An intelligent online optimization method for converter steelmaking process oriented to low-carbon and low-cost smelting

Abstract

The present invention belongs to the technical field of iron and steel metallurgy, and specifically relates to an intelligent online optimization method for converter steelmaking process oriented to low-carbon and low-cost smelting. By applying on-site manufacturing standards and historical heat production data, comprehensively considering indicators such as converter hot metal conditions, smelting targets, upper and lower limits of manufacturing conditions, smelting costs, and carbon emission intensity, and combining intelligent algorithms such as local online modeling and multi-objective optimization, it realizes the intelligent online optimization of converter steelmaking smelting and operation process parameters, improves the intelligent control level of on-site converters, and is conducive to achieving the dual goals of energy conservation, emission reduction, cost reduction, and efficiency increase in converter smelting.

Description

Technical Field

[0001] The present invention relates to the technical field of iron and steel metallurgy, and specifically to an intelligent online optimization method for converter steelmaking process for low-carbon and low-cost smelting. Background Art

[0002] Low-carbon and low-cost smelting in converters is a key path for the current iron and steel industry to pursue both green transformation and economic benefits. With the increasingly severe global climate change problem, reducing carbon emissions has gradually become the consensus of the international community. As a major carbon emitter, the iron and steel industry is facing huge emission reduction pressure. At the same time, the rising raw material prices and increasing energy costs also prompt iron and steel enterprises to seek low-cost production methods. Therefore, low-carbon and low-cost smelting in converters has become an important direction for the transformation and upgrading of the iron and steel industry. Through technological innovation and process optimization, it aims to achieve the dual goals of energy conservation, emission reduction, cost reduction and efficiency improvement.

[0003] Currently, the optimization of converter smelting and operation process parameters in iron and steel enterprises is mainly based on on-site manufacturing regulations and manual experience, and it is difficult to adapt to refined production control and diverse smelting goals. For this reason, the present invention proposes an intelligent online optimization method for converter steelmaking process for low-carbon and low-cost smelting. Using the on-site manufacturing standards and historical heat production data in the big data platform of iron and steel enterprises, considering multiple indicators such as converter hot metal conditions, smelting goals, upper and lower limits of manufacturing conditions, smelting costs and carbon emission intensity during the converter smelting process, combined with intelligent algorithms such as local online modeling and multi-objective optimization, the converter steelmaking process is intelligently optimized online to achieve low-carbon emission and low-cost smelting production in the converter, and to guide and improve the converter smelting level of on-site operators. Summary of the Invention

[0004] To solve the problems existing in the prior art, the main object of the present invention is to propose an intelligent online optimization method for converter steelmaking process for low-carbon and low-cost smelting. Based on the on-site manufacturing standards of the converter and the historical heat production data, considering multiple indicators such as converter hot metal conditions, smelting goals, upper and lower limits of manufacturing conditions, smelting costs and carbon emission intensity, using intelligent algorithms such as local online modeling and multi-objective optimization, the converter smelting and operation process parameters are intelligently optimized online to guide the on-site realization of low-carbon and low-cost smelting production in the converter. Specifically,

[0005] An intelligent online optimization method for converter steelmaking process oriented to low-carbon and low-cost smelting, characterized in that, aiming at the converter smelting target of the current heat, applying on-site manufacturing standards and production data of historical heats, instantaneously optimizing the converter production process parameters of the current heat online to meet the requirements of low-carbon and low-cost converter smelting; the optimization model used comprehensively considers indicators such as converter hot metal conditions, smelting targets, upper and lower limits of manufacturing conditions, smelting costs and carbon emission intensity, and flexibly adjusts the importance of different indicators in the optimization process through weight coefficients; combining intelligent algorithms such as local online modeling and multi-objective optimization, outputting the best smelting and operation process parameters of the current heat, and realizing the intelligent online optimization of the converter steelmaking process.

[0006] Furthermore, the construction of the optimization model includes three modules:

[0007] 1) Initial value selection module: Comprehensively considering the converter hot metal conditions, smelting targets, smelting costs and carbon emission intensity, selecting heat samples from historical heats that are similar to the smelting target of the current heat and have lower costs and carbon emissions as the initial values for optimization;

[0008] 2) Local weighted regression modeling module: According to the obtained initial values for optimization, performing local weighted regression modeling near them, and the solution process includes parts such as sample weight determination, fitting equation construction, equation parameter optimization and loss function evaluation;

[0009] 3) Optimal condition derivation module: Combining multi-objective conditions such as converter smelting targets, upper and lower limits of manufacturing conditions, smelting costs and carbon emission intensity, iteratively solving the optimization model until convergence is reached, and outputting the best smelting and operation process parameters of the current heat.

