Accurate coking coal blending method and system

Abstract

The invention relates to the technical field of coal coking, in particular to a precise coking coal blending method and system.The method comprises the steps that indexes of single coking coal are obtained to establish a single coking coal resource information database; based on the single coking coal resource information database, establishing a blended coal quality prediction model according to the linear model and the SVR fusion model to obtain a blended coal quality index; based on the blended coal quality index and the historical coking process data set, establishing a coke quality prediction model through a KAN neural network to obtain a coke quality index; on the basis of the single coal quality parameters and the coking process parameters, establishing a sampling ratio optimization model by taking the expected mixed coal quality index, the expected coke quality index and the stock as constraint conditions, so as to obtain candidate ratio schemes; based on the blended coal quality prediction model, the coke quality prediction model and the blended coal cost, the candidate matching scheme is screened to obtain the optimal coal blending scheme, and the problem that the coking coal blending method is not accurate is solved.

Description

Technical Field

[0001] The present application relates to the technical field of coal coking, and particularly to a precise coking coal blending method and system. Background Art

[0002] Coking coal is an important raw material for steel enterprises to produce coke, and the stability of coke quality is crucial for blast furnace ironmaking. Coke quality is determined by the coking process, and coking coal blending is a key action in the coking process, which is the process of preparing coke by mixing different types and proportions of coal. Common coal types include fat coal, coking coal, lean coal, and 1 / 3 coking coal, etc. Reasonable proportioning of these coal types can optimize the physical and chemical properties of coke and meet the needs of different industrial production.

[0003] The advantages and disadvantages of coal blending technology directly affect coke quality and cost, which is a key problem that coking enterprises urgently need to solve. Traditional coal blending methods rely on the experience of coal blenders and lack systematic scientific guidance, resulting in low production efficiency, long calculation cycle, unstable coke quality, and it is difficult to convert experience into data, making it impossible for employees to master coal blending technology in a short time. The complexity and variability of the production site make it difficult for empirical coal blending to quickly adapt to demand changes, resulting in frequent adjustments and increased costs.

[0004] With the upgrading of technology and the improvement of market demand, enterprises have put forward higher requirements for coke quality and its stability. The coal blending methods in related technologies mainly focus on coke quality prediction, but there are still inaccurate problems in coal blending even when only meeting coke quality prediction. Summary of the Invention

[0005] The present application provides a precise coking coal blending method and system to solve the problem of inaccurate coking coal blending method.

[0006] The first aspect of the present application provides a precise coking coal blending method, and the method includes: Obtain the indexes of single coking coal to establish a single coking coal resource information database; Based on the single coking coal resource information database, establish a blended coal quality prediction model according to the linear model and the SVR fusion model to obtain the blended coal quality indexes; Based on the blended coal quality indexes and the historical coking process data set, establish a coke quality prediction model through the KAN neural network to obtain the coke quality indexes; Based on the single coal quality parameters and coking process parameters, and taking the expected blended coal quality indexes, expected coke quality indexes and inventory as constraint conditions, establish a sampling ratio optimization model to obtain a candidate ratio plan; Based on the blended coal quality prediction model, the coke quality prediction model and the blended coal cost, screen the candidate ratio plans to obtain the optimal coal blending plan.

[0007] The above method first analyzes the production data to find the correlation between the quality indicators of coke and the quality indicators of single coals, so as to establish a coke quality prediction model; then, according to the requirements of coke quality, the requirements of blended coal quality, and the constraints such as the quality, price, and inventory of each single coal, the single coals to be blended and the optimal ratio of each coal are obtained to obtain an optimal coal blending plan. This method can not only intelligently predict the quality indicators of coke according to the characteristics and ratios of each single coal, but also intelligently and quickly calculate the optimal coal blending plan according to the actual production demand and the characteristics of each single coal, thereby improving the quality stability and production efficiency of coke, reducing production costs, and solving the problem of inaccurate coking coal blending methods.

[0008] Optionally, the indicators of the single coking coal include: the maximum thickness of the plastic layer, the caking index, the coal petrographic index of the single coal, and the working score data.

[0009] Optionally, the steps of establishing a blended coal quality prediction model based on the single coking coal resource information database and according to the linear model and the SVR fusion model include: Generating a single coal derivative feature MBI through the single coking coal resource information database; Based on the single coking coal resource information database, the single coal derivative feature MBI, and the coal blending list, and establishing a blended coal quality prediction model according to the linear model and the SVR fusion model.

[0010] The above method generates a single coal derivative feature MBI through the single coking coal resource information database, which can deeply explore the coal quality characteristics of single coals and extract comprehensive indicators that can better reflect the key characteristics of single coals in the coking process. Using a linear model and an SVR fusion model to construct a blended coal quality prediction model can give full play to the advantages of the two models. It can not only quickly capture the linear relationship in the data, but also map the non-linear problem in the low-dimensional input space to the high-dimensional feature space by using a kernel function, so that the data can be fitted by a linear model in the high-dimensional feature space, effectively processing the data that originally shows a non-linear relationship in the low-dimensional space, thereby effectively fitting these non-linear relationships and improving the prediction accuracy of the model.

