A method for optimizing pellet roasting system

Abstract

The present invention provides a method for optimizing a pellet roasting system, comprising obtaining process data of a current pellet, using the raw material composition of the current pellet and the raw material ratio of the current pellet in the process data to determine the theoretical component content of the current pellet, determining the theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet, calculating a first similarity between the theoretical slag phase index of the current pellet and the theoretical slag phase index of each sample pellet using multiple similarity algorithms and the weight of each similarity algorithm, and based on the correlation between the theoretical slag phase index of the sample pellets, the quality grade of the sample pellets and the roasting system of the sample pellets, selecting the roasting system of the sample pellet corresponding to the highest sample pellet quality grade under the theoretical slag phase index of the sample pellet with the maximum first similarity as the roasting system of the current pellet, thereby achieving rapid and accurate adjustment of the pellet roasting system and improving the production efficiency and quality of the pellets.

Description

Technical Field

[0001] The present invention relates to the technical field of pellet roasting, and in particular to a method for optimizing a pellet roasting system. Background Art

[0002] Pellets are an important raw material for ironmaking, and their quality has a significant impact on the ironmaking process. Changes in pellet composition and roasting schedules can lead to significant fluctuations in pellet quality.

[0003] In the current actual production of pellets, the composition of the pellets is controlled within a range according to the requirements of the final product, but the adjustment of the pellet roasting system often relies on manual experience, which requires a long cycle and high cost to adapt to the raw material conditions of the pellets, and the error rate is high.

[0004] Therefore, how to quickly and accurately adjust the pellet roasting system is a technical problem that needs to be solved urgently by those skilled in the art. Summary of the Invention

[0005] The present invention provides a method for optimizing a pellet roasting system, which is used to solve the defects of low adjustment efficiency and low accuracy of the pellet roasting system in the prior art.

[0006] In one aspect, the present invention provides a method for optimizing a pellet roasting system, comprising:

[0007] Acquire process data of the current pelletizing process; wherein the process data of the current pelletizing process includes the raw material composition of the current pelletizing process and the raw material ratio of the current pelletizing process;

[0008] Determining the theoretical component content of the current pellets according to the raw material components of the current pellets and the raw material ratio of the current pellets;

[0009] Determining a theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet;

[0010] Calculating a first similarity between the theoretical slag phase index of the current pellet and the theoretical slag phase index of each sample pellet using a plurality of similarity algorithms and a weight of each similarity algorithm;

[0011] Based on the correlation between the theoretical slag phase index of the sample pellets, the quality grade of the sample pellets and the roasting system of the sample pellets, the roasting system of the sample pellets corresponding to the highest quality grade of the sample pellets under the theoretical slag phase index of the sample pellets with the maximum first similarity is selected as the roasting system of the current pellets.

[0012] According to a pellet roasting system optimization method provided by the present invention, the process of setting the weight of each similarity algorithm includes:

[0013] Setting multiple weight configuration schemes for the multiple similarity algorithms, selecting the weight configuration scheme with the highest accuracy for configuration, and obtaining the weight of each similarity algorithm; wherein each weight configuration scheme includes the configuration weight of each similarity algorithm;

[0014] The process of selecting the weight configuration scheme with the highest accuracy includes:

[0015] Using each of the similarity algorithms, respectively calculating a second similarity between the theoretical slag phase index of the target training pellet and the theoretical slag phase index of each of the sample pellets;

[0016] Performing a comprehensive calculation based on each second similarity and the respective configuration weight to determine a third similarity;

[0017] Select the highest quality grade of the sample pellet under the theoretical slag phase index of the sample pellet with the maximum comprehensive similarity;

[0018] Calculating the absolute difference between the quality grade of the selected highest sample pellet and the quality grade of the training pellet under the theoretical slag phase index of the target training pellet;

[0019] The weight configuration scheme corresponding to the minimum absolute difference is used as the weight configuration scheme with the highest accuracy.

