Parameter generation method, coke quality prediction method and device

Abstract

The invention discloses a parameter generation method and a coke quality prediction method and device, and relates to the technical field of artificial intelligence and the technical field of coking industry, and the method comprises the steps: obtaining a minimum value and a maximum value of as-fired coal index data as input data in all training samples of a coke quality prediction model, calculating a coke index contribution parameter representing the influence of the unit change of the coal as fired index data on the coke index data on the basis of the numerical value interval represented by the minimum value and the maximum value and a model function of the coke quality prediction model, and calculating the coke index contribution parameter on the basis of the coke index contribution parameter. And determining a low-area coke index adjustment parameter and a high-area coke index adjustment parameter, wherein the low-area coke index adjustment parameter and the high-area coke index adjustment parameter are used for adjusting the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical interval. By adopting the scheme, the accuracy of the prediction result obtained by using the coke quality prediction model can be improved.

Description

Technical Field

[0001] The present application relates to the fields of artificial intelligence technology and coking industry technology, and in particular to a parameter generation method, a coke quality prediction method and a device. Background Art

[0002] The coal blending coking process directly determines the quality of coke. If the coke quality can be accurately predicted based on single coal information, coke oven process and other parameters before coking, then by adjusting the single coal ratio, more suitable single coal can be used for coking as much as possible to achieve better coke indicators, thereby reducing raw material costs and improving coke quality.

[0003] With the development of artificial intelligence technology and machine learning technology, a coke quality prediction model for predicting coke quality has emerged. Before starting the actual coking operation, the coke quality of coking using a certain type of coal into the furnace can be predicted through the coke quality prediction model.

[0004] However, in practical applications, due to the properties of the coking industry itself, such as complex processes, long time and high costs, it is difficult to accumulate sufficient historical coking data as sample data for model training. For example, if coking is conducted once a week, only about 40 pieces of data can be accumulated in a year, which cannot meet the conventional model training needs.

[0005] Moreover, due to the small amount of sample data, when actually predicting the coke quality, the value of the input coal index data may exceed the value range of the sample data for model training, resulting in inaccurate coke quality prediction values ​​output by the model.

[0006] Therefore, there is a problem in the prior art that the prediction results obtained using the coke quality prediction model are of low accuracy. Summary of the invention

[0007] The embodiments of the present application provide a parameter generation method, a coke quality prediction method and a device, which are used to solve the problem of low accuracy of prediction results obtained using a coke quality prediction model in the prior art.

[0008] The present application embodiment provides a parameter generation method for coke quality prediction, comprising: Obtain the minimum and maximum values ​​of the coal index data used as input data in all training samples of the coke quality prediction model; Based on the numerical range represented by the minimum value and the maximum value, and the model function of the coke quality prediction model, a coke index contribution parameter is calculated, wherein the coke index contribution parameter represents the influence of a unit change of the feed coal index data on the coke index data, the input of the model function is the feed coal index data, and the output of the model function is the coke index data; Based on the coke index contribution parameter, a low-zone coke index adjustment parameter and a high-zone coke index adjustment parameter are determined, and the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter are used to adjust the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical range.

[0009] Further, based on the numerical range represented by the minimum value and the maximum value, and the model function of the coke quality prediction model, the coke index contribution parameter is calculated, including: Based on a plurality of preselected positions in the numerical interval represented by the minimum value and the maximum value, respectively calculating the derivative of the model function of the coke quality prediction model at each of the preselected positions, and obtaining a plurality of coke index contribution parameters corresponding to the plurality of preselected positions respectively; The step of determining the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter based on the coke index contribution parameter comprises: According to the low-zone weighted method, the multiple coke index contribution parameters are weighted and summed to obtain the low-zone coke index adjustment parameter, wherein the weight values ​​corresponding to the multiple preselected positions in the low-zone weighted method show a downward trend as the preselected positions increase from small to large; According to the high-zone weighted method, the multiple coke index contribution parameters are weighted and summed to obtain the high-zone coke index adjustment parameter. The weight values ​​corresponding to the multiple pre-selected positions in the high-zone weighted method show an upward trend from small to large as the pre-selected position increases.

[0010] Further, the feed coal index data includes a plurality of feed coal test index data, and the coke index data includes a plurality of coke test index data indicating the quality of the coke; The method of obtaining the minimum and maximum values ​​of the coal index data used as input data in all training samples of the coke quality prediction model includes: Obtaining a minimum value vector and a maximum value vector of the coal index data used as input data in all training samples of the coke quality prediction model, wherein the minimum value vector includes the minimum value of each type of coal test index data in all training samples, and the maximum value vector includes the maximum value of each type of coal test index data in all training samples; The method of calculating the derivative of the model function of the coke quality prediction model at each of the preselected positions based on the preselected positions in the numerical interval represented by the minimum value and the maximum value, and obtaining a plurality of coke index contribution parameters corresponding to the preselected positions, respectively, includes: Based on multiple preselected positions in the numerical interval represented by the minimum value vector and the maximum value vector, the derivatives of the model function of the coke quality prediction model at each of the preselected positions are calculated respectively, and multiple coke index contribution matrices corresponding to the multiple preselected positions are obtained. The elements in the coke index contribution matrix represent the influence of a unit change in the test index data of each type of coal entering the furnace on each type of coke test index data.