[0011] Optionally, the steps of establishing a coke quality prediction model based on the blended coal quality indicators and the historical coking process data set and through the KAN neural network include: Screening out historical coking process data samples in the historical coking process data set; the historical coking process data samples include the coking process condition parameters corresponding to the blended coal produced each time and the historical coke quality indicators; The robust normalization method is used to perform variable scaling on the blended coal quality indexes and coking process condition parameters to obtain a model feature sample set; Based on the model feature sample set and the historical coke quality indexes, a coke quality prediction data set is established; A coke quality prediction model is established through the coke quality prediction data set and the KAN neural network.

[0012] The above method can reduce the influence of outliers through sample screening and robust normalization, effectively handle the differences in variable dimensions, balance the variable contributions, reduce the influence of noise, and improve the model stability. By establishing a coke quality prediction model through the coke quality prediction data set and the KAN neural network, it can automatically learn the internal laws from the data, and can adapt to different data sets and problems by adjusting the network structure and parameters, with good flexibility and adaptability to achieve accurate prediction of coke quality.

[0013] Optionally, the input variables of the coke quality prediction model include caking index, volatile matter, moisture, ash content, sulfur content, coking time, coke side temperature, and ash composition; the outputs of the coke quality prediction model include ash content, sulfur content, shatter strength, abrasion resistance strength, reactivity index, and post-reaction strength of coke.

[0014] The input variables and outputs of the coke quality prediction model can accurately consider coal quality, coking process, and ash composition, comprehensively evaluate coke quality, meet the requirements of different application scenarios, and provide strong support for coking process optimization and quality control.

[0015] Optionally, the steps to obtain the optimal coal blending plan include: A sampling ratio optimization model is established with the inventory of each single coal and the sum of all ratios being 1 as the constraint conditions to obtain a candidate ratio plan; The candidate ratio plan, single coal quality parameters, and coking process parameters are input into the blended coal quality prediction model and the coke quality prediction model to screen out the feasible solutions that meet the blended coal quality constraints and coke quality constraints; Calculate the prices of the feasible solutions, and screen out the optimal coal blending plan with the lowest blended coal cost as the objective function.

[0016] The above method not only completely solves the problem that the sum of ratios is not 1, intelligently and quickly calculates the optimal ratio that meets the constraints, can more quickly and accurately find the solution with the sum of ratios being 1 and meeting the inventory upper limit, especially for the case where the inventory ratio is slightly greater than the coal blending requirement. It can also meet the selection of various ratio calculation precisions, and on the premise of matching the work requirements of coal blending engineers, can compare the changes in blended coal quality, coke quality, and cost under different calculation precisions.

[0017] Optionally, the calculation formula of the objective function is: ; where n is the total amount of single coals participating in coal blending; is the cost price of the th single coal; is the proportion of the th single coal.

[0018] Optionally, the calculation formula of the sampling ratio optimization model includes: ; where n is the total amount of single coals participating in coal blending; C is the cost price of each single coal; is the proportion of each single coal; W is the blended coal quality index; V is the coke quality index; is the proportion of the th single coal; h j is the inventory ratio of the th single coal.

[0019] Optionally, the blended coal quality constraints include the caking index, plastic layer thickness, and volatile matter of the blended coal index; the coke quality constraints include ash content, sulfur content, shatter strength, abrasion resistance, reactivity index, and post-reaction strength.

[0020] By strictly controlling the caking index, plastic layer thickness, and volatile matter of the blended coal, it can ensure that the coal can be fully softened, melted, and solidified during the coking process, thereby improving the cold strength, thermal stability, and overall quality of the coke. By restricting the ash content and sulfur content of the coke, the impurity content can be reduced, and the harmful components in the coke can be decreased. The blended coal quality constraints and the coke quality constraints also help to guide the design of the coal blending ratio, making the selection of raw coal more scientific and reasonable, and then optimizing the coking process parameters and improving production efficiency. By controlling indicators such as shatter strength, abrasion resistance, reactivity index, and post-reaction strength, it is beneficial to ensure the behavior of the coke during the blast furnace smelting process, reduce powder generation, improve the permeability of the burden, and increase the smelting efficiency. By controlling the volatile matter and sulfur content, it helps to reduce pollutant emissions during the coking process, especially the emissions of sulfides and nitrogen oxides, to meet environmental protection requirements and reduce the environmental liability risk of enterprises.