[0020] According to a pellet roasting system optimization method provided by the present invention, multiple weight configuration schemes are set for the multiple similarity algorithms, including:

[0021] Setting respective initial weights for each similarity algorithm;

[0022] Taking the respective initial weights as a reference and adjusting the respective initial weights according to a preset step size, the respective initial weights are obtained to obtain a plurality of adjusted weights for each similarity algorithm;

[0023] generating the plurality of weight configuration schemes based on the respective initial weights and the respective plurality of adjusted weights;

[0024] Wherein, the sum of the respective configuration weights in each weight configuration scheme is 1.

[0025] A method for optimizing a pellet roasting system according to the present invention further includes:

[0026] Acquire historical pelletizing process data; wherein the historical pelletizing process data includes raw material composition of the historical pelletizing, raw material ratio of the historical pelletizing, quality grade of the historical pelletizing, and roasting system of the historical pelletizing;

[0027] Determining the theoretical component content of the historical pellets according to the raw material composition of the historical pellets and the raw material ratio of the historical pellets;

[0028] Determining a theoretical slag phase index of the historical pellets based on the theoretical component content of the historical pellets;

[0029] Matching the theoretical slag phase index of the historical pellets, the quality grade of the historical pellets, and the roasting system of the historical pellets, establishing a correlation between the theoretical slag phase index of the historical pellets, the quality grade of the historical pellets, and the roasting system of the historical pellets, and forming the historical pellet database;

[0030] According to a set ratio, the historical pellet database is divided into a sample database and a training database, wherein the sample database includes the correlation between the theoretical slag phase index of the sample pellets, the quality grade of the sample pellets, and the roasting system of the sample pellets, and the training database includes the correlation between the theoretical slag phase index of the training pellets, the quality grade of the training pellets, and the roasting system of the training pellets.

[0031] According to a method for optimizing a pellet roasting system provided by the present invention, the theoretical component content of the current pellet is determined based on the raw material composition and the raw material ratio of the current pellet, including:

[0032] Obtaining the content of the raw material components of the current pellet in each raw material;

[0033] Calculate the product of the content of each current pellet's raw material component and the corresponding raw material ratio;

[0034] The sum of all products is calculated to obtain the theoretical component content of the current pellet.

[0035] According to a method for optimizing a pellet roasting system provided by the present invention, the method determines a theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet, including:

[0036] The theoretical component content of the current pellet, the production temperature of the current pellet, the production pressure of the current pellet and the production gas ratio of the current pellet are input into a thermodynamic calculation model for calculation to obtain the theoretical slag phase index of the current pellet.

[0037] According to a method for optimizing a pellet roasting system provided by the present invention, the method determines a theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet, including:

[0038] The theoretical component content of the current pellet, the production temperature of the current pellet, the production pressure of the current pellet and the production gas ratio of the current pellet are input into a pre-built regression model for calculation to obtain the theoretical slag phase index of the current pellet.

[0039] According to a pellet roasting system optimization method provided by the present invention, the component process of the regression model includes:

[0040] Set the variation range and variation step of the raw material component content of all historical pellets to form a component content variation matrix;

[0041] Calculate the theoretical slag phase index corresponding to each row of the component content in the component content change matrix using a thermodynamic calculation model based on the historical pellet production temperature, the historical pellet production pressure, and the historical pellet production gas ratio, to form a component content and theoretical slag phase index data set;

[0042] According to the component content and theoretical slag phase index data set, with the component content as input and the theoretical slag phase index as the target, a machine learning model is trained to obtain the trained model as the regression model.

