Method for optimizing blast furnace burden distribution matrix based on burden distribution characteristic parameters

Abstract

The invention relates to a method for optimizing a blast furnace burden distribution matrix based on burden distribution characteristic parameters, and belongs to the technical field of blast furnace ironmaking. Calculating the correlation between the characteristic parameters and the fuel ratio; the characteristic parameters with correlation larger than a preset value are screened out to serve as input variables; training a learning model by using the input variables to obtain a fuel ratio prediction model; and taking the fuel ratio prediction model as a fitness function of a genetic algorithm, and finding an optimal set value of the blast furnace burden distribution matrix characteristic parameter value through genetic optimization. According to the method, through organic combination of characteristic parameter correlation screening, fuel ratio prediction model construction and genetic algorithm global optimization, quantitative and intelligent optimization of the blast furnace burden distribution matrix parameters is achieved, the fuel ratio can be remarkably reduced, and the operation stability and economical efficiency of the blast furnace are improved.

Description

Technical Field

[0001] This invention relates to the field of blast furnace ironmaking technology, and specifically to a method for optimizing the blast furnace charging matrix based on charging characteristic parameters. Background Technology

[0002] In the blast furnace smelting process, operators use a charging matrix to control the distribution of the furnace charge, thereby optimizing the gas flow pattern, enhancing the reduction reaction, and improving energy utilization efficiency, thus reducing fuel consumption and achieving energy conservation and emission reduction. With the advancement of computer technology, numerical simulation, modeling calculation, and machine learning methods have gradually replaced traditional methods and become the main tools for charging matrix research.

[0003] Existing charge distribution matrix models are mostly constructed based on the blast furnace charge distribution equation, as shown in Figure 1:

[0004]

[0005] Where V1 is the velocity of the furnace charge at the end of the chute, l0 is the length of the chute, α is the material distribution angle, i.e., the angle between the chute and the xy plane, ω is the rotational speed of the chute around the z-axis, and r is the radius of the circle of the furnace charge in the xy plane with the z-axis as the center.

[0006] Some researchers based their work on the shape function f of the furnace throat. r The concentric circles formed are used to solve for the optimal combination of chute inclination angle α and the number of charging rings, and an optimization model of chute rotation speed ω and inclination angle α is established with the ore-to-coke ratio in the furnace throat as the optimization objective. Other scholars use the thickness of the material layer as the objective, employing particle swarm optimization or genetic algorithms to optimize the material surface profile, or using autoregressive moving average models and regularized limit learning machines to improve the gas flow distribution, thereby guiding the adjustment of the charging matrix. However, all these models are based on the blast furnace charging equation. Since the material surface is in a dynamic state in actual production, the simulated material surface established by these models often deviates significantly from reality and is difficult to accurately reflect the actual working conditions. Summary of the Invention

[0007] To address the aforementioned problems, the purpose of this invention is to provide a method for optimizing the blast furnace charging matrix based on charging characteristic parameters.

[0008] A method for optimizing the blast furnace charging matrix based on charging characteristic parameters includes:

[0009] Step 1: Select the characteristic parameters that control the blast furnace charging matrix;

[0010] Step 2: Calculate the correlation between characteristic parameters and fuel ratio;

[0011] Step 3: Select feature parameters with a correlation greater than a preset value as input variables;

[0012] Step 4: Use the input variables to train the learning model to obtain the fuel ratio prediction model;

[0013] Step 5: Use the fuel ratio prediction model as the fitness function of the genetic algorithm, and find the optimal setting value of the characteristic parameter value of the blast furnace charging matrix through genetic optimization.

[0014] Preferably, in step 1, the characteristic parameters include: coke outer ring angle, ore outer ring angle, number of coke rings, number of ore rings, coke and ore outer ring angle difference, coke angle difference, coke platform width, ore angle difference, edge ore-coke ratio, center ore-coke ratio, and ore-coke angle difference ratio.