[0021] The second aspect of this application provides a precise coking coal blending system, which is applicable to the precise coking coal blending method described in the first aspect. The system includes: a resource information database construction module, which is used to obtain the indexes of single coking coals to establish a single coking coal resource information database; The blended coal quality prediction model construction module is used to establish a blended coal quality prediction model based on the single coking coal resource information database and according to the linear model and the SVR fusion model, so as to obtain the blended coal quality index; The coke quality prediction model construction module is used to establish a coke quality prediction model based on the blended coal quality index and the historical coking process data set and through the KAN neural network, so as to obtain the coke quality index; The ratio optimization module is used to establish a sampling ratio optimization model based on the single coal quality parameters and the coking process parameters and with the expected blended coal quality index, the expected coke quality index and the inventory as constraint conditions, so as to obtain a candidate ratio plan; screen the candidate ratio plan based on the blended coal quality prediction model, the coke quality prediction model and the blended coal cost, so as to obtain the optimal coal blending plan.

[0022] The above system first analyzes the production data to find out the correlation between the quality indexes of coke and the quality indexes of single coal, so as to establish a coke quality prediction model; then, according to the requirements of coke quality, the requirements of blended coal quality and the constraints of the quality, price and inventory of each single coal, the single coal to be blended and the optimal ratio of each coal are obtained, so as to obtain the optimal coal blending plan. This method can not only intelligently predict the quality indexes of coke according to the characteristics and ratio of each single coal type, but also intelligently and quickly calculate the optimal coal blending plan according to the actual production demand and the characteristics of each single coal type, so as to improve the quality stability and production efficiency of coke, reduce the production cost, and solve the problem of inaccurate coking coal blending method.

[0023] As can be seen from the above technical solutions, the present application provides a precise coking coal blending method and system. By obtaining the indexes of single coking coal, a single coking coal resource information database is established; based on the single coking coal resource information database, a blended coal quality prediction model is established according to the linear model and the SVR fusion model, so as to obtain the blended coal quality index; based on the blended coal quality index and the historical coking process data set, a coke quality prediction model is established through the KAN neural network, so as to obtain the coke quality index; based on the single coal quality parameters and the coking process parameters, a sampling ratio optimization model is established with the expected blended coal quality index, the expected coke quality index and the inventory as constraint conditions, so as to obtain a candidate ratio plan; screen the candidate ratio plan based on the blended coal quality prediction model, the coke quality prediction model and the blended coal cost, so as to obtain the optimal coal blending plan, and solve the problem of inaccurate coking coal blending method. Description of the Drawings

[0024] To more clearly illustrate the technical solutions of this application, the following will briefly introduce the drawings required in the embodiments. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0025] Figure 1 It is a schematic flow chart of the precise coking coal blending method described in the embodiments of this application; Figure 2 It is a schematic flow chart of establishing a blended coal quality prediction model in the precise coking coal blending method described in the embodiments of this application; Figure 3 It is a schematic flow chart of establishing a coke quality prediction model in the precise coking coal blending method described in the embodiments of this application; Figure 4 It is a schematic flow chart of the calculation of the sampling ratio optimization model in the precise coking coal blending method described in the embodiments of this application. Detailed implementation manners

[0026] The following will describe the embodiments in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following embodiments do not represent all implementation manners consistent with this application. They are only examples of systems and methods consistent with some aspects of this application.

[0027] To solve the problem of inaccurate coking coal blending methods, refer to Figure 1 , some embodiments of this application provide a precise coking coal blending method, and the method includes: S100: Obtain the indexes of single coking coals to establish a resource information database of single coking coals.

[0028] It should be understood that by inputting the indexes of the single coking coals into the resource information database through multi-system docking, a resource information database of single coking coals can be established. The resource information database of single coking coals realizes the centralized management of information, can store the information orderly, and is convenient for querying, calling, and updating the indexes of the single coking coals, providing a solid data foundation for subsequent analysis and model establishment.

[0029] In some embodiments, the indexes of the single coking coals include: the maximum thickness of the plastic layer, the caking index, the coal petrographic indexes of the single coal, and the working score data.

[0030] It should be understood that when bituminous coal is heated, a viscous liquid and semi-liquid substance will form on its surface, which is called the plastic body. When the temperature further rises, the plastic body will solidify and form a solid substance, that is, coke. The maximum thickness of the plastic layer (Y value) refers to the maximum thickness of the plastic body formed by bituminous coal under specific conditions. The maximum thickness of the plastic layer of different coal types varies significantly. For example, fat coal has a large amount of plastic body and high fluidity, and the Y value is usually relatively high, generally between 20 mm and 40 mm; while the Y value of gas coal and coking coal is relatively low. The Y value of gas coal may be below 15 mm, and the Y value of coking coal is between 10 mm and 25 mm. A relatively thick plastic layer usually means that the coal has good caking property, and good caking property helps to form coke with a stable structure.