[0010] Furthermore, the initial value selection module can be expressed as:

[0011]

[0012] In the formula, y n is the hot metal condition and smelting end point of the historical heat, y 0 is the hot metal condition and smelting target of the current heat, is a diagonal matrix for appropriately scaling the output variable, x is the converter smelting and operation process parameters, q is the weight of the deviation between the regression model fitting result and the smelting target, r is the weight of the smelting cost, c ≡ [c1 c2 … c M T represents the cost value corresponding to each process parameter, p is the weight of the carbon emission intensity, e ≡ [e1 e2 … e M T represents the carbon emission intensity value corresponding to each process parameter.

[0013] Furthermore, the local weighted regression modeling module can be expressed as: ​​

[0014] 1) Sample weight

[0015]

[0016] Wherein, is the Euclidean distance between the query sample x and the i-th sample; σ is a hyperparameter that determines the fitting degree of the model.

[0017] 2) Fitting equation

[0018]

[0019] Wherein, is the predicted value of the i-th output; b (i) is the constant term of the m-th input; a m(i) is the i-th parameter of the m-th input; x m is the m-th input.

[0020] 3) Equation parameters

[0021] θ (i) = [b (i) a 1(i) a 2(i) …a M(i) T

[0022] Wherein, b (i) is the constant term of the m-th input; a M(i) is the i-th parameter of the M-th input.

[0023] 4) Loss function

[0024]

[0025] Wherein, X is in the form of an input matrix,

[0026] Furthermore, the optimal condition derivation module can be expressed as:

[0027]

[0028] x LB ≤ x ≤ x UB

[0029] y LB ≤ y ≤ y UB

[0030] Wherein, x LB represents the lower limit value of the process parameter, x UB represents the upper limit value of the process parameter, y LB represents the lower limit value of the smelting target, y UB ​Represents the upper limit value of the smelting target.

[0031] Furthermore, the cost value corresponding to each process parameter is provided by the steel enterprise ERP management system. The upper and lower limit values of the process parameters and the smelting target are specified by the on-site manufacturing standards. The carbon emission intensity value is calculated based on the energy consumption and material consumption corresponding to each process parameter.

[0032] Furthermore, the value range of the weights is: 0.5 ≤ q ≤ 1, 0 ≤ c ≤ 1, 0 ≤ e ≤ 1.

[0033] Furthermore, the smelting and operation process parameters include scrap ratio, lime addition amount, dolomite addition amount, oxygen blowing amount, lance position, smelting time, deoxidizing alloy addition amount, etc.

[0034] The beneficial effects of the present invention are as follows:

[0035] The present invention proposes an intelligent online optimization method for converter steelmaking process for low-carbon and low-cost smelting. By using the on-site manufacturing standards and historical heat production data in the steel enterprise big data platform, and considering multiple indicators such as converter hot metal conditions, smelting targets, upper and lower limits of manufacturing conditions, smelting costs, and carbon emission intensity during the converter smelting process, combined with intelligent algorithms such as local online modeling and multi-objective optimization, the converter steelmaking process is intelligently optimized online, improving the intelligent control level of converter smelting, and facilitating the realization of the dual goals of energy conservation, emission reduction, cost reduction, and efficiency increase in converter smelting. Specific embodiments

[0036] The technical solutions in the embodiments will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0037] The technical solutions of the present invention will be further described below in conjunction with specific embodiments.

[0038] Embodiment

[0039] Taking the converter steelmaking production of silicon steel in a certain steel plant as an example, 2620 furnace historical heat production data are collected for modeling. At the start of converter smelting for a certain heat, the hot metal conditions are [C] = 4.32%, [Si] = 0.41%, [Mn] = 0.3%, [P] = 0.016%, T = 1278 °C. The converter smelting end point carbon and temperature targets are [C] = 0.06%, T = 1665 °C.

[0040] According to the initial value selection formula (1), furnace samples with smelting targets similar to the current furnace and lower costs and carbon emissions are selected from the historical furnaces.