[0015] Preferably, in step 2, the Spearman correlation coefficient is used to analyze the correlation between the characteristic parameters and the fuel ratio.

[0016] Preferably, in step 4, the learning model is an RBF network model, and the number of cluster centers of the RBF network model is 20.

[0017] Preferably, in the genetic algorithm, the initial population size is set to 100, the maximum number of generations is 40, the crossover probability is 0.6, the mutation probability is 0.01, and the number of new individuals is 40.

[0018] The present invention also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, characterized in that the computer program, when executed by the processor, implements the steps in the above-described method for optimizing a blast furnace charging matrix based on charging feature parameters.

[0019] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the steps in the above-described method for optimizing a blast furnace charging matrix based on charging feature parameters.

[0020] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0021] This invention relates to a method for optimizing the blast furnace charging matrix based on charging feature parameters. Compared with the prior art, this invention achieves quantitative and intelligent optimization of blast furnace charging matrix parameters by organically combining feature parameter correlation screening, fuel ratio prediction model construction, and global optimization using a genetic algorithm. This can significantly reduce the fuel ratio and improve the stability and economy of blast furnace operation.

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 is a schematic diagram of the blast furnace charging equation and charging process provided by the present invention; wherein, (a) is the front view and (b) is the top view;

[0025] Figure 2 The flowchart of the genetic algorithm provided by this invention;

[0026] Figure 3 shows the training results provided by the present invention; where (a) is the training result of the model and (b) is the test result of the model.

[0027] Figure 4 The diagram shows the optimization results of the genetic algorithm provided by this invention. Detailed Implementation

[0028] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0029] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0030] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0031] Please refer to Figure 1. A method for optimizing the blast furnace charging matrix based on charging characteristic parameters includes:

[0032] Step 1: Select the characteristic parameters for controlling the blast furnace charging matrix; select the following parameters: coke outer ring angle (m1), ore outer ring angle (m2), number of coke rings (m3), number of ore rings (m4), angle difference between coke and ore outer rings (m5), coke angle difference (m6), coke platform width (m7), ore angle difference (m8), edge ore-coke ratio (m9), and center ore-coke ratio (m1). 10 ), coke angle difference ratio (m) 11 A total of 11 parameters were used as the initial feature set.

[0033] Step 2: Calculate the correlation between characteristic parameters and fuel ratio;

[0034] In step 2, the Spearman correlation coefficient is used to analyze the correlation between each parameter and the fuel ratio, and key variables are selected as model inputs to reduce model complexity. The calculation formula is as follows:

[0035]

[0036] Where ρ is the Spearman correlation coefficient; x i y i Samples of input and output variables; The mean of the input and output variables.

[0037] Step 3: Select feature parameters with a correlation greater than a preset value as input variables;

[0038] Data with |ρ|≥0.20 are considered as variables with good correlation and used as key input variables to establish a fuel ratio prediction model.

[0039] Step 4: Use the input variables to train the learning model to obtain the fuel ratio prediction model;

[0040] In step 4, when the number of cluster centers is selected as 20, the fuel ratio prediction model has good accuracy. Based on production experience, a fuel ratio RBF network prediction model is established. The performance of the RBF network training model and the test model is quantitatively evaluated using the average relative error and hit rate. When the performance meets the set conditions, the training is completed.

[0041] RBF network: Radial Basis Functions Neural Network, is a three-layer feedforward structure that features rapid training, global convergence, and good nonlinear fitting capabilities. Its hidden layers use radial basis functions to achieve a nonlinear mapping from the input to the hidden layers, while the output layer is a linear combination.

[0042] (1) Hidden layer basis functions:

[0043]

[0044] Where i = 1, 2, 3, ..., m; p = 1, 2, 3, ..., n; r—number of hidden layer nodes; Xp—the p-th input sample; ci—center of the i-th hidden node; σ—shape parameter of the i-th center vector; ||X p -c i ||2—represents the Euclidean distance between Xp and ci.