[0011] Furthermore, the multiple preselected positions are evenly distributed in the numerical range.

[0012] Furthermore, the calculation formula of the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter for adjusting the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical range is as follows: Y=Y X +△X l M l +△X h M h , where M l The parameter for adjusting the coke index in the low zone, M h is the high-zone coke index adjustment parameter, and Y is the adjusted coke quality prediction value; Y X = f(X`), wherein f() is the model function of the coke quality prediction model, and X` is obtained by changing the value of the multiple test index data X of the feed coal to be predicted that is less than the minimum test index data of the feed coal to the corresponding minimum value, and changing the value of the multiple test index data X of the feed coal to be predicted that is greater than the maximum test index data of the feed coal to the corresponding maximum value, while keeping the other test index data of the feed coal unchanged; △X l To XX min The result is that the elements greater than 0 are changed to 0, and the elements less than or equal to 0 remain unchanged, X min is the minimum value vector; △X h To XX max The result is that the elements less than 0 are changed to 0, and the elements greater than or equal to 0 remain unchanged. max is the maximum value vector.

[0013] The present application also provides a method for predicting coke quality, comprising: Obtaining the coal index data of the coal to be predicted; When the feed coal index data of the feed coal to be predicted exceeds the numerical range represented by the minimum value and the maximum value, based on the coke quality prediction model and the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter in any of the above-mentioned parameter generation methods for coke quality prediction, the coke quality of coking using the feed coal to be predicted is predicted; The minimum value and the maximum value are the minimum value and the maximum value of the coal index data used as input data in all training samples of the coke quality prediction model.

[0014] Further, based on the feed coal index data of the feed coal to be predicted, the coke quality prediction model is adopted, and the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter in any of the above parameter generation methods for coke quality prediction are used to predict the coke quality using the feed coal to be predicted, including: Based on the feed coal index data of the feed coal to be predicted, the quality of coke produced by coking using the feed coal to be predicted is predicted using the following formula: Y=Y X +△X l M l +△X h M h , where M l The parameter for adjusting the coke index in the low zone, M h is the high-zone coke index adjustment parameter, and Y is the adjusted coke quality prediction value; Y X = f(X`), wherein f() is the model function of the coke quality prediction model, and X` is obtained by changing the value of the multiple test index data X of the feed coal to be predicted that is less than the minimum test index data of the feed coal to the corresponding minimum value, and changing the value of the multiple test index data X of the feed coal to be predicted that is greater than the maximum test index data of the feed coal to the corresponding maximum value, while keeping the other test index data of the feed coal unchanged; △X l To XX min The result is that the elements greater than 0 are changed to 0, and the elements less than or equal to 0 remain unchanged, X min is the minimum value vector; △X h To XX max The result is that the elements less than 0 are changed to 0, and the elements greater than or equal to 0 remain unchanged. max is the maximum value vector.

[0015] The present application also provides a parameter generation device for coke quality prediction, comprising: A numerical acquisition module is used to obtain the minimum and maximum values ​​of the coal index data used as input data in all training samples of the coke quality prediction model; a contribution parameter calculation module, for calculating a coke index contribution parameter based on the numerical range represented by the minimum value and the maximum value, and a model function of the coke quality prediction model, wherein the coke index contribution parameter represents the influence of a unit change of the feed coal index data on the coke index data, the input of the model function is the feed coal index data, and the output of the model function is the coke index data; An adjustment parameter determination module is used to determine the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter based on the coke index contribution parameter, and the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter are used to adjust the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical range.

[0016] Further, the contribution parameter calculation module is specifically used to calculate the derivative of the model function of the coke quality prediction model at each of the preselected positions based on the preselected positions in the numerical interval represented by the minimum value and the maximum value, and obtain a plurality of coke index contribution parameters corresponding to the preselected positions respectively; The adjustment parameter determination module is specifically used to perform weighted summation on the multiple coke index contribution parameters according to the low-zone weighted method to obtain the low-zone coke index adjustment parameter, wherein the weight values ​​corresponding to the multiple preselected positions in the low-zone weighted method show a downward trend as the preselected position increases from small to large; According to the high-zone weighted method, the multiple coke index contribution parameters are weighted and summed to obtain the high-zone coke index adjustment parameter. The weight values ​​corresponding to the multiple pre-selected positions in the high-zone weighted method show an upward trend from small to large as the pre-selected position increases.

[0017] Further, the feed coal index data includes a plurality of feed coal test index data, and the coke index data includes a plurality of coke test index data indicating the quality of the coke; The numerical acquisition module is specifically used to obtain the minimum value vector and the maximum value vector of the coal index data used as input data in all training samples of the coke quality prediction model, wherein the minimum value vector includes the minimum value of each type of coal test index data in all training samples, and the maximum value vector includes the maximum value of each type of coal test index data in all training samples; The contribution parameter calculation module is specifically used to calculate the derivative of the model function of the coke quality prediction model at each preselected position based on multiple preselected positions in the numerical interval represented by the minimum value vector and the maximum value vector, and obtain multiple coke index contribution matrices corresponding to the multiple preselected positions respectively. The elements in the coke index contribution matrix represent the influence of a unit change in the test index data of each type of coal entering the furnace on each type of coke test index data.

