A method and system for controlling coal blending for coking based on artificial intelligence

Abstract

The present invention relates to the field of coking coal blending, and more specifically, the present invention relates to a method and system for controlling coal coking coal blending based on artificial intelligence. The method comprises: collecting a plurality of pre-processed coal coke block grayscale images, constructing a grayscale matching matrix of the coal coke block grayscale images; calculating the importance of each element in the grayscale matching matrix, calculating the coal coke block loss value after each experiment according to the element value and importance, and taking the ratio of the coal coke block loss values ​​of two adjacent experiments as the coal coke block stability; inputting the coal coke block loss value and the coal coke block stability into a preset neural network to obtain a coal bonding index; constructing a historical coal blending sample, clustering all historical coal blending samples, and coking coal blending according to the blending ratio in the clustering result. Through the technical solution of the present invention, the accuracy of the coking coal blending results can be improved, and the quality and characteristics of the final coke can be guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of coal blending for coking, and more specifically, to a method and system for controlling coal blending for coking based on artificial intelligence. Background Art

[0002] Coking is a high-temperature dry distillation process that converts coal into coke. As the main fuel and reducing agent for blast furnace smelting of iron ore, coke occupies a core position in steel production. In the coking process, coal blending is a key step, and its purpose is to achieve the best coking effect by reasonably matching different types and qualities of coal. The quality of coal blending directly affects the quality and output of coke, and thus affects the efficiency and cost of the entire steel production. High-quality coke should have suitable block size, strength, low ash content, sulfur content and other characteristics. Therefore, achieving accurate coal blending control is of great significance for improving coke quality, reducing energy consumption and reducing environmental pollution. With the development of automation technology and information technology, modern coal blending control technology is developing in the direction of intelligence, automation and precision. Through automation equipment and image data processing, dynamic adjustment and optimization control of the coal blending process can be achieved.

[0003] The existing Chinese patent application document with publication number CN1749358A discloses a computer-controlled coking coal blending optimization system. The method integrates the blending coal and coke information through the coke production information system to form a production information database, providing regression data for the coke quality prediction system; the coal yard information system integrates the single coal information to form a raw coal information library; the coal blending optimization system ensures the coke quality through the coke quality prediction system, and optimizes the practical coal blending plan with the lowest cost through the coal yard information system; the coking coal blending optimization system monitors the coal blending production operation of the automatic coal blending device through the interface control system.

[0004] However, the physical properties of coal will also affect the final coking coal blending results. Different types of coal have different coking properties, such as adhesiveness, volatile matter, ash content, etc. The above application documents only consider factors that can affect the quality of coke, such as crushing particle size, final coking temperature and coal pile density, which may result in inaccurate coking coal blending results. Summary of the invention

[0005] In order to solve the problem of inaccurate coking coal blending results, the present invention proposes a coal coking coal blending control method and system based on artificial intelligence.

[0006] In a first aspect, the present invention discloses a coal coking coal blending control method based on artificial intelligence, comprising: collecting multiple pre-processed coal coke grayscale images, and constructing a grayscale matching matrix of the coal coke grayscale images; calculating the importance of each element in the grayscale matching matrix, calculating the coal coke loss value after each experiment according to the element value and the importance, and taking the ratio of the coal coke loss values ​​of two adjacent experiments as the coal coke stability; inputting the coal coke loss value and the coal coke stability into a preset neural network to obtain a coal bonding index; constructing the final coal blending bonding index in history and the batching ratio for obtaining the final coal blending bonding index as a historical coal blending sample, wherein the batching ratio is the ratio of the bonding indexes of different coals; clustering all historical coal blending samples to obtain multiple clustering clusters, each clustering cluster containing multiple batching ratio schemes, one clustering cluster corresponding to one final coal blending bonding index, and performing coking coal blending according to the batching ratio in the clustering result.

[0007] By constructing the grayscale matching matrix of the coal coke block grayscale image, the microscopic changes on the surface of the coal coke block during the experiment can be revealed. By analyzing these changes, the wear degree of the coal coke block can be inferred, and then its bonding performance can be evaluated. By defining the loss value ratio of two adjacent experiments as the stability of the coal coke block, the performance changes of the coal coke block can be dynamically monitored.

[0008] Preferably, the method further comprises: dividing the grayscale value of the coal coke grayscale image into a plurality of grayscale levels; the grayscale level division satisfies the relationship:

[0009] , Indicates the grayscale image of coal coke block Line The gray level of the column, Indicates the grayscale image of coal coke block Line The grayscale value of the column, It represents the maximum value of the grayscale value in the grayscale image of the coal coke block. It represents the minimum gray value in the grayscale image of the coal coke block. Indicates the total gray level.

