Intelligent online optimization method for converter steelmaking processes for low-carbon, low-cost smelting

Abstract

This invention belongs to the field of steel metallurgy technology, and specifically provides an intelligent online optimization method for converter steelmaking processes for low-carbon, low-cost smelting. [Solution] By applying on-site manufacturing standards and past charge production data, and comprehensively considering indicators such as converter iron and water conditions, smelting targets, upper and lower limits of manufacturing conditions, smelting costs, and carbon emission intensity, intelligent algorithms such as local online modeling and multi-target optimization are combined to achieve intelligent online optimization of converter steelmaking and operation process parameters, improving the intelligent control level of on-site converters and contributing to the achievement of the dual goals of energy saving and emission reduction, and cost reduction and efficiency improvement in converter smelting.

Description

[Technical Field]

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

[0002] Converter low-carbon low-cost smelting is a crucial pathway for the current steel industry to pursue green transformation and economic benefits. As climate change challenges worsen worldwide, reducing carbon emissions is becoming a common concern in the international community. As a major carbon-emitting industry, the steel industry faces significant pressure to reduce emissions. At the same time, rising raw material prices and energy costs are prompting steel companies to explore low-cost production methods. Therefore, converter low-carbon low-cost smelting has become an important direction for the transformation and upgrade of the steel industry, aiming to achieve the dual goals of energy conservation and emission reduction, and cost reduction and efficiency through technological innovation and optimization.

[0003] Current optimization of converter smelting and operational process parameters in steel companies is primarily based on on-site manufacturing regulations and human experience, making it difficult to adapt to precise production control and diversified smelting objectives. Therefore, this invention proposes an intelligent online optimization method for converter steelmaking processes for low-carbon, low-cost smelting. Utilizing on-site manufacturing standards and historical charge production data from a steel company's big data platform, the method comprehensively considers numerous indicators for the converter smelting process, such as converter iron-water conditions, smelting objectives, upper and lower limits of manufacturing conditions, smelting costs, and carbon emission intensity. By combining intelligent algorithms such as local online modeling and multi-objective optimization, the method intelligently optimizes the converter steelmaking process online, achieving low-carbon emissions and low-cost smelting production, and guiding and improving the converter smelting level of on-site workers. [Overview of the Initiative]

[0004] To address the challenges of conventional technologies, the main objective of this invention is to propose an intelligent online optimization method for converter steelmaking processes aimed at low-carbon, low-cost smelting. Based on the manufacturing standards of the converter site and past charge production data, the invention comprehensively considers many indicators such as converter iron and water conditions, smelting targets, upper and lower limits of manufacturing conditions, smelting costs, and carbon emission intensity. Using intelligent algorithms such as local online modeling and multi-target optimization, the invention intelligently optimizes converter steelmaking and operational process parameters online, guiding the site towards achieving low-carbon, low-cost converter smelting production.

[0005] Specifically, this intelligent online optimization method for converter steelmaking processes for low-carbon, low-cost smelting is characterized by: instantly optimizing the converter production process parameters of the current charge online using on-site manufacturing standards and past charge production data in relation to the converter smelting target of the current charge, thereby meeting the requirements for low-carbon, low-cost converter smelting; the optimization model used comprehensively considers indicators such as converter iron and water conditions, smelting targets, upper and lower limits of manufacturing conditions, smelting costs, and carbon emission intensity; flexibly adjusting the importance of different indicators in the optimization process through weighting coefficients; combining intelligent algorithms such as local online modeling and multi-target optimization; outputting the optimal smelting and operation process parameters for the current charge; and realizing intelligent online optimization of the converter steelmaking process.

[0006] Furthermore, in building the optimization model,

[0007] 1) An initial value selection module that comprehensively considers converter iron-water conditions, smelting targets, smelting costs, and carbon emission intensity, and selects a charge sample from past charges that is close to the current charge smelting target and has relatively low costs and carbon emissions, to be used as the initial value for optimization.

[0008] 2) Based on the obtained optimization initial values, local weighted regression modeling is performed in that vicinity, and the solution method includes a local weighted regression modeling module that includes parts such as determining the sample weights, constructing the fitting equations, optimizing the equation parameters, and evaluating the loss function.

[0009] 3) Combine multi-objective conditions such as converter smelting targets, upper and lower limit values of manufacturing conditions, smelting costs, and carbon emission intensity, and iteratively solve the optimization model until convergence, including an optimal condition derivation module that outputs the optimal smelting and operation process parameters for the current charge.

[0010] Furthermore, the initial value selection module

[0011]

Number

[0012] Here, y n is the hot metal condition and smelting end point of the past charge, y 0 is the hot metal condition and smelting target of the current charge,

Number

Number

Number

[0013] Furthermore, the locally weighted regression modeling module can be expressed as follows.