[0018] Furthermore, the multiple preselected positions are evenly distributed in the numerical range.

[0019] Furthermore, the calculation formula of the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter for adjusting the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical range is as follows: Y=Y X +△X l M l +△X h M h , where M l The parameter for adjusting the coke index in the low zone, M h is the high-zone coke index adjustment parameter, and Y is the adjusted coke quality prediction value; Y X = f(X`), wherein f() is the model function of the coke quality prediction model, and X` is obtained by changing the value of the multiple test index data X of the feed coal to be predicted that is less than the minimum test index data of the feed coal to the corresponding minimum value, and changing the value of the multiple test index data X of the feed coal to be predicted that is greater than the maximum test index data of the feed coal to the corresponding maximum value, while keeping the other test index data of the feed coal unchanged; △X l To XX min The result is that the elements greater than 0 are changed to 0, and the elements less than or equal to 0 remain unchanged, X min is the minimum value vector; △X h To XX max The result is that the elements less than 0 are changed to 0, and the elements greater than or equal to 0 remain unchanged. max is the maximum value vector.

[0020] The present application also provides a coke quality prediction device, comprising: A data acquisition module is used to acquire the coal index data of the coal to be predicted; A quality prediction module, used for predicting the quality of coke produced by coking using the to-be-predicted feed coal, based on the coke quality prediction model and the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter in any of the above-mentioned parameter generation methods for coke quality prediction, when the feed coal index data of the to-be-predicted feed coal exceeds the numerical range represented by the minimum value and the maximum value; The minimum value and the maximum value are the minimum value and the maximum value of the coal index data used as input data in all training samples of the coke quality prediction model.

[0021] Furthermore, the quality prediction module is specifically used to predict the quality of coke produced by coking using the to-be-predicted coal based on the coal index data of the to-be-predicted coal, using the following formula: Y=Y X +△X l M l +△X h M h , where M l The parameter for adjusting the coke index in the low zone, M h is the high-zone coke index adjustment parameter, and Y is the adjusted coke quality prediction value; Y X = f(X`), wherein f() is the model function of the coke quality prediction model, and X` is obtained by changing the value of the multiple test index data X of the feed coal to be predicted that is less than the minimum test index data of the feed coal to the corresponding minimum value, and changing the value of the multiple test index data X of the feed coal to be predicted that is greater than the maximum test index data of the feed coal to the corresponding maximum value, while keeping the other test index data of the feed coal unchanged; △X l To XX min The result is that the elements greater than 0 are changed to 0, and the elements less than or equal to 0 remain unchanged, X min is the minimum value vector; △X h To XX max The result is that the elements less than 0 are changed to 0, and the elements greater than or equal to 0 remain unchanged. max is the maximum value vector.

[0022] An embodiment of the present application also provides an electronic device, including a processor and a machine-readable storage medium, wherein the machine-readable storage medium stores machine-executable instructions that can be executed by the processor, and the processor is prompted by the machine-executable instructions to: implement any of the above-mentioned parameter generation methods for coke quality prediction, or implement any of the above-mentioned coke quality prediction methods.

[0023] An embodiment of the present application also provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, it implements any of the above-mentioned parameter generation methods for coke quality prediction, or implements any of the above-mentioned coke quality prediction methods.

[0024] An embodiment of the present application also provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute any of the above-mentioned parameter generation methods for coke quality prediction, or to execute any of the above-mentioned coke quality prediction methods.

[0025] The beneficial effects of this application include: In the method provided in an embodiment of the present application, the minimum and maximum values ​​of the incoming coal index data used as input data in all training samples of the coke quality prediction model are obtained, and based on the numerical range represented by the minimum and maximum values ​​and the model function of the coke quality prediction model, the coke index contribution parameter is calculated, and the coke index contribution parameter represents the influence of a unit change in the incoming coal index data on the coke index data, the input of the model function is the incoming coal index data, and the output of the model function is the coke index data, and based on the coke index contribution parameter, a low-zone coke index adjustment parameter and a high-zone coke index adjustment parameter are determined, and the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter are used to adjust the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical range. In this method, since the calculated coke index contribution parameter can represent the impact of a unit change in the coal index data on the coke index data, the regional coke index adjustment parameter and the high-zone coke index adjustment parameter determined based on the coke index contribution parameter can effectively adjust the output coke quality prediction value when the actual input of the prediction model exceeds the numerical range, thereby improving the accuracy of the prediction results.

[0026] Other features and advantages of the present application will be described in the subsequent description, and partly become apparent from the description, or be understood by practicing the present application. The purpose and other advantages of the present application can be realized and obtained by the structures specifically pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The accompanying drawings are used to provide a further understanding of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings: Figure 1 A flow chart of a parameter generation method for coke quality prediction provided in an embodiment of the present application; Figure 2 A flow chart of a coke quality prediction method provided in an embodiment of the present application; Figure 3 A schematic diagram of the structure of a parameter generation device for coke quality prediction provided in an embodiment of the present application; Figure 4 A schematic diagram of the structure of a coke quality prediction device provided in an embodiment of the present application; Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to provide an implementation scheme to improve the accuracy of the prediction results obtained using the coke quality prediction model, the embodiments of the present application provide a parameter generation method, a coke quality prediction method and a device. The preferred embodiments of the present application are described below in conjunction with the drawings of the specification. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and are not used to limit the present application. In addition, the embodiments in the present application and the features in the embodiments can be combined with each other if there is no conflict.