[0041]

[0042] In the formula, y n represents the hot metal conditions and the smelting end point of the historical heat, and y 0 represents the hot metal conditions and the smelting target of the current heat. is a diagonal matrix for appropriately scaling the output variables, x is the converter smelting and operation process parameters, q is the weight of the deviation between the regression model fitting result and the smelting target, r is the weight of the smelting cost, and c ≡ [c1 c2 … c M T represents the cost value corresponding to each process parameter, p is the weight of the carbon emission intensity, and e ≡ [e1 e2 … e M T represents the carbon emission intensity value corresponding to each process parameter.

[0043] Set the weight q of the deviation between the regression model fitting result and the smelting target to 1, the weight r of the smelting cost to 1, and the weight e of the carbon emission intensity to 1. Determine the initial value furnace heat sample through the initial value selection module as hot metal [C] = 4.35%, [Si] = 0.44%, [Mn] = 0.31%, [P] = 0.017%, T = 1290°C, and the converter smelting end point carbon [C] = 0.05%, and the end point temperature is 1663°C.

[0044] Adopt the locally weighted modeling module to construct a local prediction model. By prioritizing the past data near the initial value, a linear prediction model with sufficient accuracy can be established around the initial value. Calculate the output corresponding to the target value by using the generated model.

[0045] Assign weights to the samples in the same way as in the locally weighted linear regression model, and the weight calculation formula is shown in Equation (2).

[0046]

[0047] In Equation (2) ‖x - x (i) ‖2 is the weighted Euclidean distance, x is the function to be queried (initial value), and x (i) is the i-th training sample (historical sample), and σ is the hyperparameter that determines the model fitting degree. If the hyperparameter σ is too small, the model will have overfitting; if σ is too large, the model will have underfitting. In actual applications, the situation where the output variables and feature vectors of the data are linearly related is rare. Therefore, the pure linear regression model has poor applicability, but the locally weighted linear regression can make a better fit to the non-linear data distribution. Randomly extract n historical samples from the database for complete cross-validation to determine the appropriate σ value.

[0048] ​​For predicting the output variable with a given feature vector as the input, the loss function after weighted processing is used in Equation (3). Equation (3) is a summation equation with weights added to the loss function. By minimizing the loss function, the algorithm can find the value of θ when the samples close to x have similar distances and the minimum loss, that is, the algorithm pays more attention to the points close to x, which is also beneficial for more accurately predicting y. Expressing Equation (3) in matrix form can be represented by Equation (4), and the local linear regression model expression is shown in Equation (5).

[0049]

[0050] In the formula, is the predicted value of the i-th output, and b (i) is the constant term of the m-th input, and a m(i) is the i-th parameter of the m-th input, and x m is the m-th input, and b (i) is the constant term of the m-th input; a M(i) is the i-th parameter of the M-th input.

[0051]

[0052] Finally, the module for deriving the optimal manufacturing conditions calculates the minimization of the objective function under given constraints using the quadratic programming method. Since the derived optimal manufacturing conditions are different from the initial conditions, it is necessary to continuously iterate in the local model to obtain the local weighted regression model and the derivation of the optimal manufacturing conditions until the optimal manufacturing conditions converge. This iteration is necessary to ensure the improvement of the accuracy of the local linear model around the derived optimal manufacturing conditions. The role of the initial value selection module is to provide the initial value for the convergence calculation.

[0053] Through Equations (2) to (6), the model parameter θ (i) =[b (i) a 1(i) a 2(i) …a M(i) T can be obtained for the values of each manufacturing parameter in it, and the model parameter values corresponding to the smelting targets (carbon content, temperature) are organized in Table 1.

[0054] Table 1 Model parameter values corresponding to operating parameters

[0055]

[0056] Next, the operating parameters x are optimized through a quadratic programming problem, and this optimization problem is shown in Equation (7).

[0057]

[0058] ​x LB x ≤ x ≤ UB

[0059] y LB y ≤ y ≤ UB

[0060] In the formula, x LB represents the lower limit value of the process parameter, and x UB represents the upper limit value of the process parameter. y LB represents the lower limit value of the smelting target, and y UB represents the upper limit value of the smelting target.