[0014] 1) Sample weight

[0015]

Number

[0016] Here,

number

[0017] 2) Fitting equation

[0018]

number

[0019] Here,

number

[0020] 3) Equation parameters

[0021]

number

[0022] Here, b (i) is the mth constant term to be entered, and a M(i) This is the i-th parameter, which is the Mth input.

[0023] 4) Loss function

[0024]

number

[0025] Here, X is the matrix form of the input, [Number] is.

[0026] Furthermore, the optimal condition derivation module

[0027] [Number] can be represented as

[0028] where 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.

[0029] 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 parameter and the smelting target are defined by the 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.

[0030] Furthermore, the value range of the weights is 0.5 ≤ q ≤ 1, 0 ≤ c ≤ 1, 0 ≤ e ≤ 1.

[0031] Furthermore, the smelting and operation process parameters include the scrap ratio, lime addition amount, dolomite addition amount, oxygen injection amount, lance position, smelting time, and deoxidizing alloy addition amount.

[0032] The beneficial effects of the present invention are as follows.

[0033] This invention proposes an intelligent online optimization method for converter steelmaking processes aimed at low-carbon, low-cost smelting. Utilizing on-site manufacturing standards and historical charge production data from a steel company's big data platform, the method comprehensively considers numerous indicators for the converter smelting process, including converter iron and water conditions, smelting targets, upper and lower limits on manufacturing conditions, smelting costs, and carbon emission intensity. By combining intelligent algorithms such as local online modeling and multi-target optimization, the method intelligently optimizes the converter steelmaking process online, improving the level of intelligent control in converter smelting and contributing to the achievement of the dual goals of energy saving, emission reduction, cost reduction, and efficiency improvement in converter smelting. [Modes for carrying out the invention]

[0034] The following will be a clearer and more complete explanation of the technical concepts presented in the examples, but it is clear that the examples described are only a partial representation of the present invention and not all of its embodiments. All other embodiments that can be obtained by those skilled in the art without any creative work based on the examples of the present invention are within the scope of the protection of the present invention.

[0035] The technical solution of the present invention will be further explained below with reference to specific examples.

[0036] Examples Using silicon steel converter production at a certain steel mill as an example, historical charge production data for 2620 furnaces was collected and used for modeling. Smelting began in a certain charge converter, with iron-water conditions of [C]=4.32%, [Si]=0.41%, [Mn]=0.3%, [P]=0.016%, and T=1278℃. The target end carbon temperature for converter smelting was [C]=0.06% and T=1665℃.

[0037] According to the initial value selection formula (1), a charge sample is selected from past charges that is close to the current charge smelting target and has low cost and carbon emissions.

[0038]

number

[0039] Here, y n is the past charge iron water condition and smelting endpoint, y 0 These are the current charge conditions and smelting targets.

number

number

number

[0040] Based on the fitting results from the regression model, and setting the weights q=1 for the smelting target deviation, r=1 for the smelting cost, and e=1 for the carbon emission intensity, the charge sample for which initial values ​​were searched through the initial value selection module was determined to have the following initial values: iron / water [C]=4.35%, [Si]=0.44%, [Mn]=0.31%, [P]=0.017%, T=1290℃, converter smelting endpoint carbon [C]=0.05%, and endpoint temperature 1663℃.

[0041] By constructing a local prediction model using the local weighting modeling module and prioritizing historical data near the initial value, a linear prediction model with sufficient accuracy for the initial value can be constructed. The generated model is then used to calculate the output corresponding to the target value.

[0042] The samples are weighted using the method for locally weighted linear regression models, and the weight calculation formula is shown in equation (2).

[0043]

number

[0044] Formula (2)

number

[0045] To predict the output variable from a given feature vector as input, we use the weighted loss function shown in equation (3). Equation (3) is an additive equation that adds weights to the loss function. By minimizing the loss function, this algorithm can find the value of θ that is close to the sample distance close to the x distance and has the smallest loss. In other words, this algorithm places more emphasis on points close to x, which is also advantageous for predicting y more accurately. When equation (3) is expressed in matrix form, it can be expressed as equation (4), and the local linear regression model equation is shown as equation (5).

[0046]

number

[0047] Here,

number

[0048]

number

[0049] Finally, the module for deriving the optimal manufacturing conditions uses quadratic programming to minimize the target function under given constraints. Since the derived optimal manufacturing conditions differ from the initial conditions, the local model needs to iterate through to obtain a locally weighted regression model and the derivation of the optimal manufacturing conditions until the optimal manufacturing conditions converge. This iteration is necessary to improve the accuracy of the locally linear model around the derived optimal manufacturing conditions. The selection of an initial value module is to provide initial values ​​for the convergence calculation.