[0029] The present application embodiment provides a parameter generation method for coke quality prediction, such as Figure 1 As shown, including: Step 11, obtaining the minimum and maximum values ​​of the coal index data used as input data in all training samples of the coke quality prediction model; Step 12: based on the numerical range represented by the minimum value and the maximum value, and the model function of the coke quality prediction model, calculate the coke index contribution parameter, the coke index contribution parameter represents the influence of a unit change of the feed coal index data on the coke index data, the input of the model function is the feed coal index data, and the output of the model function is the coke index data; Step 13: Based on the coke index contribution parameter, determine the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter, which are used to adjust the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical range.

[0030] By using the parameter generation method for coke quality prediction provided in the embodiment of the present application, since the calculated coke index contribution parameter can represent the influence of a unit change in the index data of the coal entering the furnace on the coke index data, the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter determined based on the coke index contribution parameter can effectively adjust the output coke quality prediction value when the actual input of the prediction model exceeds the numerical range, thereby improving the accuracy of the prediction result.

[0031] Accordingly, the present application also provides a coke quality prediction method, such as Figure 2 As shown, including: Step 21, obtaining the coal index data of the coal to be predicted; Step 22: when the coal index data of the to-be-predicted coal exceeds the numerical range represented by the minimum value and the maximum value, based on the coal index data of the to-be-predicted coal, a coke quality prediction model is used, as well as the low-zone coke index adjustment parameters and the high-zone coke index adjustment parameters in the parameter generation method for coke quality prediction, to predict the coke quality of coking using the to-be-predicted coal; The minimum value and the maximum value are the minimum value and the maximum value of the coal index data used as input data in all training samples of the coke quality prediction model.

[0032] By adopting the above-mentioned coke quality prediction method provided in the embodiment of the present application, for the case where the input coal index data of the coke quality prediction model exceeds the numerical range of the model input during training, the case where the index data is less than the minimum value and the case where the index data is greater than the maximum value is distinguished, and the low-zone coke index adjustment parameters and the high-zone coke index adjustment parameters are used to adjust the coke quality prediction value output by the model, thereby improving the accuracy of the prediction results.

[0033] The method and device provided by the present application are described in detail below with specific embodiments.

[0034] In one embodiment of the present application, for the above-mentioned step 12, based on the numerical range represented by the minimum value and the maximum value, and the model function of the coke quality prediction model, the coke index contribution parameter is calculated, and multiple preselected positions can be determined in advance in the numerical range represented by the minimum value and the maximum value, and each preselected position corresponds to a corresponding numerical value in the numerical range. Then, the derivative of the model function of the coke quality prediction model at each preselected position is calculated respectively, and multiple coke index contribution parameters corresponding to the multiple preselected positions are obtained. Each coke index contribution parameter can represent the influence of a unit change of the in-furnace coal index data at the preselected position on the coke index data.

[0035] Accordingly, for the above step 13, based on the coke index contribution parameter, determining the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter may specifically include: According to the low-zone weighted method, the multiple coke index contribution parameters are weighted and summed to obtain the low-zone coke index adjustment parameter, and the weight values ​​corresponding to the multiple pre-selected positions in the low-zone weighted method show a downward trend as the pre-selected position increases from small to large; According to the high-zone weighted method, the multiple coke index contribution parameters are weighted and summed to obtain the high-zone coke index adjustment parameter. In the high-zone weighted method, the weight values ​​corresponding to the multiple pre-selected positions respectively show an upward trend as the pre-selected position increases from small to large.

[0036] In this embodiment, the number of the plurality of preselected positions may be determined based on the number of actual applications, for example, may be 8 or 16, including a minimum value and a maximum value, and the plurality of preselected positions may be evenly distributed in the numerical range; further, in order to reduce the amount of calculation, there may be two preselected positions, one of which corresponds to the minimum value and the other corresponds to the maximum value; The weight value corresponding to each preselected position is also the weight value of the coke index contribution parameter corresponding to the preselected position. The specific setting of the weight value can be set based on the needs of the actual application, and can also be generated by a normalized exponential function softmax function. It only needs to satisfy that the weight values ​​corresponding to the multiple preselected positions in the low-zone weighted method show a downward trend as the preselected position increases from small to large, and the weight values ​​corresponding to the multiple preselected positions in the high-zone weighted method show an upward trend as the preselected position increases from small to large.

[0037] In the embodiment of the present application, the input data of the coke quality prediction model is the coal index data entering the furnace, and the output data is the coke index data. The coke index data, as the coke quality prediction value, can represent the coke quality.

[0038] In practical applications, the index data of the coal entering the furnace can be obtained by actual testing of the coal entering the furnace, or a coal entering the furnace index prediction model can be used to predict the index data of the mixed coal serving as the coal entering the furnace. The mixed coal is obtained by mixing multiple single coals in a certain proportion. The input of the coal entering the furnace index prediction model is the index data of multiple single coals and the proportion of each single coal, and the output is the index data of the coal entering the furnace (i.e., the mixed coal).