[0045] (2) Network output:

[0046]

[0047] Where j = 1, 2, 3, ..., s; wij—output layer weights; bj—output layer thresholds.

[0048] Step 5: Use the fuel ratio prediction model as the fitness function of the genetic algorithm, and find the optimal setting value of the characteristic parameter value of the blast furnace charging matrix through genetic optimization.

[0049] In step 5, the initial population size is 100, the maximum number of generations is 40, the crossover probability is 0.6, the mutation probability is 0.01, and the number of new individuals is 40. The trained RBF network model is used as the fitness function of the genetic algorithm. Through genetic optimization, the optimal settings of the key feature parameters of the cloth matrix are found.

[0050] Please see Figure 2 The steps of the genetic algorithm are as follows:

[0051] (1) The parent population P of size N t With offspring population Q t Merged into a population R of size 2N t Perform a non-dominated sort to obtain the non-dominated solution set Z1–Z n And calculate individual crowding levels;

[0052] (2) Sort by non-dominant level and aggregation distance, and select individuals in sequence to fill the new parent population P. t+1 , until its size reaches N;

[0053] (3) Generate new offspring Q through selection, crossover and mutation operations. t+1 ;

[0054] (4) Merge P t+1 With Q t+1 For R t+1 Re-sort the non-dominated items;

[0055] (5) Iterate through steps (1) to (4) until the maximum number of iterations is reached.

[0056] Table 1 provides an example of Spearman correlation coefficient calculation, listing 4000m... 3 The Spearman correlation coefficients of 11 charging parameters and fuel ratio in a blast furnace were calculated.

[0057] Table 1. Calculation results of Spearman correlation coefficient

[0058]

[0059] Based on the standard of |ρ|≥0.20, six key parameters were selected: coke outer ring angle, coke-ore outer ring angle difference, coke angle difference, ore angle difference, edge ore-coke ratio, and center ore-coke ratio. These parameters respectively control the size of the central funnel, the area of ​​the ore-free zone, and the amount of central coke, and have clear physical significance.

[0060] A fuel ratio (RBF) prediction model was established based on the above six parameters, with the number of cluster centers set to 20.

[0061]

[0062] From the 265 preprocessed data sets, 240 sets were selected as the training set and 25 sets as the test set. The results are shown in Figure 3. The model performance metrics are shown in Table 2.

[0063] Table 2 Model Performance Indicators

[0064] Parameter name Relative error / % Mean square error / % Root mean square error / % Hit probability / % <![CDATA[Fuel ratio / kg·t -1 > 0.01 88.23 9.41 88.12

[0065] The model has a very small relative error, a high hit rate, and good performance. Some prediction biases are caused by blast furnace shutdowns or abnormal furnace conditions. While correcting these as outliers can improve the fit, it may weaken the model's ability to reflect actual operating conditions.

[0066] Using the RBF model as the fitness function, the genetic algorithm parameters were set and optimized, and the results are as follows: Figure 4 As shown. The fuel-to-total ratio reached a minimum of 520.11 kg·t when evolved to the third generation.-1 The corresponding optimized values ​​for key parameters are shown in Table 3.

[0067] Table 3 Optimization results of input variables

[0068]

[0069] It is evident that the optimized fabric matrix parameters have solved the problems of easy melting, easy solidification, and difficult remelting of minerals, thereby promoting a reduction in fuel ratio.

[0070] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0071] 1. This invention uses correlation analysis to extract parameters with correlation coefficients ρ greater than a certain standard (|ρ|≥0.20) from 11 material distribution matrix parameters as key feature parameters of the material distribution matrix to establish RBF network training and testing models. Then, a genetic algorithm is applied for optimization to obtain the set values ​​of the key feature parameters of the material distribution matrix. After adjusting the material distribution matrix according to the set values, the blast furnace fuel ratio is reduced by 5.58 kg / tHTM.