[0010] The amount of calculation can be reduced by reducing the matrix size from 255 times 255 to N times N, thereby improving the efficiency of calculating the coal caking index.

[0011] Preferably, the importance includes: constructing a coordinate system with the upper left corner of the grayscale matching matrix as the origin, the horizontal left direction as the horizontal axis, and the horizontal downward direction as the vertical axis; for any element in the grayscale matching matrix, calculating the shortest distance between any element and the diagonal of the grayscale matching matrix according to the point-to-line distance formula, and taking the normalized shortest distance as the importance.

[0012] It can reveal the change of gray value at the same position in the two images before and after the experiment. The diagonal line indicates that the gray value at the same position has not changed, and the impact on the coal bonding index can be ignored. Therefore, the more the value deviates from the diagonal line, the greater the change of gray value before and after the experiment, and the greater the impact on the calculation of coal bonding index, so the greater the importance.

[0013] Preferably, the importance also satisfies the relationship:

[0014] , Represents the grayscale matching matrix Line The importance of column elements, Indicates the number of rows, Indicates the number of columns, Represents the normalization function.

[0015] Preferably, the coal coke loss value satisfies the relationship:

[0016] , Indicates the loss value of coal coke. Represents the grayscale matching matrix Line The importance of column elements, Represents the grayscale matching matrix Line The value of the column element.

[0017] The coke loss value can evaluate the wear degree of the coke. The ratio of the coke loss values ​​of two adjacent experiments is defined as the stability of the coke. It can dynamically monitor the performance changes of the coke and detect performance degradation in time.

[0018] Preferably, the preset neural network is a BP neural network, including an input layer, a hidden layer and an output layer; the input layer is used to receive the input of coal coke loss value and coal coke stability, and after the hidden layer extracts features from the input information, the extracted features are input into the output layer to output the coal bonding index.

[0019] Preferably, the training process of the preset neural network includes: taking the coal coke loss value and coal coke stability after the same experiment in history as input information, and taking the actual coal bonding index after the same experiment as a label, to obtain a set of training data; inputting the input information in the training data into the preset neural network to obtain an output result; calculating the loss value of the preset neural network through the output result and the label, the loss function of the preset neural network uses mean square error loss, back-propagating the error signal according to the loss value, updating the network parameters of the preset neural network, and making the loss value smaller; iteratively updating the network parameters of the preset neural network, and when the preset neural network reaches the set maximum number of training times or the network loss value is less than the set loss value, stopping the update to obtain a trained preset neural network.

[0020] By taking the loss value and stability of coal coke as input variables and the actual coal caking index as output label, the caking behavior of different types of coal can be simulated and learned. This helps to optimize the coking process, improve the quality of coke, reduce production costs, and provide accurate data support for the efficient use of coal resources and the automated control of the coking process.

[0021] In a second aspect, the present invention discloses an artificial intelligence-based coal coking coal blending control system, comprising: a processor; and a memory, wherein the memory stores computer instructions, and when the computer instructions are executed by the processor, the system executes the above-mentioned artificial intelligence-based coal coking coal blending control method.

[0022] Beneficial effects of the present invention:

[0023] By collecting the grayscale image of coal coke and constructing a grayscale matching matrix, the loss value and stability of coal coke are calculated, and then the coal bonding index is predicted, and finally the optimization of coking coal blending is achieved through cluster analysis. It can accurately simulate and quantify the physical changes of coal coke during the coking process.

[0024] By converting the image data of coal coke blocks into quantifiable grayscale values ​​and further calculating their loss values ​​and stability, the microscopic changes of coal coke blocks during the coking process can be effectively captured, thereby predicting their macroscopic properties.

[0025] By clustering and analyzing historical coal blending samples, we can find the internal connection between different proportion schemes, provide a scientific basis for coking coal blending, optimize the coal blending ratio, improve the quality of coke and production efficiency, and reduce costs. It not only improves the accuracy and efficiency of coal blending, but also helps save resources and protect the environment, and promotes the development of the coal coking industry towards intelligence and automation. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0027] Figure 1 It is a production flow chart of an embodiment of the present invention.

[0028] Figure 2 It is a flow chart of the detection link of an embodiment of the present invention.

[0029] Figure 3 It is a flow chart of the coal coke drum experiment in an embodiment of the present invention.