[0039] In one embodiment of the present application, the feed coal index data may include a plurality of feed coal test index data, and the coke index data may include a plurality of coke test index data representing the quality of the coke.

[0040] Furthermore, the plurality of test index data of the incoming coal may include at least two of the following test index data: Ash content on dry basis Ad; Dry basis volatile matter Vd; Combustible base volatile matter Vdaf; Sulfur content on dry ash-free basis Std; Adhesion index G; The ultimate shrinkage of coal is x; Maximum thickness of the glial layer y; The plurality of coke test index data may include at least two of the following test index data: Ash content on dry basis Ad; Combustible base volatile matter Vdaf; Crushing strength M40; Wear resistance: M10; Reactive CRI; Post-reaction strength CSR.

[0041] Correspondingly, for the above step 11, the minimum value and maximum value of the coal index data used as input data in all training samples of the coke quality prediction model are obtained. Specifically, the minimum value vector and the maximum value vector of the coal index data used as input data in all training samples of the coke quality prediction model can be obtained, wherein the minimum value vector includes the minimum value of each type of coal test index data in all training samples, and the maximum value vector includes the maximum value of each type of coal test index data in all training samples; Xmin=[X1 min ,X2 min ,X3 min ,X4 min ,X5 min ,X6 min ,X7 min ]=[Ad min ,Vd min ,Vdaf min ,Std min ,G min ,x min ,y min ], represents the minimum value vector, each element in the vector represents the minimum value of the test index data of the coal entering the furnace; Xmax=[X1 max ,X2 max ,X3 max ,X4 max ,X5 max ,X6 max ,X7 max ]=[Ad max ,Vd max ,Vdaf max ,Std max ,G max ,x max ,y max ], which represents the maximum value vector, and each element in the vector represents the maximum value of the test index data of this type of coal entering the furnace.

[0042] Since it includes multiple coal test index data and multiple coke test index data, the coke index contribution parameter is a coke index contribution matrix, wherein the elements in the coke index contribution matrix represent the influence of a unit change of each coal test index data on each coke test index data, as shown below: [ [R ad-ad ,R ad-Vd ,R ad-Vdaf ,Rad-Std ,R ad-G ,R ad-x ,R ad-y ], [R Vdaf-ad ,R Vdaf-Vd ,R Vdaf-Vdaf ,R Vdaf-Std ,R Vdaf-G ,R Vdaf-x ,R Vdaf-y ], [R M40-ad ,R M40-Vd ,R M40-Vdaf ,R M40-Std ,R M40-G ,R M40-x ,R M40-y ], [R M10-ad ,R M10-Vd ,R M10-Vdaf ,R M10-Std ,R M10-G ,R M10-x ,R M10-y ], [R CRI-ad ,R CRI-Vd ,R CRI-Vdaf ,R CRI-Std ,R CRI-G ,R CRI-x ,R CRI-y ], [R CSR-ad ,R CSR-Vd ,R CSR-Vdaf ,R CSR-Std ,R CSR-G ,R CSR-x ,R CSR-y ] ] Among them, R j-i The meaning of is the change in the coke test index data j caused by a unit change in the test index data i of the coal entering the furnace.

[0043] Accordingly, for the above step 12, based on the numerical range represented by the minimum value and the maximum value, and the model function of the coke quality prediction model, the coke index contribution parameter is calculated, which may specifically include: Based on multiple preselected positions in the numerical interval represented by the minimum vector and the maximum vector, the derivative of the model function of the coke quality prediction model at each preselected position is calculated respectively, and multiple coke index contribution matrices corresponding to the multiple preselected positions are obtained. For example, n coke index contribution matrices are obtained, M=[M 1 ,M 2 ...M n], n is the number of multiple pre-selected positions.

[0044] Accordingly, the following formula is used to calculate the low-zone coke index adjustment parameters: M l =W l ·M=Wl1*M1+Wl2*M2+....Wln*Mn; Among them, W l is the weight vector corresponding to the low area weighting method, M l Adjust parameters for low zone coke indicators.

[0045] The following formula is used to calculate the low-zone coke index adjustment parameters: M h =W h ·M=Wh1*M1+Wh2*M2+....Whn*Mn; Among them, W h is the weight vector corresponding to the high zone weighted method, and Mh is the high zone coke index adjustment parameter.

[0046] In the embodiment of the present application, the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter are used to adjust the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical range. Specifically, the coke quality prediction value can be adjusted using the following formula: Regarding step 22 in the above coke quality prediction method, when there is only one coal index data of the coal to be predicted, it may specifically include: When the coal index data is less than the minimum value, the following formula is used to adjust the coke quality prediction value using the low-zone coke index adjustment parameter to obtain the adjusted coke quality prediction value: Y l =Y min + (X l -X min )·M l ; Among them, X min is the minimum value, X l is the input of the coke quality prediction model, and the input is less than the minimum value, Y min The input is X min The output of the coke quality prediction model is l Y is the low-zone coke index adjustment parameter, l is the adjusted predicted value of coke quality; When the coal index data of the furnace is greater than the maximum value, the high-zone coke index adjustment parameters are used to adjust the coke quality prediction value to obtain the adjusted coke quality prediction value: Y h =Y max + (Xh -X max )·M h ; Among them, X max is the maximum value, X h is the input of the coke quality prediction model, which is greater than the maximum value, Y max The input is X max The output of the coke quality prediction model is h Y is the high-zone coke index adjustment parameter, h is the adjusted coke quality prediction value.