[0031] The caking index (G value) is an important indicator to measure the caking ability of coal, which reflects the ability of coal to form coke during heating. Specifically, after mixing a certain amount of coal sample with special anthracite evenly, it is heated in a specific experimental device, and then the caking index is calculated according to the strength of the obtained coke. Coal with a high caking index can better bind together to form massive coke, which improves the mechanical strength and stability of the coke and is beneficial to subsequent processes such as blast furnace smelting.

[0032] The coal petrographic index (Rmax) of single coal refers to the maximum swelling pressure that single coal can reach during heating, and it is one of the important indicators reflecting the swelling and plasticity of coal. By measuring the swelling pressure of coal at different temperatures, a swelling pressure curve can be drawn to determine the value of Rmax. By understanding the Rmax of different single coals, the differences in the metamorphic degrees of various coals can be accurately grasped, so as to reasonably adjust the proportion of different coal types to make the blended coal reach the best quality. For example, when it is necessary to produce coke with specific strength and quality, appropriate coking coal and steam coal can be selected according to the Rmax of single coal for matching to ensure the stable quality of coke. Rmax can also predict the quality indicators such as the strength and reactivity of coke to a certain extent. Generally speaking, within a certain range of Rmax, the reactivity of coke is relatively low and the mechanical strength is relatively high. A corresponding mathematical model or empirical formula can be established through the Rmax of single coal to predict the quality of coke after coal blending.

[0033] Proximate analysis data of coal refers to the industrial analysis data of coal, including the contents of components such as moisture, ash, volatile matter and fixed carbon. These data are crucial for evaluating the quality and applicability of coal. For example, low ash content and appropriate volatile matter content are usually characteristics of high-quality coking coal.

[0034] S200: Based on the single coking coal resource information database, a blended coal quality prediction model is established according to the linear model and the SVR fusion model to obtain the blended coal quality index.

[0035] In some embodiments, referring to Figure 2 , the steps of establishing a blended coal quality prediction model based on the single coking coal resource information database and according to a linear model and an SVR fusion model include: S210: Generate a single coal-derived feature MBI through the single coking coal resource information database.

[0036] It should be understood that the single coal-derived feature MBI (Mineral Base Index) is used to characterize the content of alkaline metals in coal ash. The higher the MBI value, the stronger the reactivity (CRI) of coke usually is, because the alkaline metals in coal ash have a catalytic effect on the reaction of coke with carbon dioxide. As the MBI value increases, the structure of coke will be damaged to a greater extent under the state of being eroded by carbon dioxide and alkali metals, resulting in a decrease in the strength after reaction (CSR). When the MBI value is large, the internal structure of coke is more damaged during the coking process, and the wear resistance of coke decreases, manifested as an increase in the abrasion strength of coke. When the MBI value is large, the shatter strength of coke usually decreases because alkaline metals will penetrate into the carbon structure of coke, deforming and cracking the carbon structure. Therefore, coke with a lower MBI value also has relatively low reactivity. When used in a blast furnace, it can better withstand high temperatures and chemical erosion, maintain high strength and stability, which is conducive to the smooth operation of the blast furnace and the extension of the service life of coke. Coking enterprises can formulate a more scientific and reasonable coal blending plan according to the MBI values of different single coals, combined with other coal quality indicators. By adjusting the coal blending ratio, reducing the MBI value of the blended coal, improving the quality of coke, and at the same time, resources can be reasonably utilized and production costs can be reduced.

[0037] The calculation formula of the single coal-derived feature MBI can be: .

[0038] Among them, ω(A d ） is the ash content in the single coal; ω(K 2 O） is the potassium oxide content in the single coal; ω(Na 2 O） is the sodium oxide content in the single coal; ω(CaO) is the calcium oxide content in the single coal; ω(MgO) is the magnesium oxide content in the single coal; ω(Fe 2 O 3 ） is the iron(III) oxide content in the single coal; ω(V daf） is the content of volatile matter in single coal; ω (SiO 2 ） is the content of silicon dioxide in single coal; ω(Al 2 O 3 ） is the content of aluminum oxide in single coal.

[0039] S220: Based on the single coking coal resource information database, the derived characteristics MBI of single coal and the coal blending list, and according to the linear model and the SVR fusion model, a prediction model for the quality of blended coal is established.

[0040] It should be understood that in the process of coke quality prediction, it is necessary to predict the quality index of blended coal through the quality index and ratio of single coal, and then predict the quality index of coke. The roles of different coal types in blended coal are different. The coking property of blended coal depends on the performance and proportion of each single coal. The coking property of blended coal is the result of the comprehensive action of the coking properties of each single coal. Give full play to the characteristics of various coals, learn from each other's strengths and make up for each other's weaknesses to improve the coke quality. The linear model is used to predict the quality of blended coal, but the quality of the same batch of single coal is prone to fluctuations, and it is necessary to correct the linear model according to the test results of blended coal. Therefore, it is necessary to fuse the linear model and the SVR (Support Vector Regression) model to predict the quality of blended coal and improve the prediction accuracy of the quality of blended coal.