[0061] During the experiment, the predictive equation fitted out has been obtained in the modeling process. y 0 and c T have been informed in the previous conditions. For the upper and lower limit values of the smelting parameters and end-point composition in formula (7), they have been determined according to the actual historical data, and the upper and lower limit values of the relevant parameters are shown in Table 2.

[0062] Table 2 Upper and lower limit values of smelting process parameters

[0063]

[0064] Through the secondary optimization iteration solution of formula (7), the predicted optimal manufacturing parameter results are shown in Table 3.

[0065] Table 3 Model-predicted smelting process parameters

[0066]

[0067] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention. Any equivalent structural transformation made using the content of the specification of the present invention under the inventive concept of the present invention, or direct / indirect application in other related technical fields, is included in the patent protection scope of the present invention.

Claims

1. An intelligent online optimization method for converter steelmaking process oriented to low-carbon and low-cost smelting, characterized in that, For the converter smelting target of the current heat, applying on-site manufacturing standards and historical heat production data, instantaneously optimize the converter production process parameters of the current heat online to meet the requirements of low-carbon emissions and low-cost converter smelting; the optimization model comprehensively considers indicators including converter hot metal conditions, smelting targets, upper and lower limits of manufacturing conditions, smelting costs, and carbon emission intensity, and flexibly adjusts the importance of different indicators in the optimization process through weight coefficients; combining local online modeling and multi-objective optimization intelligent algorithms, output the best smelting and operation process parameters of the current heat to achieve intelligent online optimization of the converter steelmaking process; The construction of the optimization model includes three modules: 1) Initial value selection module: Comprehensively considering converter hot metal conditions, smelting targets, smelting costs, and carbon emission intensity, select heat samples from historical heats that are similar to the smelting target of the current heat and have lower costs and carbon emissions as the initial values for optimization; the initial value selection module is expressed as: In the formula, y n is the hot metal condition and the smelting end point of the historical heat, y 0 is the hot metal condition and the smelting target of the current heat, is a diagonal matrix for appropriately scaling the output variable, x are the converter smelting and operation process parameters, q is the weight of the deviation between the regression model fitting result and the smelting target, r is the weight of the smelting cost, represents the cost value corresponding to each process parameter, p represents the weight of the carbon emission intensity, represents the carbon emission intensity value corresponding to each process parameter; 2) Local weighted regression modeling module: According to the obtained initial values for optimization, perform local weighted regression modeling in its vicinity, and the solution process includes sample weight determination, fitting equation construction, equation parameter optimization, and loss function evaluation parts; 3) Optimal condition derivation module: Combining multi-objective conditions including converter smelting targets, upper and lower limits of manufacturing conditions, smelting costs, and carbon emission intensity, iteratively solve the optimization model until convergence is reached, and output the best smelting and operation process parameters of the current heat.

2. The intelligent online optimization method for converter steelmaking process for low-carbon and low-cost smelting according to claim 1, characterized in that, The local weighted regression modeling module is expressed as: 1) Sample weight In the formula, is the query sample and the the i-th Euclidean distance of the sample; is a hyperparameter that determines the fitting degree of the model; 2) Fitting equation Wherein, is the predicted value of the i -th output; is the constant term of the m -th input; is the m -th parameter of the i -th input; is the m -th input; 3) Equation parameters Wherein, is the m th constant term of the input; is the M th i th parameter of the input; 4) Loss function Where X is the input in matrix form, .

3. An intelligent online optimization method for converter steelmaking process for low-carbon and low-cost smelting according to claim 1, characterized in that, The optimal condition derivation module is expressed as: In the formula, represents the lower limit value of the process parameter, represents the upper limit value of the process parameter, represents the lower limit value of the smelting target, represents the upper limit value of the smelting target.

4. The intelligent online optimization method for the converter steelmaking process for low-carbon and low-cost smelting according to claim 3, characterized in that, The cost value corresponding to each process parameter is provided by the steel enterprise ERP management system, the upper and lower limit values of the process parameters and smelting targets are specified by on-site manufacturing standards, and the carbon emission intensity value is calculated based on the energy consumption and material consumption corresponding to each process parameter.

5. An intelligent online optimization method for a converter steelmaking process oriented to low-carbon and low-cost smelting, according to any one of claims 1-3, characterized in that The smelting and operation process parameters include scrap ratio, lime addition amount, dolomite addition amount, oxygen blowing amount, smelting time, deoxidizing alloy addition amount.

Images