[0047] Regarding step 22 in the above coke quality prediction method, when the incoming coal index data to be predicted includes multiple data, the data may specifically include: Y=Y X +△X l M l +△X h M h , where M l Adjustment parameters for low-zone coke index, M h is the coke index adjustment parameter of the area, and Y is the predicted value of coke quality after adjustment; Y X = f(X`), where f() is the model function of the coke quality prediction model, and X` is obtained by changing the test index data X of the multiple test index data of the feed coal to be predicted, which is less than the minimum test index data of the feed coal to the corresponding minimum value, and which is greater than the maximum test index data of the feed coal to the corresponding maximum value, while other test index data of the feed coal remain unchanged; △X l To XX min The result is that the elements greater than 0 are changed to 0, and the elements less than or equal to 0 remain unchanged, X min is the minimum value vector; △X h To XX max The result is that the elements less than 0 are changed to 0, and the elements greater than or equal to 0 remain unchanged. max is the maximum value vector; For example, suppose that the multiple coal test index data X of the coal to be predicted are as follows: X=[X1,X2,X3,X4,X5,X6,X7], where X1, X3 and X7 are smaller than X1 respectively. min , X3 min and X7 min , X2, X4 and X6 correspond to max , X4 max and X6max ,but: X`=[X1 min ,X2 max ,X3 min ,X4 max ,X5,X6 max ,X7 min ]; XX min =[X1-X1 min ,X2-X2 min ,X3-X3 min ,X4-X4 min ,X5-X5 min ,X6-X6 min ,X7-X7 min ], where X2-X2 min , X4-X4 min , X5-X5 min and X6-X6 min All are greater than 0, so change them to 0, △X l =[X1-X1 min ,0,X3-X3 min ,0,0,0,X7-X7 min ]; XX max =[X1-X1 max ,X2-X2 max ,X3-X3 max ,X4-X4 max ,X5-X5 max ,X6-X6 max ,X7-X7 max ], where X1-X1 max , X3-X3 max , X5-X5 max and X7-X7 max are all less than 0, so change them to 0, △X h =[0,X2-X2 max ,0,X4-X4 max ,0,X6-X6 max ,0].

[0048] Based on the above formula for the case where there are multiple incoming coal index data, it can be known that when there is only one incoming coal index data of the incoming coal to be detected, the formula is also substantially applicable.

[0049] Based on the same inventive concept, according to the parameter generation method for coke quality prediction provided in the above embodiment of the present application, correspondingly, another embodiment of the present application also provides a parameter generation device for coke quality prediction, and its structural schematic diagram is shown as follows: Figure 3 As shown, specifically including: A numerical acquisition module 31 is used to obtain the minimum and maximum values ​​of the coal index data used as input data in all training samples of the coke quality prediction model; A contribution parameter calculation module 32, for calculating a coke index contribution parameter based on the numerical range represented by the minimum value and the maximum value and the model function of the coke quality prediction model, wherein the coke index contribution parameter represents the influence of a unit change of the feed coal index data on the coke index data, the input of the model function is the feed coal index data, and the output of the model function is the coke index data; The adjustment parameter determination module 33 is used to determine the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter based on the coke index contribution parameter, and the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter are used to adjust the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical range.

[0050] Further, the contribution parameter calculation module 32 is specifically used to calculate the derivative of the model function of the coke quality prediction model at each of the preselected positions based on the preselected positions in the numerical interval represented by the minimum value and the maximum value, and obtain a plurality of coke index contribution parameters corresponding to the preselected positions respectively; The adjustment parameter determination module 33 is specifically used to perform weighted summation on the multiple coke index contribution parameters according to the low-zone weighted method to obtain the low-zone coke index adjustment parameter, wherein the weight values ​​corresponding to the multiple preselected positions in the low-zone weighted method show a downward trend as the preselected position increases from small to large; According to the high-zone weighted method, the multiple coke index contribution parameters are weighted and summed to obtain the high-zone coke index adjustment parameter. The weight values ​​corresponding to the multiple pre-selected positions in the high-zone weighted method show an upward trend from small to large as the pre-selected position increases.

[0051] Further, the feed coal index data includes a plurality of feed coal test index data, and the coke index data includes a plurality of coke test index data indicating the quality of the coke; The numerical acquisition module 31 is specifically used to obtain the minimum value vector and the maximum value vector of the coal index data used as input data in all training samples of the coke quality prediction model, wherein the minimum value vector includes the minimum value of each coal test index data in all training samples, and the maximum value vector includes the maximum value of each coal test index data in all training samples; The contribution parameter calculation module 32 is specifically used to calculate the derivative of the model function of the coke quality prediction model at each preselected position based on multiple preselected positions in the numerical interval represented by the minimum value vector and the maximum value vector, and obtain multiple coke index contribution matrices corresponding to the multiple preselected positions respectively. The elements in the coke index contribution matrix represent the influence of a unit change in the test index data of each type of coal entering the furnace on each type of coke test index data.

[0052] Furthermore, the multiple preselected positions are evenly distributed in the numerical range.

