changes of the three electrodes of the LF furnace in real time, and build a model of the relationship between the voltage of the three electrodes and the argon gas flow and slag layer state based on historical data.

[0012] S5. Based on the real-time monitored voltage data and the identification of influencing factors, the demand for argon gas flow and slag layer status is predicted using the relationship model between the three electrode voltages and the argon gas flow and slag layer status.

[0013] S6. Generate and execute adjustment strategies for argon gas flow and slag layer state based on the predicted argon gas flow and slag layer state requirements.

[0014] Furthermore, step S2 specifically includes: using a binary classification model as the basis for the error identification model, using the error source parameters of the electrode voltage as the input of the error identification model, and outputting the probability of the existence of accuracy error influence.

[0015] The error identification model is trained using historical records of electric arc heating of molten steel to confirm the model coefficients. Then, the error source parameters of the electrode voltage acquired in real-time during the electric arc heating process are input into the error identification model. The error identification model is as follows:

[0016]

[0017] Where P(y=PrecisionErrors) represents the probability of the presence of precision error, α1, α2, and α3 represent the characteristic indicators of electrode corrosion parameters, environmental parameters, and slag layer interference parameters, respectively; β1, β2, and β3 represent the coefficients of the error identification model, respectively.

[0018] The probability of the presence of accuracy error based on the error identification model is compared with the error identification threshold. If the probability of the presence of accuracy error is greater than or equal to the error identification threshold, it is determined that there is accuracy error in the electrode voltage measurement process; if the probability of the presence of accuracy error is less than the error identification threshold, it is determined that there is no accuracy error in the electrode voltage measurement process.

[0019] Furthermore, the specific methods for obtaining the characteristic indicators of the electrode corrosion parameters, environmental parameters, and slag layer interference parameters are as follows:

[0020] The electrode corrosion parameter is specifically the change in electrode diameter; the environmental parameters include electromagnetic interference intensity and ambient humidity; and the slag layer interference parameter is specifically the slag layer conductivity.

[0021] The change in electrode diameter is recorded as a characteristic index of electrode corrosion parameters.

[0022] The electromagnetic interference intensity and ambient humidity are normalized, and the sum of the normalized electromagnetic interference intensity and ambient humidity is recorded as the characteristic index of the environmental parameter.

[0023] The electrical conductivity of the slag layer is recorded as a characteristic index of the slag layer interference parameters.

[0024] Furthermore, the specific analysis of influencing factors based on error analysis is as follows: for electrode corrosion parameters, a linear relationship is fitted between the change in electrode diameter and the voltage measurement error:

[0025] ΔV corrosion =ε1*α1

[0026] Where, ΔV corrosion This represents the voltage measurement error caused by electrode corrosion parameters, and ε1 represents the linearity coefficient of the fit.

[0027] For environmental parameters, a linear relationship is fitted between the characteristic indices of the environmental parameters and the voltage measurement error:

[0028] ΔV env =ε2*α2

[0029] Where, ΔV env This represents the voltage measurement error caused by environmental parameters, and ε2 represents the linearity coefficient of the fit.

[0030] For the slag layer interference parameters, a linear relationship is fitted between the slag layer conductivity and the voltage measurement error:

[0031] ΔV slag =ε3*α3

[0032] Where, ΔV slagThis represents the voltage measurement error caused by interference parameters from the slag layer, and ε3 represents the linearity coefficient of the fit.

[0033] Weights are assigned based on the degree of influence of each error source, and then the influence factors are obtained by combining the voltage measurement errors caused by electrode corrosion parameters, voltage measurement errors caused by environmental parameters, and voltage measurement errors caused by slag layer interference parameters:

[0034] IF=κ1*ΔV corrosion +κ2*ΔV env +κ3*ΔV slag

[0035] Where IF represents the impact factor, and κ1, κ2, and κ3 represent the weight values ​​of each error source.

[0036] Furthermore, the model constructing the relationship between the three electrode voltages and the argon gas flow and slag layer state based on historical data is specifically as follows:

[0037] Collect voltage data and corresponding argon flow rate and slag layer thickness at various historical time points. The voltage data includes the voltage values ​​of the three electrodes of the LF furnace.

[0038] The voltage data and corresponding argon flow rate and slag layer thickness at each historical time point are processed, and all data are mapped to a range of 0 to 1.

[0039] Based on a multinomial regression model, the voltage data at each historical time point were fitted with the argon gas flow rate or the voltage data at each historical time point with the slag layer thickness, respectively, to obtain the relationship model between the three electrode voltages and the argon gas flow rate or the three electrode voltages and the slag layer state.

[0040] Furthermore, step S5 specifically includes:

[0041] Real-time voltage data is obtained through real-time monitoring of production records. The real-time voltage data includes the real-time voltage values ​​of the three electrodes of the LF furnace.

[0042] The system identifies the impact of accuracy errors during electrode voltage measurement. When it is determined that there are no accuracy errors during electrode voltage measurement, the real-time voltage values ​​of the three electrodes of the LF furnace are input into the relationship model between the three electrode voltages and the argon gas flow, and the relationship model between the three electrode voltages and the slag layer state, respectively. The expected values ​​of argon gas flow and slag layer thickness are then output.

[0043] When it is determined that there is an accuracy error in the electrode voltage measurement process, the influencing factor is introduced into the relationship model between the three electrode voltages and the argon gas flow, and the relationship model between the three electrode voltages and the slag layer state. The optimized relationship models between the three electrode voltages and the argon gas flow, and the optimized relationship models between the three electrode voltages and the slag layer state are updated to obtain the expected values ​​of the argon gas flow and the slag layer thickness. The real-time voltage values ​​of the three electrodes of the LF furnace are respectively input into the optimized relationship models between the three electrode voltages and the argon gas flow, and the optimized relationship models between the three electrode voltages and the slag layer state, and the expected values ​​of the argon gas flow and the slag layer thickness are output.