[0041] The calculation formula of the linear model is: ; where is the vector of dependent variables, that is, the observed values; is the observation matrix of independent variables, and each column is an independent variable; is the vector containing the parameters to be estimated usually; is the error term, representing the influence of other factors on the dependent variable except the independent variable.

[0042] The calculation formula of the SVR fusion model is: ; where and are Lagrange multipliers; is the sum function, and b is the bias term.

[0043] The above method generates the single - coal derivative feature MBI through the single - coking - coal resource information database, which can deeply explore the coal quality characteristics of single - coal and extract a comprehensive index that can better reflect the key characteristics of single - coal in the coking process. By constructing a blended - coal quality prediction model using a linear model and an SVR fusion model, the advantages of both models can be comprehensively utilized. It can not only quickly capture the linear relationship in the data but also map the non - linear problems in the low - dimensional input space to the high - dimensional feature space by using kernel functions, enabling the use of a linear model to fit the data in the high - dimensional feature space, effectively dealing with the data that originally shows non - linear relationships in the low - dimensional space, thus effectively fitting these non - linear relationships and improving the prediction accuracy of the model.

[0044] S300: Based on the blended - coal quality index and the historical coking process data set, and through the KAN neural network, a coke quality prediction model is established to obtain the coke quality index.

[0045] It should be understood that the KAN neural network is an innovative neural network architecture. Different from traditional neural networks that use fixed activation functions, it can use learnable activation functions at the edges of the network. This design enables each weight parameter in the KAN to be replaced by a univariate function, which is usually parameterized in the form of a spline function, thus providing extremely high flexibility and being able to simulate complex functions with fewer parameters. It is not only efficient and accurate but also enhances the interpretability of the model. Therefore, when using the KAN neural network to predict the coke quality index, the calculation formula of the KAN neural network can be expressed by the following formula: ; Among them, and are univariate continuous functions, and The calculation formulas of ; .

[0046] Among them, the B - spline mechanism is the most important core learning mechanism of the KAN neural network. They replace the traditional weight parameters often used in neural networks. Their flexibility enables them to adaptively fit the complex relationships in the data by adjusting their shapes, thereby minimizing the approximation error and enhancing the network's ability to learn subtle patterns from high - dimensional data sets. The B - spline can be expressed by the formula: ; Among them, are the coefficients optimized during training, are the B - spline basis functions defined on the grid, which are hyperparameters affecting the network accuracy.

[0047] In some embodiments, the input variables of the coke quality prediction model include adhesion index, volatile matter, moisture, ash content, sulfur content, coking time, coke side temperature and ash composition; the output of the coke quality prediction model includes the ash content, sulfur content, crushing strength, abrasion resistance, reactivity index and post-reaction strength of the coke.

[0048] It should be understood that the cohesiveness index in the input is directly related to whether the coal can form coke of sufficient strength during the coking process; the volatile matter, moisture, ash and sulfur content reflect the composition of the coal, which has an important impact on the yield, quality and furnace life of the coke. The coking time determines the residence time of the coal in the coke oven, which directly affects the maturity and structure of the coke; the coke side temperature affects the pyrolysis and solidification process of the coal, and plays a key role in the properties of the coke. By considering the ash composition, the quality characteristics of the coke can be predicted more accurately, which can provide more accurate results. The ash and sulfur content in the output are important indicators to measure the purity of the coke, which directly affects the application effect of the coke in blast furnace smelting; the crushing strength and wear resistance reflect the ability of the coke to resist crushing and wear during transportation and use; the reactivity index and post-reaction strength are closely related to the behavior of the coke in the chemical reaction, which has an important impact on the reaction process and efficiency in the blast furnace to ensure the normal operation and efficient production of the blast furnace.

[0049] The input variables and outputs of the coke quality prediction model can accurately consider coal quality, coking process and ash composition, comprehensively evaluate coke quality, meet the needs of different application scenarios, and provide strong support for coking process optimization and quality control.

[0050] In some embodiments, see Figure 3 The steps of establishing a coke quality prediction model based on the blended coal quality index and the historical coking process data set by using a KAN neural network include: S310: Filter out historical coking process data samples from the historical coking process data set.

[0051] The historical coking process data samples include the coking process condition parameters and historical coke quality indicators corresponding to each production of the blended coal. Due to the presence of various interference factors in the industrial field, it is necessary to remove some samples that are obviously inconsistent with the physical meaning from the historical coking process data set so that the historical coking process data samples do not contain abnormal data.

[0052] S320: Using a robust normalization method to perform variable scaling on the blended coal quality index and coking process condition parameters to obtain a model feature sample set.