[0053] Furthermore, the calculation formula of the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter for adjusting the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical range is as follows: Y=Y X +△X l M l +△X h M h , where M l The parameter for adjusting the coke index in the low zone, M h is the high-zone coke index adjustment parameter, and Y is the adjusted coke quality prediction value; Y X = f(X`), wherein f() is the model function of the coke quality prediction model, and X` is obtained by changing the value of the multiple test index data X of the feed coal to be predicted that is less than the minimum test index data of the feed coal to the corresponding minimum value, and changing the value of the multiple test index data X of the feed coal to be predicted that is greater than the maximum test index data of the feed coal to the corresponding maximum value, while keeping the other test index data of the feed coal unchanged; △X l To XX min The result is that the elements greater than 0 are changed to 0, and the elements less than or equal to 0 remain unchanged, X min is the minimum value vector; △X h To XX max The result is that the elements less than 0 are changed to 0, and the elements greater than or equal to 0 remain unchanged. max is the maximum value vector.

[0054] Based on the same inventive concept, according to the coke quality prediction method provided in the above embodiment of the present application, correspondingly, another embodiment of the present application also provides a coke quality prediction device, whose structural schematic diagram is shown in FIG. Figure 4 As shown, specifically including: The data acquisition module 41 is used to acquire the coal index data of the coal to be predicted; The quality prediction module 42 is used to predict the quality of coke produced by coking using the to-be-predicted coal for coaling based on the coke quality prediction model and the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter in any of the above-mentioned parameter generation methods for coke quality prediction when the to-be-predicted coal for coaling index data exceeds the numerical range represented by the minimum value and the maximum value; The minimum value and the maximum value are the minimum value and the maximum value of the coal index data used as input data in all training samples of the coke quality prediction model.

[0055] Furthermore, the quality prediction module 42 is specifically used to predict the quality of coke produced by coking using the to-be-predicted coal based on the coal index data of the to-be-predicted coal, using the following formula: Y=Y X +△X l M l +△X h M h , where M l The parameter for adjusting the coke index in the low zone, M h is the high-zone coke index adjustment parameter, and Y is the adjusted coke quality prediction value; Y X = f(X`), wherein f() is the model function of the coke quality prediction model, and X` is obtained by changing the value of the multiple test index data X of the feed coal to be predicted that is less than the minimum test index data of the feed coal to the corresponding minimum value, and changing the value of the multiple test index data X of the feed coal to be predicted that is greater than the maximum test index data of the feed coal to the corresponding maximum value, while keeping the other test index data of the feed coal unchanged; △X l To XX min The result is that the elements greater than 0 are changed to 0, and the elements less than or equal to 0 remain unchanged, X min is the minimum value vector; △X h To XX max The result is that the elements less than 0 are changed to 0, and the elements greater than or equal to 0 remain unchanged. max is the maximum value vector.

[0056] The functions of the above modules can correspond to Figure 1 and Figure 2 The corresponding processing steps in the shown process will not be repeated here.

[0057] The parameter generation device for coke quality prediction and the coke quality prediction device provided in the embodiments of the present application can be implemented by a computer program. Those skilled in the art should be able to understand that the above-mentioned module division method is only one of many module division methods. If it is divided into other modules or no modules are divided, as long as the parameter generation device for coke quality prediction and the coke quality prediction device have the above-mentioned functions, they should be within the protection scope of the present application.

[0058] The present application also provides an electronic device, such as Figure 5 As shown, it includes a processor 51 and a machine-readable storage medium 52, wherein the machine-readable storage medium 52 stores machine-executable instructions that can be executed by the processor 51, and the processor 51 is prompted by the machine-executable instructions to: implement any of the above-mentioned parameter generation methods for coke quality prediction, or implement any of the above-mentioned coke quality prediction methods.

[0059] An embodiment of the present application also provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, it implements any of the above-mentioned parameter generation methods for coke quality prediction, or implements any of the above-mentioned coke quality prediction methods.

[0060] An embodiment of the present application also provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute any of the above-mentioned parameter generation methods for coke quality prediction, or to execute any of the above-mentioned coke quality prediction methods.

[0061] The machine-readable storage medium in the electronic device may include a random access memory (RAM) or a non-volatile memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the processor.

[0062] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0063] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, electronic device, computer-readable storage medium, and computer program product embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0064] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0065] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0066] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0067] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0068] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A parameter generation method for coke quality prediction, characterized in that: include: Obtain the minimum and maximum values ​​of the coal index data used as input data in all training samples of the coke quality prediction model; Based on the numerical range represented by the minimum value and the maximum value, and the model function of the coke quality prediction model, a coke index contribution parameter is calculated, wherein the coke index contribution parameter represents the influence of a unit change of the feed coal index data on the coke index data, the input of the model function is the feed coal index data, and the output of the model function is the coke index data; Based on the coke index contribution parameter, a low-zone coke index adjustment parameter and a high-zone coke index adjustment parameter are determined, and the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter are used to adjust the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical range.