[0044] Furthermore, step S6 specifically includes:

[0045] Receive the expected value of argon gas flow and the expected value of slag layer thickness, and identify the slag layer thickness and argon gas flow input in the real-time LF furnace;

[0046] Based on the expected values ​​of argon gas flow and slag layer thickness, the argon gas and slag layer input materials in the LF furnace are controlled. The control includes, but is not limited to, increasing the argon gas supply and decreasing the slag layer input material supply, or decreasing the argon gas supply and increasing the slag layer input material supply.

[0047] An adjustment system for an argon-slag layer in molten steel heated by electric arc includes:

[0048] Error judgment module: Based on a binary classification model, an error identification model is constructed. The error source parameters of electrode voltage during the electric arc heating of molten steel are used to determine whether there is an accuracy error in the electrode voltage measurement process.

[0049] Influence Factor Acquisition Module: When it is determined that there is an accuracy error in the electrode voltage measurement process, the module identifies the influence factors based on error analysis. The influence factors are used to quantify the degree of influence of the error on the electrode voltage measurement.

[0050] Relationship model building module: Real-time monitoring of voltage changes at the three electrodes of the LF furnace, and construction of a relationship model between the voltage of the three electrodes and the argon gas flow and slag layer state based on historical data;

[0051] Argon gas slag layer state identification module: Based on real-time monitored voltage data and the identification status of influencing factors, the module uses a model of the relationship between the three electrode voltages and the argon gas flow and slag layer state to predict the demand for argon gas flow and slag layer state.

[0052] Adjustment Execution Module: Generates and executes adjustment strategies for argon gas flow and slag layer state based on the predicted argon gas flow and slag layer state requirements.

[0053] Compared with the prior art, the beneficial effects of the present invention are:

[0054] By identifying error source parameters and constructing an error identification model, it is possible to accurately determine whether there are accuracy errors affecting the electrode voltage measurement process. When such errors exist, the influencing factors are further identified to quantify the degree of error impact, thereby effectively assessing and correcting measurement errors, improving the accuracy of voltage measurement, and providing a reliable basis for subsequent voltage data-based analysis and control. Real-time monitoring of electrode voltage changes, combined with historical data to construct a relationship model with argon gas flow and slag layer state, can accurately predict the demand for argon gas flow and slag layer state. Based on these accurate predictions, adjustment strategies are generated and executed, enabling precise control of the argon-slag layer, optimizing the process of electric arc heating molten steel, and improving steel quality and production efficiency. Considering multiple error source parameters, the system can adapt to electrode voltage measurements under different working conditions and complex environments. Through error processing and accurate prediction and adjustment, system fluctuations caused by errors are reduced, enhancing the stability and reliability of the entire electric arc heating molten steel system and reducing uncertainties in the production process. Attached Figure Description

[0055] Figure 1 This is a flowchart illustrating the method described in this invention.

[0056] Figure 2 This is a schematic diagram of the system described in this invention. Detailed Implementation

[0057] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings:

[0058] This invention discloses a method for adjusting the argon-slag layer in molten steel heated by electric arc. It comprehensively identifies error sources affecting electrode voltage measurement during the electric arc heating process, including electrode corrosion parameters, environmental parameters, and slag layer interference parameters. A binary classification model is used to construct an error identification model to determine if accuracy errors exist. When an impact is found, error analysis identifies influencing factors and quantifies the degree of error's influence on voltage measurement, providing a basis for subsequent data correction. The method monitors the voltage changes of the three electrodes in the LF furnace in real time and collects a large amount of historical data. Through data processing and analysis, a relationship model is constructed between the three electrode voltages and the argon gas flow and slag layer state. This model reflects the intrinsic connection between electrode voltage and the argon-slag layer, providing a foundation for predicting the argon gas flow and slag layer state requirements. Based on the real-time monitored voltage data and the identification of influencing factors, the voltage data is corrected or the relationship model is directly used to predict the argon gas flow and slag layer state requirements, achieving precise adjustment of the argon-slag layer, optimizing the electric arc heating process, and improving production efficiency and steel quality.

[0059] See Figure 1 This is a flowchart illustrating the method described in this invention. Specifically, it includes the following steps:

[0060] S1. Identify the error source parameters of electrode voltage during the electric arc heating of molten steel. The error source parameters include electrode corrosion parameters, environmental parameters, and slag layer interference parameters.

[0061] S2. Based on the binary classification model, an error identification model is constructed. The error source parameters of electrode voltage during the electric arc heating of molten steel are used to determine whether there is an accuracy error in the electrode voltage measurement process.

[0062] An error identification model is constructed based on a binary classification model. The specific analysis of determining whether there is an accuracy error affecting the electrode voltage measurement process using the error source parameters of the electrode voltage during electric arc heating of molten steel is as follows: The binary classification model is used as the basis for the error identification model, the error source parameters of the electrode voltage are used as the input of the error identification model, and the output of the error identification model is the probability of the existence of an accuracy error. The error identification model is trained using historical records of electric arc heating of molten steel to confirm the coefficients of the error identification model. Then, the error source parameters of the electrode voltage during the real-time electric arc heating of molten steel are input into the error identification model. The specific error identification model is as follows:

[0063]

[0064] Where P(y=PrecisionErrors) represents the probability of the presence of precision error, α1, α2, and α3 represent the characteristic indicators of electrode corrosion parameters, environmental parameters, and slag layer interference parameters, respectively; β1, β2, and β3 represent the coefficients of the error identification model, respectively.

[0065] The probability of the presence of accuracy error impact obtained from the error identification model is compared with the error identification judgment threshold. When the probability of the presence of accuracy error impact is greater than or equal to the error identification judgment threshold, it is determined that there is accuracy error impact in the electrode voltage measurement process; when the probability of the presence of accuracy error impact is less than the error identification judgment threshold, it is determined that there is no accuracy error impact in the electrode voltage measurement process.