[0053] It should be understood that the robust normalization method is a normalization technique used to handle outliers in data preprocessing, aiming to make the data insensitive to outliers. Due to the different dimensions in the blended coal quality indexes and coking process condition parameters, the sizes of different variable data vary greatly. At the same time, the data distribution ranges are different, which will exaggerate the influence of some variables on the target and mask the contributions of some variables. Also, due to the existence of some noise in the data, directly applying the min-max normalization and standardization methods will also affect the accuracy of the model. Therefore, the robust normalization method can be used to perform variable scaling on the blended coal quality indexes and coking process condition parameters, which can effectively reduce the influence of outliers on the normalization result and better maintain the distribution characteristics of the data.

[0054] S330: Establish a coke quality prediction data set based on the model feature sample set and the historical coke quality indexes.

[0055] S340: Establish a coke quality prediction model through the coke quality prediction data set and the KAN neural network.

[0056] It should be understood that the coke quality prediction uses the KAN neural network for prediction, with the blended coal quality indexes and coking process condition parameters as inputs and the coke quality indexes as outputs.

[0057] The above method can reduce the influence of outliers through sample screening and robust normalization, effectively handle the differences in variable dimensions, balance the contributions of variables, reduce the influence of noise, and improve the stability of the model. Establishing a coke quality prediction model through the coke quality prediction data set and the KAN neural network can automatically learn the internal laws from the data, and can adapt to different data sets and problems by adjusting the network structure and parameters, with good flexibility and adaptability to achieve accurate prediction of coke quality.

[0058] S400: Based on the single coal quality parameters and coking process parameters, and with the expected blended coal quality indexes, expected coke quality indexes and inventory as constraint conditions, establish a sampling ratio optimization model to obtain a candidate ratio plan.

[0059] It should be understood that the coking process parameters include the coking time and the coke side temperature.

[0060] S500: Screen the candidate ratio plans based on the blended coal quality prediction model, the coke quality prediction model and the blended coal cost to obtain the optimal coal blending plan.

[0061] The above method first analyzes the production data to find the correlation between the quality indexes of coke and the quality indexes of single coals, so as to establish a coke quality prediction model; then, according to the requirements of coke quality, the requirements of blended coal quality, and the constraints such as the quality, price, and inventory of each single coal, the single coals to be blended and the optimal ratio of each coal are obtained to obtain an optimal coal blending plan. This method can not only intelligently predict the quality indexes of coke according to the characteristics and ratios of each single coal, but also intelligently and quickly calculate the optimal coal blending plan according to the actual production requirements and the characteristics of each single coal, thereby improving the quality stability and production efficiency of coke, reducing the production cost, and solving the problem of inaccurate coke blending method.

[0062] In some embodiments, the steps of obtaining the optimal coal blending plan include: Taking the sum of the inventory of each single coal and all ratios as 1 as a constraint condition to establish a sampling ratio optimization model to obtain a candidate ratio plan.

[0063] Inputting the candidate ratio plan, the quality parameters of single coals, and the coking process parameters into the blended coal quality prediction model and the coke quality prediction model to screen out the feasible solutions that meet the blended coal quality constraint and the coke quality constraint.

[0064] Calculating the price of the feasible solutions, and screening out the optimal coal blending plan with the lowest blended coal cost as the objective function.

[0065] The above method not only completely solves the problem that the sum of ratios is not 1, intelligently and quickly calculates the optimal ratio that meets the constraints, and can more quickly and accurately find the solution with the sum of ratios being 1 and meeting the inventory upper limit, especially for the case where the inventory ratio is slightly greater than the coal blending requirement. It can also meet the selection of multiple ratio calculation precisions, and on the premise of matching the work requirements of coal blending engineers, it can compare the changes in the quality of blended coal, the quality of coke, and the cost under different calculation precisions.

[0066] In some embodiments, the calculation formula of the objective function is: ; where n is the total amount of single coals participating in coal blending; is the cost price of the th single coal; is the ratio of the th single coal.

[0067] In some embodiments, the calculation formula of the sampling ratio optimization model includes: ; where n is the total amount of single coals participating in coal blending; C is the cost price of each single coal; is the proportion of each single coal; W is the quality index of blended coal; V is the quality index of coke; is the proportion of the th single coal; hj is the inventory ratio of the th single coal.

[0068] It should be understood that V min is the minimum value of the coke quality index ，V max is the maximum value of the coke quality index ，W max is the maximum value of the blended coal quality index ，W min is the minimum value of the blended coal quality index.

[0069] See Figure 4 , the main steps of the calculation process of the sampling ratio optimization model are as follows 7 steps: (1) Initialize parameters: the uncompleted ratio xn = 1; the completed ratio xy = 0; the inventory of single coals KC = [kc1, kc2,..., kcn]; the remaining inventory xs = sum(KC).

[0070] (2) Update the solution, randomly and uniformly generate the coal amount x = xi of the i-th coal from the inventory of single coals KC, calculate the uncompleted ratio xn = xn - xi, the completed ratio xy = xy + xi, and the remaining inventory xs = xs - kci.