[0030] Figure 4 The present invention is a flowchart of a method for controlling coal blending for coking based on artificial intelligence. DETAILED DESCRIPTION

[0031] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0032] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0033] The invention provides a coal blending control method for coal coking based on artificial intelligence.

[0034] like Figure 1 The production flow chart is shown in Figure 1. The process of coal coking coal blending is to input the coal sample into the detection link, and the detection link outputs the coal bonding index; enter the coal blending link, refer to the coal bonding index to adjust the proportion of ingredients, complete the coal blending process, and produce according to the coal after the coal blending.

[0035] It should be noted that the coal sample will be burned into coal char before the drum test, and this process is common knowledge to those skilled in the art.

[0036] like Figure 2The figure shows the flow chart of the detection process. The detection process includes a camera, a host computer and a display screen. The camera is used to collect the grayscale image of the coal coke block before and after the drum test; the host computer stores a system for calculating the coal caking index; and the display screen displays the output coal caking index.

[0037] like Figure 3 The figure shows the flow chart of the drum test for coal coke. When calculating the coal caking index, each coal sample should be tested twice, so there is a cyclic process.

[0038] like Figure 4 As shown, a coal blending control method for coal coking based on artificial intelligence includes steps S1 to S4, which are described in detail below.

[0039] S1, collecting multiple pre-processed coal coke block grayscale images and constructing a grayscale matching matrix of the coal coke block grayscale images.

[0040] In one embodiment, in the drum experiment of coal coke, a CCD camera is used to capture RGB (RedGreenBlue) images before and after the experiment. With its high resolution and precise imaging capabilities, the CCD camera can capture subtle changes and features on the surface of the coal coke. By applying Gaussian filtering to these images for denoising, the image quality can be further improved.

[0041] Gaussian filtering is a commonly used image processing technique that reduces noise by smoothing the pixel values ​​in an image while retaining edge information as much as possible. The physical meaning of Gaussian filtering is that it simulates the way the human visual system perceives images, that is, when observing a point, the surrounding environment will also affect the perception of that point, and Gaussian filtering achieves the effect of denoising by fusing the pixel values ​​around each pixel through weighted averaging. This filtering method is widely used in the field of image processing, especially in industrial inspection, medical imaging and other fields, it can significantly improve the quality of images and the accuracy of analysis.

[0042] The RGB image is grayed to obtain the coal coke block grayscale image. The coal coke block grayscale image includes the coal coke block grayscale image before the experiment and the coal coke block grayscale image after the experiment. Before the drum experiment, the coal sample to be tested is photographed to obtain the coal coke block grayscale image before the experiment. After the first experiment, the coal coke block grayscale image after the experiment is taken. Since the drum experiment needs to be carried out at least twice, the coal coke block grayscale image after the first experiment is also used as the coal coke block grayscale image before the second experiment.

[0043] The probability of occurrence of pixel pairs at the same position in the grayscale image of the coal coke block before the experiment and the grayscale image of the coal coke block after the experiment is constructed as a grayscale matching matrix. The rows of the grayscale matching matrix represent the pixel values ​​in the grayscale image of the coal coke block before the experiment, and the columns of the grayscale matching matrix represent the pixel values ​​in the grayscale image of the coal coke block after the experiment. The elements in the grayscale matching matrix represent the number or probability of occurrence of the object pixel pairs.

[0044] Since the pixel value range of the grayscale image is 0 to 255, directly constructing the grayscale matching matrix will result in a 255 by 255 matrix, which is computationally intensive. Therefore, the grayscale values ​​in the grayscale image of the coal coke block can be divided into multiple grayscale levels before constructing the grayscale matching matrix. The grayscale level division satisfies the relationship:

[0045] , Indicates the grayscale image of coal coke block Line The gray level of the column, Indicates the grayscale image of coal coke block Line The grayscale value of the column, It represents the maximum value of the grayscale value in the grayscale image of the coal coke block. It represents the minimum gray value in the grayscale image of the coal coke block. Indicates the total gray level.

[0046] The total grayscale level is the side length of the grayscale matching matrix. For example, the total grayscale level is set to 16, and the grayscale matching matrix is ​​a matrix of 16 times 16. The total grayscale level can be set by those skilled in the art.

[0047] S2, calculate the importance of each element in the grayscale matching matrix, calculate the coal coke loss value after each experiment based on the element value and importance, and take the ratio of the coal coke loss values ​​of two adjacent experiments as the coal coke stability.