[0066] The specific methods for obtaining the characteristic indices of electrode corrosion parameters, environmental parameters, and slag layer interference parameters are as follows: the electrode corrosion parameter is specifically the change in electrode diameter; the environmental parameters specifically include electromagnetic interference intensity and ambient humidity; and the slag layer interference parameter is specifically the slag layer conductivity. The change in electrode diameter is recorded as the characteristic index of the electrode corrosion parameter. The electromagnetic interference intensity and ambient humidity are normalized, and the sum of the normalized electromagnetic interference intensity and ambient humidity is recorded as the characteristic index of the environmental parameter. The slag layer conductivity is recorded as the characteristic index of the slag layer interference parameter.

[0067] The change in electrode diameter reflects the change in the diameter of the electrode due to factors such as corrosion during the arc heating of molten steel. Electrode corrosion affects its electrical properties, which in turn affects the accuracy of electrode voltage measurement. Before and after arc heating, the diameter of the electrode is measured using high-precision calipers and other measuring tools, and then the difference between the two is calculated and expressed in length units (such as millimeters). The larger the change in diameter, the more severe the electrode corrosion, and the greater the potential impact on voltage measurement.

[0068] Electromagnetic interference intensity refers to the strength of electromagnetic interference in the electric arc heating environment. Electromagnetic interference may interfere with the transmission and measurement of electrode voltage signals and introduce errors. An electromagnetic interference tester is used to measure the electromagnetic interference at a suitable location near the electrode. The tester will display the current electromagnetic interference intensity value in real time, in microtesla (μT). The larger the value, the stronger the electromagnetic interference.

[0069] Ambient humidity reflects the water vapor content in the electric arc heating workshop environment. Humidity may affect the conductivity of the electrode surface and the performance of electrical equipment, thus affecting voltage measurement. A humidity sensor is used for real-time monitoring. The sensor converts humidity information into an electrical signal output, which is expressed as a percentage (%) of relative humidity. The higher the humidity, the greater the potential impact on electrode voltage measurement.

[0070] The conductivity of the slag layer reflects its ability to conduct current. Changes in the conductivity of the slag layer will affect the current distribution and voltage drop between the electrode and the molten steel, and thus affect the electrode voltage measurement. A slag layer sample is taken from the ladle and measured using a specialized conductivity measuring instrument. The unit is Siemens per meter (S / m). The conductivity value reflects the strength of the slag layer's electrical conductivity.

[0071] The specific steps for training the error identification model and confirming the coefficients are as follows: Extract error source parameters such as electrode diameter variation, electromagnetic interference intensity, ambient humidity, and slag conductivity from historical records of electric arc heating of molten steel. Simultaneously, mark whether the corresponding electrode voltage measurement is affected by accuracy errors (marked as 1 if present, 0 if absent); normalize the electromagnetic interference intensity and ambient humidity, scaling the data to a suitable range to eliminate differences in units and numerical ranges between different parameters; select a suitable binary classification model (logistic regression model) as the error identification model and initialize the model coefficients to random values; divide the compiled historical data into training and testing sets, using the training set data to train the model, and continuously adjust the model coefficients to ensure that the probability of accuracy errors in the model output closely matches the actual labeled values. Approximate optimization algorithms, such as gradient descent, aim to minimize the loss function (e.g., logarithmic loss function). The model's performance is evaluated using test set data, such as calculating accuracy and recall. If the model performance does not meet requirements, the model parameters are adjusted or the model structure is changed, and retraining is performed. After multiple iterations of training and evaluation, when the model performance reaches a satisfactory level, the model coefficients at this point are determined as the final error identification model coefficients. These coefficients represent the degree and direction of the influence of electrode corrosion parameters, environmental parameters, and slag layer interference parameters on the probability of accuracy error. For example, a positive and large β1 indicates that the greater the change in electrode diameter, the higher the probability of accuracy error in electrode voltage measurement; conversely, a negative β1 indicates a negative correlation between the change in electrode diameter and the probability of error.

[0072] The error identification threshold is obtained as follows: Based on past production experience and expert knowledge, an initial threshold is set, and then continuously adjusted and optimized in practical applications. Different thresholds can also be used to test the trained model on a validation dataset. Performance indicators such as accuracy, recall, and F1 score are calculated for each threshold. The threshold that achieves the best overall performance is selected as the error identification threshold. The receiver operating characteristic (ROC) curve of the model is obtained or plotted. By calculating indicators such as the area under the curve (AUC), the optimal threshold point is found, which can achieve a good balance between the false positive rate and the false negative rate.

[0073] By comprehensively considering error sources such as electrode corrosion, environment, and slag layer, the accuracy of electrode voltage measurement can be identified more comprehensively and accurately, thereby improving the reliability of measurement results and providing a solid foundation for subsequent process control and quality assurance. An error identification model based on a binary classification model transforms the complex error judgment problem into a probability calculation problem. By comparing the probability of the accuracy error's impact with the judgment threshold, it can quickly and automatically determine whether an error influence exists, achieving intelligent decision-making and reducing human interference. The model is trained using historical records of electric arc heating of molten steel, enabling it to learn the intrinsic relationship between error source parameters and measurement errors under different operating conditions. With continuous data accumulation and model optimization, the accuracy of error identification will continuously improve, providing strong support for production process optimization. Normalization and other methods are used to transform different types of parameters into unified feature indicators, enabling the model to adapt to different production environments and equipment conditions, enhancing its versatility and adaptability.

[0074] S3. When it is determined that there is an accuracy error in the electrode voltage measurement process, the influencing factors are identified based on error analysis. The influencing factors are used to quantify the degree of influence of the error on the electrode voltage measurement.

[0075] The specific analysis based on error analysis to identify influencing factors is as follows: For electrode corrosion parameters, the linear relationship between the change in electrode diameter and voltage measurement error is fitted:

[0076] ΔV corrosion =ε1*α1

[0077] Where, ΔV corrosion The voltage measurement error caused by electrode corrosion parameters is represented by ε1, which represents the linear coefficient of the fit. ε1 represents the change in voltage measurement error caused by electrode corrosion parameters for each unit change in electrode diameter. When using the least squares method for linear fitting, it is determined by solving the system of linear equations. α1 represents the change in electrode diameter.