[0071] (3) Judge the condition xn - xs > 0 or xy > 1; (4) If the condition in (3) is not satisfied, then i = i + 1, judge whether i is equal to the maximum number of iterations m, if so, output the candidate solution (i.e., the candidate ratio plan); (5) If the condition in (3) is satisfied, then calculate the uncompleted ratio xn = xn - xi, the completed ratio xy = xy - xi, and the remaining inventory xs = xs + kci.

[0072] (6) Update the solution according to the method of updating the solution in (2); (7) Execute step (3), perform loop calculation until the candidate solution (i.e., the candidate ratio plan) is output.

[0073] In some embodiments, the blended coal quality constraints include the caking index, the plastic layer thickness, and the volatile matter of the blended coal index; the coke quality constraints include the ash content, the sulfur content, the shatter strength, the abrasion resistance, the reactivity index, and the post-reaction strength.

[0074] By strictly controlling the caking index, plastic layer thickness and volatile matter of the blended coal, it is possible to ensure that the coal can be fully softened, melted and solidified during the coking process, thereby improving the cold strength, thermal stability and overall quality of the coke. By restricting the ash content and sulfur content of the coke, the impurity content can be reduced and the harmful components in the coke can be lowered. The above-mentioned blended coal quality constraints and coke quality constraints also help to guide the design of the coal blending ratio, making the selection of raw coal more scientific and reasonable, and further optimizing the coking process parameters (such as the temperature on the coke side, coking time, etc.) to improve production efficiency. By controlling indexes such as the shatter strength, abrasion strength, reactivity index and post-reaction strength, it is beneficial to ensure the behavior of the coke during the blast furnace smelting process, reduce the generation of powder, improve the permeability of the burden, and increase the smelting efficiency. By controlling the volatile matter and sulfur content, it helps to reduce the pollutant emissions during the coking process, especially the emissions of sulfides and nitrogen oxides, to meet the environmental protection requirements and reduce the environmental liability risk of the enterprise.

[0075] Some embodiments of the present application further provide a precise coking coal blending system, which is applicable to the precise coking coal blending method described in the above embodiments. The system includes: A resource information database construction module, which is used to obtain the indexes of single coking coal to establish a single coking coal resource information database.

[0076] A blended coal quality prediction model construction module, which is used to establish a blended coal quality prediction model based on the single coking coal resource information database and according to a linear model and an SVR fusion model to obtain the blended coal quality indexes.

[0077] A coke quality prediction model construction module, which is used to establish a coke quality prediction model based on the blended coal quality indexes and the historical coking process data set and through a KAN neural network to obtain the coke quality indexes.

[0078] A ratio optimization module, which is used to establish a sampling ratio optimization model based on the single coal quality parameters and coking process parameters and with the expected blended coal quality indexes, expected coke quality indexes and inventory as constraint conditions to obtain a candidate ratio plan; and screen the candidate ratio plan based on the blended coal quality prediction model, the coke quality prediction model and the blended coal cost to obtain the optimal coal blending plan.

[0079] The above system first analyzes the production data to find the correlation between the quality indicators of coke and the quality indicators of single coals, so as to establish a coke quality prediction model; then, based on the requirements of coke quality, the requirements of blended coal quality, and the constraints such as the quality, price, and inventory of each single coal, the single coals to be blended and the optimal ratio of each coal are obtained to obtain an optimal coal blending plan. This method can not only intelligently predict the quality indicators of coke according to the characteristics and ratios of each single coal, but also intelligently and quickly calculate the optimal coal blending plan according to the actual production requirements and the characteristics of each single coal, thereby improving the quality stability and production efficiency of coke, reducing production costs, and solving the problem of inaccurate coking coal blending methods.

[0080] As can be seen from the above technical solutions, the embodiments of the present application provide a precise coking coal blending method and system. By obtaining the indicators of single coking coals, a resource information database of single coking coals is established; based on the resource information database of single coking coals and according to a linear model and an SVR fusion model, a blended coal quality prediction model is established to obtain the blended coal quality indicators; based on the blended coal quality indicators and the historical coking process data set, and through a KAN neural network, a coke quality prediction model is established to obtain the coke quality indicators; based on the quality parameters of single coals and the coking process parameters, and taking the expected blended coal quality indicators, the expected coke quality indicators, and the inventory as constraint conditions, a sampling ratio optimization model is established to obtain a candidate ratio plan; based on the blended coal quality prediction model, the coke quality prediction model, and the blended coal cost, the candidate ratio plan is screened to obtain an optimal coal blending plan, solving the problem of inaccurate coking coal blending methods.

[0081] For the similarities between the embodiments provided in the present application, reference can be made to each other. The specific embodiments provided above are only several examples under the general concept of the present application and do not constitute a limitation on the protection scope of the present application. For those skilled in the art, any other embodiments extended based on the solution of the present application without creative efforts belong to the protection scope of the present application.