2. The method according to claim 1, characterized in that Based on the numerical range represented by the minimum value and the maximum value, and the model function of the coke quality prediction model, the coke index contribution parameter is calculated, including: Based on a plurality of preselected positions in the numerical interval represented by the minimum value and the maximum value, respectively calculating the derivative of the model function of the coke quality prediction model at each of the preselected positions, and obtaining a plurality of coke index contribution parameters corresponding to the plurality of preselected positions respectively; The step of determining the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter based on the coke index contribution parameter comprises: According to the low-zone weighted method, the multiple coke index contribution parameters are weighted and summed to obtain the low-zone coke index adjustment parameter, wherein the weight values ​​corresponding to the multiple preselected positions in the low-zone weighted method show a downward trend as the preselected positions increase from small to large; According to the high-zone weighted method, the multiple coke index contribution parameters are weighted and summed to obtain the high-zone coke index adjustment parameter. The weight values ​​corresponding to the multiple pre-selected positions in the high-zone weighted method show an upward trend from small to large as the pre-selected position increases.

3. The method according to claim 2, characterized in that The feed coal index data includes a plurality of feed coal test index data, and the coke index data includes a plurality of coke test index data indicating the quality of the coke; The method of obtaining the minimum and maximum values ​​of the coal index data used as input data in all training samples of the coke quality prediction model includes: Obtaining a minimum value vector and a maximum value vector of the coal index data used as input data in all training samples of the coke quality prediction model, wherein the minimum value vector includes the minimum value of each type of coal test index data in all training samples, and the maximum value vector includes the maximum value of each type of coal test index data in all training samples; The method of calculating the derivative of the model function of the coke quality prediction model at each of the preselected positions based on the preselected positions in the numerical interval represented by the minimum value and the maximum value, and obtaining a plurality of coke index contribution parameters corresponding to the preselected positions, respectively, includes: Based on multiple preselected positions in the numerical interval represented by the minimum value vector and the maximum value vector, the derivatives of the model function of the coke quality prediction model at each of the preselected positions are calculated respectively, and multiple coke index contribution matrices corresponding to the multiple preselected positions are obtained. The elements in the coke index contribution matrix represent the influence of a unit change in the test index data of each type of coal entering the furnace on each type of coke test index data.

4. The method according to claim 2, characterized in that The plurality of preselected positions are evenly distributed in the numerical interval.

5. The method according to claim 3, characterized in that The calculation formula of the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter for adjusting the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical range is as follows: Y=Y X +△X l M l +△X h M h , where M l The parameter for adjusting the coke index in the low zone, M h is the high-zone coke index adjustment parameter, and Y is the adjusted coke quality prediction value; Y X = f(X`), wherein f() is the model function of the coke quality prediction model, and X` is obtained by changing the value of the multiple test index data X of the feed coal to be predicted that is less than the minimum test index data of the feed coal to the corresponding minimum value, and changing the value of the multiple test index data X of the feed coal to be predicted that is greater than the maximum test index data of the feed coal to the corresponding maximum value, while keeping the other test index data of the feed coal unchanged; △X l To XX min The result is that the elements greater than 0 are changed to 0, and the elements less than or equal to 0 remain unchanged, X min is the minimum value vector; △X h To XX max The result is that the elements less than 0 are changed to 0, and the elements greater than or equal to 0 remain unchanged. max is the maximum value vector.

6. A method for predicting coke quality, characterized in that: include: Obtaining the coal index data of the coal to be predicted; When the feed coal index data of the feed coal to be predicted exceeds the numerical range represented by the minimum value and the maximum value, based on the feed coal index data of the feed coal to be predicted, the coke quality prediction model is adopted, as well as the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter in any one of the methods described in claims 1-5, to predict the coke quality of coking using the feed coal to be predicted; The minimum value and the maximum value are the minimum value and the maximum value of the coal index data used as input data in all training samples of the coke quality prediction model.

7. A parameter generation device for coke quality prediction, characterized in that: include: A numerical acquisition module is used to obtain the minimum and maximum values ​​of the coal index data used as input data in all training samples of the coke quality prediction model; a contribution parameter calculation module, for calculating a coke index contribution parameter based on the numerical range represented by the minimum value and the maximum value, and a model function of the coke quality prediction model, wherein the coke index contribution parameter represents the influence of a unit change of the feed coal index data on the coke index data, the input of the model function is the feed coal index data, and the output of the model function is the coke index data; An adjustment parameter determination module is used to determine the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter based on the coke index contribution parameter, and the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter are used to adjust the coke quality prediction value output when the input of the coke quality prediction model exceeds the numerical range.

8. A coke quality prediction device, characterized in that: include: A data acquisition module is used to acquire the coal index data of the coal to be predicted; A quality prediction module, used for predicting the quality of coke produced by coking using the to-be-predicted coal for coaling, based on the coke quality prediction model and the low-zone coke index adjustment parameter and the high-zone coke index adjustment parameter in the method described in any one of claims 1 to 5, when the coal index data of the to-be-predicted coal for coaling exceeds the numerical range represented by the minimum value and the maximum value; The minimum value and the maximum value are the minimum value and the maximum value of the coal index data used as input data in all training samples of the coke quality prediction model.

9. An electronic device, characterized in that: The invention comprises a processor and a machine-readable storage medium, wherein the machine-readable storage medium stores machine-executable instructions that can be executed by the processor, and the processor is prompted by the machine-executable instructions to implement any one of the methods described in claims 1-5, or to implement the method described in claim 6.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in any one of claims 1 to 5 is implemented, or the method described in claim 6 is implemented.

Images