[0078] For environmental parameters, a linear relationship is fitted between the characteristic indices of the environmental parameters and the voltage measurement error:

[0079] ΔV env =ε2*α2

[0080] Where, ΔV env ε2 represents the voltage measurement error caused by environmental parameters, α2 represents the linear coefficient of the fit, and α2 represents the change in voltage measurement error caused by environmental parameters for each unit change in the characteristic index of the environmental parameters. When using the least squares method for linear fitting, it is determined by solving the system of linear equations.

[0081] For the slag layer interference parameters, a linear relationship is fitted between the slag layer conductivity and the voltage measurement error:

[0082] ΔV slag =ε3*α3

[0083] Where, ΔV slag ε3 represents the voltage measurement error caused by the interference parameters of the slag layer, ε3 represents the linear coefficient of the fit, and α3 represents the change in voltage measurement error caused by the interference parameters of the slag layer for every unit change in the conductivity of the slag layer. When using the least squares method for linear fitting, it is determined by solving the system of linear equations.

[0084] Weights are assigned based on the degree of influence of each error source, and then the influence factors are obtained by combining the voltage measurement errors caused by electrode corrosion parameters, voltage measurement errors caused by environmental parameters, and voltage measurement errors caused by slag layer interference parameters:

[0085] IF=κ1*ΔV corrosion +κ2*ΔV env +κ3*ΔV slag

[0086] Where IF represents the impact factor, and κ1, κ2, and κ3 represent the weight values ​​of each error source, respectively;

[0087] In this implementation plan, the specific steps for fitting the linear relationship between the change in electrode diameter and the voltage measurement error are as follows: A series of experiments are designed to measure the corresponding electrode voltage under different electrode diameter changes, and the actual voltage measurement error is recorded. During the experiments, other conditions, such as environmental parameters and slag layer conditions, must be kept as consistent as possible to accurately reflect the relationship between the change in electrode diameter and the voltage measurement error. The collected data is cleaned to remove outliers and noise. Simultaneously, the completeness of the data is checked; if missing values ​​are found, interpolation or other suitable methods can be used to fill them. A scatter plot is drawn with the change in electrode diameter as the abscissa and the voltage measurement error as the ordinate to observe the approximate relationship between the two and determine if a linear trend exists. Linear regression methods, such as least squares, are used to fit the data to obtain a linear equation. Evaluation indicators, such as the correlation coefficient and determination coefficient of the fitted line, are calculated to assess the fitting effect. If the fitting effect is not ideal, the data needs to be re-examined or other fitting methods should be considered.

[0088] The specific steps for fitting the linear relationship between the characteristic indices of environmental parameters and voltage measurement error are as follows: Change the environmental parameters electromagnetic interference intensity and ambient humidity, simultaneously measure the electrode voltage and record the voltage measurement error. Normalize the electromagnetic interference intensity and ambient humidity, then sum them to obtain the characteristic indices of the environmental parameters. Similar to the data preprocessing steps for fitting the linear relationship between the change in electrode diameter and voltage measurement error, plot a scatter plot with the characteristic indices of the environmental parameters on the x-axis and the voltage measurement error on the y-axis to observe the linear trend. Use the least squares method for linear fitting to obtain the linear equation. Calculate relevant evaluation indicators to assess the fitting effect.

[0089] The specific steps for fitting the linear relationship between slag layer conductivity and voltage measurement error are as follows: Under different slag layer conductivity conditions, measure the electrode voltage and record the voltage measurement error; adjust the slag layer conductivity by changing factors such as slag layer composition and temperature; similar to the data preprocessing steps above, plot a scatter plot with slag layer conductivity on the x-axis and voltage measurement error on the y-axis to observe the linear trend; use the least squares method to perform linear fitting, obtain the linear equation, calculate relevant evaluation indicators, and evaluate the fitting effect.

[0090] The weight values ​​are obtained as follows: A hierarchical model is constructed, with each error source treated as a hierarchical element. By comparing the relative importance of different error sources, a judgment matrix is ​​constructed, and then the eigenvectors of the judgment matrix are solved to obtain the weight value of each error source. Alternatively, historical data can be used, and machine learning algorithms, such as multiple linear regression and decision trees, can be employed to determine the contribution of each error source to the voltage measurement error. This contribution level is then used as the weight value. For example, using a multiple linear regression model, the voltage measurement error is used as the dependent variable, while the electrode diameter change, characteristic indicators of environmental parameters, and slag conductivity are used as independent variables. The absolute value of the regression coefficient reflects the influence of each independent variable on the dependent variable, and the weight value is obtained after normalization.

[0091] By fitting the linear relationships between electrode corrosion parameters, environmental parameters, slag interference parameters, and voltage measurement errors, the specific impact of each error source on electrode voltage measurement can be accurately quantified, providing precise data support for subsequent error correction and adjustment. By assigning weights based on the impact of each error source and combining the voltage measurement errors caused by different sources to obtain influence factors, this method, which comprehensively considers multiple factors, can more comprehensively and objectively reflect the actual situation, avoiding focusing on only a single factor while ignoring the influence of other important factors. Clear influence factors can help operators quickly understand the main sources and severity of electrode voltage measurement errors, thereby taking targeted measures for adjustment and optimization, improving the stability of the production process and product quality. The linear fitting method is simple, intuitive, and easy to understand and implement. Furthermore, this method has a certain degree of scalability, allowing error source parameters to be added or adjusted according to actual conditions, making it suitable for different electric arc heating molten steel production scenarios.

[0092] S4. Monitor the voltage changes of the three electrodes of the LF furnace in real time, and build a model of the relationship between the voltage of the three electrodes and the argon gas flow and slag layer state based on historical data.