Claims

1. A precise coking coal blending method, characterized in that, The method includes: Obtaining the indexes of single coking coal to establish a resource information database of single coking coal; Based on the resource information database of single coking coal, and establishing a blended coal quality prediction model according to the linear model and the SVR fusion model to obtain the blended coal quality indexes; Based on the blended coal quality indexes and the historical coking process data set, and establishing a coke quality prediction model through the KAN neural network to obtain the coke quality indexes; Based on the quality parameters of single coal and the coking process parameters, and establishing a sampling ratio optimization model with the expected blended coal quality indexes, the expected coke quality indexes and the inventory as constraint conditions to obtain the candidate ratio schemes; Based on the blended coal quality prediction model, the coke quality prediction model and the blended coal cost, screening the candidate ratio schemes to obtain the optimal coal blending scheme.

2. The precise coking coal blending method according to claim 1, characterized in that, The indexes of the single coking coal include: the maximum thickness of the plastic layer, the caking index, the petrographic indexes of single coal and the working points data.

3. The precise coking coal blending method according to claim 1, characterized in that, The steps of establishing a blended coal quality prediction model based on the resource information database of single coking coal and according to the linear model and the SVR fusion model include: Generating the single coal derived feature MBI through the resource information database of single coking coal; Based on the resource information database of single coking coal, the single coal derived feature MBI and the coal blending list, and establishing a blended coal quality prediction model according to the linear model and the SVR fusion model.

4. The precise coking coal blending method according to claim 1, characterized in that, The steps of establishing a coke quality prediction model based on the blended coal quality indexes and the historical coking process data set and through the KAN neural network include: Screening out the historical coking process data samples in the historical coking process data set; the historical coking process data samples include the coking process condition parameters corresponding to the blended coal produced each time and the historical coke quality indexes; Using the robust normalization method to scale the variables of the blended coal quality indexes and the coking process condition parameters to obtain the model feature sample set; Based on the model feature sample set and the historical coke quality indexes, establishing a coke quality prediction data set; Establishing a coke quality prediction model through the coke quality prediction data set and the KAN neural network.

5. The precise coking coal blending method according to claim 4, wherein The input variables of the coke quality prediction model include the caking index, volatile matter, moisture, ash, sulfur, coking time, coke side temperature and ash composition; the output of the coke quality prediction model includes the ash, sulfur, shatter strength, abrasion resistance strength, reactivity index and post-reaction strength of coke.

6. The precise coking coal blending method according to claim 1, characterized in that The steps of obtaining the optimal coal blending scheme include: Establishing a sampling ratio optimization model with the inventory of each single coal and the sum of all ratios being one as constraint conditions to obtain the candidate ratio schemes; Inputting the candidate ratio schemes, the single coal quality parameters and the coking process parameters into the blended coal quality prediction model and the coke quality prediction model to screen out the feasible solutions that meet the blended coal quality constraints and the coke quality constraints; Calculating the price of the feasible solutions, and screening out the optimal coal blending scheme with the lowest blended coal cost as the objective function.

7. The precise coking coal blending method according to claim 6, wherein, The calculation formula of the objective function is: ； Among them, n is the total amount of single coals participating in coal blending; is the cost price of the th single coal; is the ratio of the th single coal.

8. The precise coking coal blending method according to claim 7, characterized in that The calculation formula of the sampling ratio optimization model includes: ； where n is the total amount of single coals participating in coal blending; C is the cost price of each single coal; is the proportion of each single coal; W is the quality index of blended coal; V is the quality index of coke; is the proportion of the h j is the inventory ratio of the 9. The precise coking coal blending method according to claim 6, characterized in that The blended coal quality constraints include the caking index, plastic layer thickness, and volatile matter of the blended coal; the coke quality constraints include ash content, sulfur content, shatter strength, abrasion strength, reactivity index, and post-reaction strength.

10. A precise coking coal blending system, characterized in that, Applicable to the precise coking coal blending method described in any one of claims 1-9, the system includes: A resource information database construction module for obtaining the indexes of single coking coals to establish a single coking coal resource information database; A blended coal quality prediction model construction module for establishing a blended coal quality prediction model based on the single coking coal resource information database and according to a linear model and an SVR fusion model to obtain the blended coal quality indexes; A coke quality prediction model construction module for establishing a coke quality prediction model based on the blended coal quality indexes and the historical coking process data set and through a KAN neural network to obtain the coke quality indexes; A ratio optimization module for establishing a sampling ratio optimization model based on the single coal quality parameters and the coking process parameters and with the expected blended coal quality indexes, the expected coke quality indexes, and the inventory as constraint conditions to obtain a candidate ratio plan; screening the candidate ratio plan based on the blended coal quality prediction model, the coke quality prediction model, and the blended coal cost to obtain the optimal coal blending plan.

Images