[0050] According to equations (2) to (6), the model parameter θ (i) =[b (i) a 1(i) a 2(i) …a M(i) ] T By obtaining the values ​​for each manufacturing parameter, the model parameter values ​​corresponding to the smelting targets (carbon content, temperature) can be summarized in Table 1.

[0051] [Table 1]

[0052] Next, the operational parameter x is optimized using a quadratic programming problem as shown in equation (7), which is shown in equation (7).

[0053]

number

[0054] Here, x LBx indicates the lower limit of the process parameter, UB indicates the upper limit of the process parameter, y LB indicates the lower limit of the smelting target, y UB This indicates the upper limit of the smelting target.

[0055] In the experiment, the predictive equation fitted during modeling.

number

[0056] [Table 2]

[0057] Table 3 shows the predicted optimal manufacturing parameter results obtained through the quadratic optimization iterative solution method of equation (7).

[0058] [Table 3]

[0059] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. All equivalent structural transformations performed using the content of the specification of the present invention, or those used directly or indirectly in other related fields of the art, under the concept of the present invention, are included within the scope of the patent protection of the present invention.

Claims

1. To meet the current charge's converter smelting target, the system instantly optimizes the converter production process parameters online using on-site manufacturing standards and past charge production data, satisfying the requirements for low-carbon, low-cost converter smelting. The optimization model used comprehensively considers indicators such as converter iron-water conditions, smelting targets, upper and lower limits of manufacturing conditions, smelting costs, and carbon emission intensity. It flexibly adjusts the importance of different indicators in the optimization process through weighting coefficients, and combines intelligent algorithms such as local online modeling and multi-target optimization to output the current charge's optimal smelting and operation process parameters, achieving intelligent online optimization of the converter steelmaking process. To build an optimization model, 1) Considering the converter iron-water conditions, smelting targets, smelting costs, and carbon emission intensity comprehensively, select a charge sample from past charges that is close to the current charge smelting target and has relatively low costs and carbon emissions, and use it as the initial value for optimization. [Number 24] Here, y ｎ is the past charge iron water condition and the smelting endpoint, y ０ These are the current charge conditions and smelting targets. [Number 25] is a diagonal matrix that appropriately scales the output variables, x is the converter smelting and operating process parameters, q is the weight of the fitting result from the regression model and the deviation from the smelting target, and r is the weight of the smelting cost. [Number 26] represents the cost value corresponding to each process parameter, and p represents the weight of the carbon emission intensity. [Number 27] This includes an initial value selection module that represents the carbon emission intensity value corresponding to each process parameter, 2) Based on the obtained optimization initial values, local weighted regression modeling is performed in that vicinity, and the solution method includes a local weighted regression modeling module that includes parts such as determining the sample weights, constructing the fitting equations, optimizing the equation parameters, and evaluating the loss function. 3) An intelligent online optimization method for a converter steelmaking process for low-carbon, low-cost smelting, characterized by including an optimal condition derivation module that combines multiple target conditions such as converter smelting targets, upper and lower limits of manufacturing conditions, smelting costs, and carbon emission intensity, iteratively solves an optimization model until convergence occurs, and outputs the optimal smelting and operating process parameters for the current charge.

2. The local weighted regression modeling module is represented as follows: 1) Sample weight [Number 28] Here, [Number 29] σ is the Euclidean distance between query sample x and the i-th sample, and σ is a hyperparameter that determines the degree of model fitting. 2) Fitting equation [Number 30] Here, [Number 31] is the i-th predicted value, and b （ｉ） is the m-th constant term, and a ｍ（ｉ） is the i-th parameter that is entered as the m-th, and x ｍ It is entered as the mth entry, 3) Equation parameters [Number 32] Here, b （ｉ） is the m-th constant term, and a Ｍ（ｉ） This is the i-th parameter that is entered as the Mth, 4) Loss function [Number 33] Here, X is the matrix form of the input, [Number 34] An intelligent online optimization method for a converter steelmaking process for low-carbon, low-cost smelting, as described in claim 1.

3. The optimal condition derivation module is, [Number 35] As shown, Here, 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, and is characterized by the intelligent online optimization method for the converter steelmaking process for low-carbon and low-cost smelting according to claim 1.

4. An intelligent online optimization method for a converter steelmaking process for low-carbon, low-cost smelting, as described in claim 3, characterized in that cost values ​​corresponding to each process parameter are provided by the steel company's ERP management system, upper and lower limits for process parameters and smelting targets are defined by on-site manufacturing standards, and carbon emission intensity values ​​are calculated based on energy consumption and material consumption corresponding to each process parameter.

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

Images