[0093] The specific analysis of constructing a model of the relationship between the three electrode voltages, argon flow rate, and slag layer state based on historical data is as follows: Voltage data and corresponding argon flow rate and slag layer thickness at each historical time point are collected. The voltage data includes the voltage values ​​of the three electrodes of the LF furnace. The voltage data and corresponding argon flow rate and slag layer thickness at each historical time point are processed, mapping all data to a range of 0 to 1. Based on a multinomial regression model, the voltage data and argon flow rate or the voltage data and slag layer thickness at each historical time point are fitted to obtain the relationship model between the three electrode voltages and argon flow rate or the three electrode voltages and slag layer state.

[0094] In this implementation scheme, the voltage values ​​of the three electrodes of the LF furnace of the ladle refining furnace reflect the potential difference between each electrode and the molten steel. During the process of heating the molten steel by electric arc, an electric arc is generated between the electrode and the molten steel. The magnitude of the voltage value is closely related to the intensity and stability of the electric arc and the electrical characteristics between the electrode and the molten steel. Different combinations of electrode voltages can reflect the distribution of the electric arc and the energy input, thereby affecting the heating effect of the molten steel and the refining process.

[0095] The specific analysis of indirectly identifying the argon gas flow and slag layer requirements through electrode voltage values ​​is as follows: The argon gas flow and slag layer condition affect the stability of the electric arc and the energy transfer efficiency, thus affecting the electrode voltage. For example, a suitable argon gas flow can stir the molten steel, improve the arc distribution, and make the electrode voltage more stable. The thickness and composition of the slag layer affect the resistance between the electrode and the molten steel, thereby changing the electrode voltage. Therefore, by monitoring the changes in electrode voltage, it is possible to indirectly infer whether the argon gas flow and slag layer condition are within a suitable range, thereby identifying the requirements of the argon gas flow and slag layer.

[0096] The specific steps for constructing the relational model are as follows: Collect the voltage values ​​of the three electrodes of the LF furnace at various time points from historical production records, along with the corresponding argon flow rate and slag thickness data, ensuring data accuracy and completeness, and processing any missing or abnormal data; standardize the collected voltage, argon flow rate, and slag thickness data, mapping all data to between 0 and 1; select a multinomial regression model as the base model, with the dependent variable being either the argon flow rate or the slag thickness, and the independent variable being the combination of the three electrode voltages; initialize the model's order, generally starting with a lower order, such as second or third order; divide the standardized data into training and testing sets, typically with the training set accounting for 70%-80% of the total data and the testing set accounting for 20%-30%. 0%, the training set is used for model training, and the test set is used to evaluate the model's performance; the multinomial regression model is fitted using the training set data, and the model coefficients are determined by minimizing the sum of squared errors, specifically using optimization algorithms such as least squares method to solve for the model coefficients; the trained model is evaluated using the test set data, and evaluation metrics such as mean squared error, mean absolute error, and coefficient of determination are calculated. If the model performance does not meet the requirements, the model order can be adjusted or other optimization methods can be used, and the training and evaluation can be repeated; after multiple iterations and evaluations, the model with the best performance is selected as the final model for the relationship between the three electrode voltages and the argon gas flow or the three electrode voltages and the slag layer state. An example of the expression for the relationship between the three electrode voltages and the argon gas flow is as follows:

[0097] Q = η0 + η1 * V A +η2*V B +η3*V C +η4*V A 2

[0098] +η5*V B 2 +η6*V C 2 +η7*V A *V B +η8*V A *V C +η9*VB *V C

[0099] Where Q represents the argon gas flow rate, V A V B V C η0, η1, η2, ..., η9 represent the voltage values ​​of phase A, B, and C electrodes, respectively, and η0, η1, η2, ..., η9 represent the model coefficients, respectively.

[0100] The specific model expression for the relationship between the three electrode voltages and the slag layer state is as follows:

[0101] H = μ0 + μ1 * V A +μ2*V B +μ3*V C +μ4*V A 2

[0102] +μ5*V B 2 +μ6*V C 2 +μ7*V A *V B +μ8*V A *V C +μ9*V B *V C

[0103] Where H represents the thickness of the slag layer, and μ0, μ1, μ2, ..., μ9 represent the model coefficients, respectively;

[0104] By constructing a model relating three electrode voltages to argon gas flow and slag layer state, the required argon gas flow and slag layer state can be accurately predicted based on real-time monitored electrode voltage data. This facilitates precise control of argon gas flow and slag layer thickness during arc heating of molten steel, improving steel quality and production efficiency. Utilizing historical data in model construction allows the model to reflect the inherent patterns in actual production processes. With continuous data accumulation and model optimization, the model's accuracy and reliability can be further improved, providing strong support for production process optimization. Accurate prediction of argon gas flow and slag layer state requirements avoids excessive argon use and unreasonable slag layer adjustments, thereby reducing production costs. Simultaneously, it reduces production accidents and product quality issues caused by improper parameter adjustments, improving production stability and economic benefits. Employing a multinomial regression model as the foundation provides versatility and scalability, allowing adjustment of the model's order and parameters to adapt to different production environments and process requirements.

[0105] S5. Based on the real-time monitored voltage data and the identification of influencing factors, the demand for argon gas flow and slag layer status is predicted using the relationship model between the three electrode voltages and the argon gas flow and slag layer status.