[0048] In one embodiment, the importance includes: constructing a coordinate system with the upper left corner of the grayscale matching matrix as the origin, the horizontal left direction as the horizontal axis, and the horizontal downward direction as the vertical axis; for any element in the grayscale matching matrix, calculating the shortest distance between any element and the diagonal line of the grayscale matching matrix according to the point-to-line distance formula, and taking the normalized shortest distance as the importance.

[0049] The loss value of the coal coke block after each experiment is calculated based on the element value and importance. The loss value of the coal coke block satisfies the relationship:

[0050] , Indicates the loss value of coal coke. Represents the grayscale matching matrix Line The importance of column elements, Represents the grayscale matching matrix Line The value of the column element.

[0051] It should be noted that the greater the importance, the greater the difference between the same position of the two images before and after the experiment, the greater the loss value of the coal coke block during the experiment, the smaller the bonding index, the easier the coal coke block is to break, the worse the ability of coal to form coke during heating, and the lower the quality of the coke.

[0052] The ratio of the coke loss values ​​of two consecutive experiments is taken as the coke stability. It can be the ratio of the coke loss value of the second experiment to the coke loss value of the first experiment, or the ratio of the coke loss value of the first experiment to the coke loss value of the second experiment. The closer the ratio is to 1, the closer the coke loss values ​​of the two experiments are.

[0053] In another embodiment, the importance also satisfies the relationship:

[0054] , Represents the grayscale matching matrix Line The importance of column elements, Indicates the number of rows, Indicates the number of columns, Represents the normalization function.

[0055] S3, inputting the coal coke block loss value and the coal coke block stability into a preset neural network to obtain a coal bonding index.

[0056] In one embodiment, the preset neural network is a BP neural network, including an input layer, a hidden layer and an output layer; the input layer is used to receive the input of coal coke loss value and coal coke stability, and after the hidden layer extracts features from the input information, the extracted features are input into the output layer to output the coal bonding index.

[0057] The training process of the preset neural network includes: collecting the coal coke block grayscale image of the historical coal coke block drum test, calculating the coal coke block loss value and the stability of the coal coke block, and professionals in this field give the actual coal bonding index of the coal coke block based on the drum test results.

[0058] The coal coke loss value and the coal coke stability after the same experiment in history are used as input information, and the actual coal bonding index after the same experiment is used as a label to obtain a set of training data; the input information in the training data is input into a preset neural network to obtain an output result; the loss value of the preset neural network is calculated through the output result and the label, and the loss function of the preset neural network uses the mean square error loss. The error signal is back-propagated according to the loss value, and the network parameters of the preset neural network are updated to reduce the loss value; the network parameters of the preset neural network are iteratively updated, and when the preset neural network reaches the set maximum number of training times or the network loss value is less than the set loss value, the update is stopped to obtain a trained preset neural network.

[0059] Exemplarily, the training is completed when the preset neural network is trained 200 times or the network loss value is less than 0.0001.

[0060] The coal coke loss value and the coal coke stability of the coal coke to be tested are input into a preset neural network to obtain the coal caking index.

[0061] S4, construct historical coal blending samples, cluster all historical coal blending samples, and perform coking coal blending according to the blending ratio in the clustering results.

[0062] It should be noted that coal blending refers to the combination of coking coals of different metamorphic degrees in appropriate proportions. Coals of different metamorphic degrees have their own characteristics, and they play different roles in coal blending. By utilizing the complementary properties of various coals, the quality of the blended coal can meet the requirements of coking production, which is of great significance for the rational use of coal resources, saving high-quality coking coal, and expanding coking coal resources. If the coal blending plan is reasonable, the characteristics of various coals can be fully utilized to improve the quality of coke.

[0063] Among the coal used for coking, the proportion of coking coal with good adhesion and fat coal resources is relatively small, and the method of coal blending must be adopted to save high-quality coking coal and expand the source of coking coal.

[0064] In one embodiment, the final coal blending cohesiveness index in history and the blending ratio for obtaining the final coal blending cohesiveness index are constructed as a historical coal blending sample, wherein the blending ratio is the ratio of the cohesiveness indexes of different coals.

[0065] All historical coal blending samples are clustered to obtain multiple clusters. Each cluster contains multiple blending ratio schemes. One cluster corresponds to a final coal blending adhesion index. Coking coal is blended according to the blending ratio in the clustering results.

[0066] For example, the final coal blending caking index is 18, which is obtained by The coal samples and the coal caking index are The coal samples are obtained according to the proportion of the ingredients, that is, the historical coal blending samples constructed are .

[0067] In addition, the final coal blending adhesion index 18 can also be calculated by the coal adhesion index: The coal sample, the coal caking index is The coal samples and coal caking index are The coal samples are obtained according to the proportion of the ingredients, that is, the historical coal blending samples constructed are .