[0072] 2. After the application of this invention, the fuel cost of a certain blast furnace from January to August 2025 decreased by 5.58 kg / tHTM compared to 2024 (of which: the coke ratio decreased by 5.02 kg / tHTM, and the coal ratio decreased by 0.56 kg / tHTM). Based on a coke price of 1661.73 yuan / t, a pulverized coal price of 862.77 yuan / t, and an annual output of 3.02 million tons, the fuel cost reduction in 2025 is calculated as (5.020 / 1000×1661.73+0.565 / 1000×862.77)×302=26.66 million yuan.

[0073] 3. After the application of this invention, the fuel consumption of a certain blast furnace of Baogang in January-August 2025 was reduced by 5.58 kg / tHTM compared with 2024. The average carbon content of the fuel was 81.33%. Based on a carbon atomic weight of 12 g / mol, a CO2 molecular weight of 44 g / mol, and an annual output of 3.02 million tons, the CO2 emissions in 2025 are expected to decrease by (5.58 × 81.33% / 44 / 12) × 3.02 / 1000 = 50,300 tons.

[0074] The present invention also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor. The transceiver, the memory, and the processor are connected via the bus. The computer program, when executed by the processor, implements the steps in the aforementioned method for optimizing a blast furnace charging matrix based on charging characteristic parameters. Compared with the prior art, the beneficial effects of the electronic device provided by the present invention are the same as those of the aforementioned method for optimizing a blast furnace charging matrix based on charging characteristic parameters, and will not be elaborated upon here.

[0075] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps in the above-described method for optimizing a blast furnace charging matrix based on charging feature parameters. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present invention are the same as the beneficial effects of the above-described method for optimizing a blast furnace charging matrix based on charging feature parameters, and will not be elaborated here.

[0076] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for optimizing the blast furnace charging matrix based on charging characteristic parameters, characterized in that, include: Step 1: Select the characteristic parameters that control the blast furnace charging matrix; Step 2: Calculate the correlation between characteristic parameters and fuel ratio; Step 3: Select feature parameters with a correlation greater than a preset value as input variables; Step 4: Use the input variables to train the learning model to obtain the fuel ratio prediction model; Step 5: Use the fuel ratio prediction model as the fitness function of the genetic algorithm, and find the optimal setting value of the characteristic parameter value of the blast furnace charging matrix through genetic optimization.

2. The method for optimizing the blast furnace charging matrix based on charging characteristic parameters according to claim 1, characterized in that, In step 1, the characteristic parameters include: coke outer ring angle, ore outer ring angle, number of coke rings, number of ore rings, coke and ore outer ring angle difference, coke angle difference, coke platform width, ore angle difference, edge ore-coke ratio, center ore-coke ratio, and ore-coke angle difference ratio.

3. The method for optimizing the blast furnace charging matrix based on charging characteristic parameters according to claim 2, characterized in that, In step 2, the Spearman correlation coefficient is used to analyze the correlation between the characteristic parameters and the fuel ratio.

4. The method for optimizing the blast furnace charging matrix based on charging characteristic parameters according to claim 3, characterized in that, In step 4, the learning model is an RBF network model, and the number of cluster centers in the RBF network model is 20.

5. The method for optimizing the blast furnace charging matrix based on charging characteristic parameters according to claim 4, characterized in that, In the genetic algorithm, the initial population size is set to 100, the maximum number of generations is 40, the crossover probability is 0.6, the mutation probability is 0.01, and the number of new individuals is 40.

6. An electronic device comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, characterized in that, When the computer program is executed by the processor, it implements the steps in the method for optimizing the blast furnace charging matrix based on charging characteristic parameters as described in any one of claims 1-5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps in the method for optimizing the blast furnace charging matrix based on charging characteristic parameters as described in any one of claims 1-5.

Images