[0106] Based on real-time monitored voltage data and the identification of influencing factors, the specific analysis for predicting the demand for argon gas flow and slag layer state using the relationship model between the three electrode voltages and argon gas flow and slag layer state is as follows: Real-time voltage data is obtained based on real-time monitoring of production records, including the real-time voltage values ​​of the three electrodes of the LF furnace; the impact of accuracy errors during electrode voltage measurement is identified; when it is determined that there are no accuracy errors during electrode voltage measurement, the real-time voltage values ​​of the three electrodes of the LF furnace are input into the relationship model between the three electrode voltages and argon gas flow, and the relationship model between the three electrode voltages and slag layer state, respectively. The expected values ​​of argon gas flow and slag layer thickness are obtained. When it is determined that there is an accuracy error in the electrode voltage measurement process, the influencing factor is introduced into the relationship model between the three electrode voltages and the argon gas flow, and the relationship model between the three electrode voltages and the slag layer state. The optimized relationship models between the three electrode voltages and the argon gas flow, and the optimized relationship models between the three electrode voltages and the slag layer state are updated. The real-time voltage values ​​of the three electrodes of the LF furnace are input into the optimized relationship models between the three electrode voltages and the argon gas flow, and the optimized relationship models between the three electrode voltages and the slag layer state, respectively, and the expected values ​​of argon gas flow and slag layer thickness are output.

[0107] The specific optimized relationship model expression between the three electrode voltages and the argon gas flow is shown as: Q'=Q+IF*η 10 Where Q' represents the argon gas flow rate output by the optimized relational model, and η 10 These are the model coefficients, obtained through least squares fitting.

[0108] The specific optimized relationship model expression between the three electrode voltages and the slag layer state is shown in the following example: H' = H + IF * μ 10 In the formula, H' represents the slag layer thickness output by the optimized relational model, μ 10 The coefficients are model coefficients, obtained through least squares fitting.

[0109] S6. Generate and execute adjustment strategies for argon gas flow and slag layer state based on the predicted argon gas flow and slag layer state requirements.

[0110] The specific analysis of generating and executing the adjustment strategy for argon gas flow and slag layer state based on the predicted argon gas flow and slag layer state requirements is as follows: receiving the expected value of argon gas flow and the expected value of slag layer thickness, and identifying the slag layer thickness and argon gas flow input in the real-time LF furnace; controlling the argon gas and slag layer input materials in the LF furnace based on the expected value of argon gas flow and the expected value of slag layer thickness, including but not limited to increasing argon gas supply and decreasing slag layer input material supply, and decreasing argon gas supply and increasing slag layer input material supply.

[0111] In this implementation plan, the specific process for controlling the argon gas and slag layer feed materials in the LF furnace based on the expected argon gas flow rate and the expected slag layer thickness is as follows: Receive the expected argon gas flow rate Q. expected Compared with the expected value H of the slag layer thickness expected Simultaneously, the argon flow rate Q in the LF furnace is identified through corresponding monitoring equipment (such as argon flow meters and slag layer thickness measuring instruments). current and slag layer thickness H current The expected value of argon gas flow Q expected Argon gas flow rate Q in the LF furnace current By comparing the values, the difference ΔQ is obtained, and the expected value H of the slag layer thickness is calculated. expected The slag layer thickness H in the LF furnace current The difference ΔH is obtained through comparison. A specific example of the control strategy is as follows: When ΔQ > 0 and ΔH < 0, a strategy of increasing argon gas supply and reducing slag layer material supply is adopted. Specifically, this could involve increasing the valve opening of the argon gas delivery pipeline to increase the argon flow rate while simultaneously reducing the amount of slag added. When ΔQ < 0 and ΔH > 0, a strategy of reducing argon gas supply and increasing slag layer material supply is adopted. For example, this could involve decreasing the valve opening of the argon gas delivery pipeline to reduce the argon flow rate and increasing the amount of slag added. When ΔQ > 0 and ΔH > 0... The adjustment of parameters with a significant impact on production needs to be comprehensively considered, taking into account factors such as production process requirements and equipment capacity. Priority should be given to adjusting parameters that have a greater impact on production, or simultaneously increasing argon supply and slag layer material supply appropriately. When ΔQ < 0 and ΔH < 0, various factors should also be comprehensively considered, prioritizing adjustments to parameters with a greater impact on production, or simultaneously appropriately reducing argon supply and slag layer material supply. When ΔQ = 0 and ΔH = 0, it indicates that the current argon flow and slag layer state meet expectations, and the current argon supply and slag layer material supply state should be maintained. According to the established control strategy, the corresponding equipment, such as argon valves and slag material feeding devices, should be operated to control the argon flow and slag layer material in the LF furnace. After executing the control strategy, the changes in the argon flow input and slag layer thickness in the LF furnace should be continuously monitored and compared with the expected values. If the actual values ​​still deviate significantly from the expected values, the cause should be analyzed and the control strategy adjusted. The control operation should be re-executed until the expected argon flow and slag layer state are achieved.

[0112] Based on the presence of accuracy errors during electrode voltage measurement, different models are flexibly selected to predict argon gas flow and slag layer status. When there are no errors, the original relationship model is used; when errors exist, an impact factor is introduced to update the model, making the prediction results more consistent with reality and improving prediction accuracy. By monitoring voltage data in real time and combining it with model prediction, the impact of electrode voltage measurement errors on the prediction of argon gas flow and slag layer status can be promptly identified and addressed, ensuring accurate demand prediction under various conditions. This leads to the generation of reasonable adjustment strategies, guaranteeing the stability of the LF furnace production process and reducing production fluctuations caused by inaccurate parameters. Based on accurate prediction results, the generation and execution of adjustment strategies enable precise control of argon gas and slag layer input materials. For example, argon gas supply and slag layer input material supply can be increased or decreased according to demand, optimizing resource allocation in the production process, improving production efficiency, and reducing production costs. The design considers the factor of electrode voltage measurement error, enabling the system to adapt to different measurement accuracy conditions. Whether the measurement accuracy is high or there are errors, appropriate processing methods can ensure the normal operation and accurate prediction of the system, enhancing the system's adaptability and robustness.

[0113] See Figure 2 The present invention provides an adjustment system for an argon slag layer in molten steel heated by electric arc, comprising an error identification module, an error judgment module, an influencing factor acquisition module, a relationship model construction module, an argon slag layer state identification module, and an adjustment execution module.