[0068] Then the final cluster with coal blending cohesion index 18 will contain two batching ratio schemes. The coal samples and the coal caking index are For coal samples, the first batching ratio scheme is used for batching.

[0069] At this point, the coal blending work can be realized automatically, and the quality of the coal ratio can be controlled, while the total coal body and raw material ratio values ​​can be controlled.

[0070] An embodiment of the present invention also discloses an artificial intelligence-based coal coking coal blending control system, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an artificial intelligence-based coal coking coal blending control method according to the present invention is implemented.

[0071] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0072] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module or both. Any such computer storage medium may be part of a device or accessible or connectable to a device.

[0073] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

[0074] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for controlling coal blending for coking based on artificial intelligence, characterized in that: include: Collect multiple pre-processed coal coke block grayscale images and construct a grayscale matching matrix of the coal coke block grayscale images; The importance of each element in the grayscale matching matrix is ​​calculated, and the loss value of the coal coke block after each experiment is calculated according to the element value and importance. The ratio of the coal coke block loss values ​​of two adjacent experiments is taken as the stability of the coal coke block. The coal coke loss value and the coal coke stability are input into a preset neural network to obtain a coal caking index; The final coal blending caking index in history and the batching ratio for obtaining the final coal blending caking index are constructed as historical coal blending samples, wherein the batching ratio is the ratio of different coal caking indexes; All historical coal blending samples are clustered to obtain multiple clusters. Each cluster contains multiple blending ratio schemes. One cluster corresponds to a final coal blending adhesion index. Coking coal is blended according to the blending ratio in the clustering results.

2. The method for controlling coal blending for coking based on artificial intelligence according to claim 1, characterized in that: Also includes: The gray value of the coal coke block gray image is divided into multiple gray levels; The division of gray levels satisfies the relationship: , Indicates the grayscale image of coal coke. Line The gray level of the column, Indicates the grayscale image of coal coke. Line The grayscale value of the column, It represents the maximum value of the grayscale value in the grayscale image of the coal coke block. It represents the minimum gray value in the grayscale image of the coal coke block. Indicates the total gray level.

3. The method for controlling coal blending for coking based on artificial intelligence according to claim 1, characterized in that: The importance includes: A coordinate system is constructed with the upper left corner of the grayscale matching matrix as the origin, the horizontal left direction as the horizontal axis, and the horizontal downward direction as the vertical axis; For any element in the grayscale matching matrix, the shortest distance between any element and the diagonal of the grayscale matching matrix is ​​calculated according to the point-to-line distance formula, and the normalized shortest distance is used as the importance.

4. The method for controlling coal blending for coking based on artificial intelligence according to claim 1, characterized in that: The importance also satisfies the relationship: , Represents the grayscale matching matrix Line The importance of column elements, Indicates the number of rows, Indicates the number of columns, Represents the normalization function.

5. The method for controlling coal blending for coking based on artificial intelligence according to claim 1, characterized in that: The coal coke loss value satisfies the relationship: , Indicates the loss value of coal coke. Represents the grayscale matching matrix Line The importance of column elements, Represents the grayscale matching matrix Line The value of the column element.

6. The method for controlling coal blending for coking based on artificial intelligence according to claim 1, characterized in that: The preset neural network is a BP neural network, including an input layer, a hidden layer and an output layer; the input layer is used to receive the input of coal coke loss value and coal coke stability, and the hidden layer extracts features from the input information and inputs the extracted features into the output layer to output the coal bonding index.

7. The method for controlling coal blending for coking based on artificial intelligence according to claim 1, characterized in that: The training process of the preset neural network includes: The coal coke loss value and the coal coke stability after the same experiment in history are used as input information, and the actual coal caking index after the same experiment is used as a label to obtain a set of training data; Input the input information in the training data into the preset neural network to obtain the output result; The loss value of the preset neural network is calculated by outputting the results and labels. The loss function of the preset neural network uses mean square error loss. The error signal is back-propagated according to the loss value, and the network parameters of the preset neural network are updated to reduce the loss value. Iteratively update the network parameters of the preset neural network. When the preset neural network reaches the set maximum number of training times or the network loss value is less than the set loss value, stop updating to obtain a trained preset neural network.

8. A coal blending control system for coking based on artificial intelligence, characterized in that: include: Processor; and A memory storing computer instructions, wherein when the computer instructions are executed by a processor, the system executes an artificial intelligence-based coal coking coal blending control method according to any one of claims 1 to 7.

Images