[0114] Error identification module: Identifies the error sources of electrode voltage during the electric arc heating of molten steel. These error sources include, but are not limited to, electrode corrosion parameters, environmental parameters, and slag layer interference parameters. For the identification of error source parameters, high-precision instruments such as calipers, laser rangefinders, temperature and humidity sensors, electromagnetic interference testers, and spectral analyzers are used to periodically or in real time detect electrode corrosion, environmental, and slag layer parameters.

[0115] Error Judgment Module: An error identification model is constructed based on a binary classification model. It uses the error source parameters of electrode voltage during the electric arc heating of molten steel to determine whether there are accuracy errors affecting the electrode voltage measurement process. In order to construct the error identification model, a large amount of error source parameter data under different working conditions is collected. A binary classification model such as support vector machine is selected. The model is trained on the training set and tested and optimized on the test set to improve the accuracy and generalization ability of the model.

[0116] Influence Factor Acquisition Module: When it is determined that there is an accuracy error in the electrode voltage measurement process, the influencing factors are identified based on error analysis. The influencing factors are used to quantify the degree of influence of the error on the electrode voltage measurement. For the error influencing factor analysis, mathematical methods such as the control variable method and regression analysis are used to determine the degree of influence of the error source parameters on the electrode voltage measurement and the value of the influencing factors through sensitivity analysis and quantitative analysis.

[0117] The relational model construction module monitors the voltage changes of the three electrodes of the LF furnace in real time and constructs a relational model between the three electrode voltages, argon flow rate, and slag layer state based on historical data. For the construction of the relational model and real-time monitoring, historical data on electrode voltage, argon flow rate, and slag layer state under different operating conditions of the LF furnace are collected, preprocessed, and used to construct the relational model using methods such as multiple linear regression. High-precision voltage, argon flow rate, and slag layer state monitoring equipment are installed to acquire data in real time.

[0118] Argon Slag Layer Status Identification Module: Based on real-time monitored voltage data and the identification of influencing factors, this module uses a model of the relationship between three electrode voltages, argon gas flow rate, and slag layer status to predict the demand for argon gas flow and slag layer status. For adjusting argon gas flow and slag layer status, it formulates adjustment strategies based on predicted demand and production processes, such as increasing or decreasing argon gas flow, slag removal, or adjusting slag layer composition. The module utilizes automated control systems such as PLCs to implement adjustments and includes a feedback mechanism to adjust strategies promptly based on comparisons between actual and target values. Regular maintenance and calibration of the system ensure the equipment and model are functioning correctly and accurately; operators must possess professional knowledge and skills.

[0119] Adjustment Execution Module: Generates and executes adjustment strategies for argon gas flow and slag layer state based on the predicted argon gas flow and slag layer state requirements.

[0120] In summary, this application has at least the following effects:

[0121] By identifying error sources in electrode voltage during electric arc heating of molten steel, such as electrode corrosion parameters, environmental parameters, and slag layer interference parameters, we can more accurately understand the factors affecting electrode voltage measurement. Using a binary classification model to construct an error identification model allows for real-time assessment of any accuracy errors affecting electrode voltage measurement, enabling timely corrective measures and improving measurement accuracy. Error analysis identifies influencing factors, quantifying the impact of errors on electrode voltage measurement and providing a scientific basis for subsequent adjustment strategies. Based on the identified influencing factors, we can more accurately predict the demand for argon gas flow and slag layer conditions, thereby generating and implementing more reasonable adjustment strategies to optimize the electric arc heating process. Real-time monitoring of voltage changes at the three electrodes of the LF furnace allows for timely detection of voltage anomalies, preventing potential production failures. Building a model of the relationship between the three electrode voltages, argon gas flow, and slag layer conditions based on historical data enables accurate prediction of argon gas flow and slag layer conditions, providing strong support for production scheduling and decision-making. Generating and implementing adjustment strategies based on predicted argon gas flow and slag layer conditions ensures the stability and efficiency of the electric arc heating process, improving production efficiency.

[0122] Those skilled in the art will understand that embodiments of the present invention can be provided as methods or systems. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0123] This invention is described with reference to flowchart illustrations and structural diagrams of methods and systems according to embodiments of the invention. It should be understood that the combination of each process and module in the flowchart and structural diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the process. Figure 1 One or more processes and structures Figure 1 A device for a function specified in one or more modules.

[0124] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and structures Figure 1 The function specified in one or more modules.

[0125] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and structures Figure 1 The steps of a specified function in one or more modules.

[0126] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0127] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for adjusting the argon slag layer in molten steel based on electric arc heating, characterized in that, Includes the following steps: S1. Identify the error source parameters of the electrode voltage during the electric arc heating of molten steel. The error source parameters include electrode corrosion parameters, environmental parameters, and slag layer interference parameters. S2. An error identification model is constructed based on a binary classification model. This model uses the error source parameters of the electrode voltage during the arc heating of molten steel to determine whether there are accuracy errors affecting the electrode voltage measurement process. Specifically: The binary classification model is used as the basis for the error identification model. The error source parameters of the electrode voltage are used as the input of the error identification model. The output of the error identification model is the probability of the existence of the accuracy error. The error identification model is trained using historical records of electric arc heating of molten steel to confirm the model coefficients. Then, the error source parameters of the electrode voltage acquired in real-time during the electric arc heating process are input into the error identification model. The error identification model is as follows: in, This indicates that the accuracy error has a probability of affecting the accuracy. , , These represent characteristic indicators of electrode corrosion parameters, environmental parameters, and slag layer interference parameters, respectively. , , These represent the coefficients of the error identification model; The probability of the presence of accuracy error impact obtained from the error identification model is compared with the error identification judgment threshold. When the probability of the presence of accuracy error impact is greater than or equal to the error identification judgment threshold, it is determined that there is accuracy error impact in the electrode voltage measurement process; when the probability of the presence of accuracy error impact is less than the error identification judgment threshold, it is determined that there is no accuracy error impact in the electrode voltage measurement process. S3. When it is determined that there is an accuracy error in the electrode voltage measurement process, the influencing factors are identified based on error analysis. The influencing factors are used to quantify the degree of influence of the error on the electrode voltage measurement. S4. Real-time monitoring of voltage changes at the three electrodes of the LF furnace, and construction of a model relating the three electrode voltages to argon gas flow and slag layer state based on historical data; specifically: Collect voltage data and corresponding argon flow rate and slag layer thickness at various historical time points. The voltage data includes the voltage values ​​of the three electrodes of the LF furnace. The voltage data and corresponding argon flow rate and slag layer thickness at each historical time point are processed, and all data are mapped to a range of 0 to 1. Based on the multinomial regression model, the voltage data at each historical time point were fitted with the argon gas flow rate or the voltage data at each historical time point were fitted with the slag layer thickness to obtain the relationship model between the three electrode voltages and the argon gas flow rate or the three electrode voltages and the slag layer state. S5. Based on real-time monitored voltage data and the identification of influencing factors, predict the demand for argon gas flow and slag layer condition using a model relating the three electrode voltages to argon gas flow and slag layer condition; specifically: Real-time voltage data is obtained through real-time monitoring of production records. The real-time voltage data includes the real-time voltage values ​​of the three electrodes of the LF furnace. The system identifies the impact of accuracy errors during electrode voltage measurement. When it is determined that there are no accuracy errors during electrode voltage measurement, the real-time voltage values ​​of the three electrodes of the LF furnace are input into the relationship model between the three electrode voltages and the argon gas flow, and the relationship model between the three electrode voltages and the slag layer state, respectively. The expected values ​​of argon gas flow and slag layer thickness are then output. When it is determined that there is an accuracy error in the electrode voltage measurement process, the influencing factor is introduced into the relationship model between the three electrode voltages and the argon gas flow, and the relationship model between the three electrode voltages and the slag layer state. The optimized relationship models between the three electrode voltages and the argon gas flow, and the optimized relationship models between the three electrode voltages and the slag layer state are updated to obtain the optimized relationship models between the three electrode voltages and the argon gas flow, and the optimized relationship models between the three electrode voltages and the slag layer state, respectively. The expected values ​​of the argon gas flow and the expected values ​​of the slag layer thickness are then output. S6. Generate and execute adjustment strategies for argon gas flow and slag layer state based on the predicted argon gas flow and slag layer state requirements; specifically: Receive the expected value of argon gas flow and the expected value of slag layer thickness, and identify the slag layer thickness and argon gas flow input in the LF furnace in real time; Based on the expected values ​​of argon gas flow and slag layer thickness, the argon gas and slag layer input materials in the LF furnace are controlled. The control includes increasing the argon gas supply and decreasing the slag layer input material supply, and decreasing the argon gas supply and increasing the slag layer input material supply.

2. The method for adjusting the argon slag layer in molten steel based on electric arc heating according to claim 1, characterized in that, The specific methods for obtaining the characteristic indicators of the electrode corrosion parameters, environmental parameters, and slag layer interference parameters are as follows: The electrode corrosion parameter is specifically the change in electrode diameter; the environmental parameters include electromagnetic interference intensity and ambient humidity; and the slag layer interference parameter is specifically the slag layer conductivity. The change in electrode diameter is recorded as a characteristic index of electrode corrosion parameters. The electromagnetic interference intensity and ambient humidity are normalized, and the sum of the normalized electromagnetic interference intensity and ambient humidity is recorded as the characteristic index of the environmental parameter. The electrical conductivity of the slag layer is recorded as a characteristic index of the slag layer interference parameters.

3. The method for adjusting the argon slag layer in molten steel based on electric arc heating according to claim 2, characterized in that, The specific analysis of influencing factors based on error analysis is as follows: for electrode corrosion parameters, a linear relationship is fitted between the change in electrode diameter and the voltage measurement error. in, This indicates the voltage measurement error caused by electrode corrosion parameters. Represents the linear coefficients of the fit; For environmental parameters, a linear relationship is fitted between the characteristic indices of the environmental parameters and the voltage measurement error: in, This indicates the voltage measurement error caused by environmental parameters. Represents the linear coefficients of the fit; To address the interference parameters of the slag layer, a linear relationship between the slag layer conductivity and the voltage measurement error was fitted: in, This indicates the voltage measurement error caused by interference parameters from the slag layer. Represents the linear coefficients of the fit; Weights are assigned based on the degree of influence of each error source, and then the influence factors are obtained by combining the voltage measurement errors caused by electrode corrosion parameters, voltage measurement errors caused by environmental parameters, and voltage measurement errors caused by slag layer interference parameters: Wherein, IF represents the impact factor. , , These represent the weight values ​​for each source of error.

4. A system for adjusting an argon-slag layer in molten steel heated by an electric arc, employing the adjustment method for an argon-slag layer in molten steel heated by an electric arc as described in any one of claims 1-3, characterized in that... include: Error judgment module: Based on a binary classification model, an error identification model is constructed. The error source parameters of electrode voltage during the electric arc heating of molten steel are used to determine whether there is an accuracy error in the electrode voltage measurement process. Influence Factor Acquisition Module: When it is determined that there is an accuracy error in the electrode voltage measurement process, the module identifies the influence factors based on error analysis. The influence factors are used to quantify the degree of influence of the error on the electrode voltage measurement. Relationship model building module: Real-time monitoring of voltage changes at the three electrodes of the LF furnace, and construction of a relationship model between the voltage of the three electrodes and the argon gas flow and slag layer state based on historical data; Argon Slag Layer State Identification Module: Based on real-time monitored voltage data and the identification status of influencing factors, the module uses a model of the relationship between the three electrode voltages, argon gas flow rate, and slag layer state to predict the demand for argon gas flow rate and slag layer state. Adjustment Execution Module: Generates and executes adjustment strategies for argon gas flow and slag layer state based on the predicted argon gas flow and slag layer state requirements.

Images