[0043] On the other hand, the present invention also provides a pellet roasting system optimization system, which includes:

[0044] An acquisition module, configured to acquire process data of the current pelletizing process; wherein the process data of the current pelletizing process includes the raw material composition and the raw material ratio of the current pelletizing process;

[0045] a determination module, configured to determine the theoretical component content of the current pellet according to the raw material composition of the current pellet and the raw material ratio of the current pellet; and determine the theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet;

[0046] a calculation module, configured to calculate the similarity between the theoretical slag phase index of the current pellet and the theoretical slag phase index of each sample pellet using a plurality of similarity algorithms and a weight of each similarity algorithm;

[0047] The optimization module is used to select, based on the correlation between the theoretical slag phase index of the sample pellets, the quality grade of the sample pellets and the roasting system of the sample pellets, the roasting system of the sample pellets corresponding to the highest sample pellet quality grade under the theoretical slag phase index of the sample pellets with the maximum first similarity as the roasting system of the current pellets.

[0048] On the other hand, the present invention also provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the above-described methods for optimizing the pellet roasting system is implemented.

[0049] On the other hand, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for optimizing the pellet roasting schedule.

[0050] On the other hand, the present invention further provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for optimizing the pellet roasting schedule as described above is implemented.

[0051] The method for optimizing the pellet roasting system provided by the present invention obtains the process data of the current pellet, and uses the raw material composition of the current pellet and the raw material ratio of the current pellet in the process data to determine the theoretical component content of the current pellet, and then determines the theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet, and then uses multiple similarity algorithms and the weight of each similarity algorithm to calculate the first similarity between the theoretical slag phase index of the current pellet and the theoretical slag phase index of each sample pellet. Further, based on the correlation between the theoretical slag phase index of the sample pellet, the quality grade of the sample pellet and the roasting system of the sample pellet, the roasting system of the sample pellet corresponding to the highest sample pellet quality grade under the theoretical slag phase index of the sample pellet with the maximum first similarity is selected as the roasting system of the current pellet, thereby realizing rapid and accurate adjustment of the pellet roasting system and improving the production efficiency and quality of the pellets. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0053] Figure 1 1 is a flow chart of a method for optimizing a pellet roasting system according to an embodiment of the present invention;

[0054] Figure 2 It is a flow chart of a method for selecting the weight configuration scheme with the highest accuracy;

[0055] Figure 3 It is a structural diagram of a pellet roasting system optimization method provided by an embodiment of the present invention;

[0056] Figure 4 It is a structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0057] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0058] Figure 1 It is a flow chart of the method for optimizing the pellet roasting system provided in an embodiment of the present invention.

[0059] like Figure 1 As shown, the pellet roasting system optimization method provided by the embodiment of the present invention can be executed by an electronic device, and the method mainly includes the following steps:

[0060] 101. Obtain the current pelletizing process data;

[0061] In a specific implementation process, the process data of the current pellets can be obtained according to the requirements of the final product, etc. The process data of the current pellets can include the raw material composition and the raw material ratio of the current pellets.

[0062] 102. Determine the theoretical component content of the current pellet according to the raw material composition and the raw material ratio of the current pellet;

[0063] In a specific implementation, pelletizing raw materials may contain several to more than a dozen mineral powders, and the number of components to be calculated is usually around 10. The collected pelletizing raw material components include, but are not limited to, the percentage content of iron (Fe2O3), silicon (SiO2), aluminum (Al2O3), calcium (CaO), magnesium (MgO), sulfur (S), phosphorus (P), and other components. For each mineral powder, a detailed composition table should be prepared.

[0064] Therefore, in this embodiment, the content of the raw material components of the current pellet in each raw material can be obtained; the content of the raw material components of each current pellet and the product of the corresponding raw material ratio are calculated; and the sum of all products is calculated to obtain the theoretical component content of the current pellet.

[0065] The specific calculation process can be found in formula (1):

[0066] (1)

[0067] in, Indicates the theoretical composition content of the current pellets, Indicates the content of the raw material component of the current pellet in the i-th raw material, It represents the product of the raw material ratios of the i-th raw material, and n represents the number of raw materials.

[0068] 103. Determine a theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet;

[0069] In a specific implementation, if a thermodynamic calculation model is available, the theoretical component content of the current pellet, the current pellet production temperature, the current pellet production pressure, and the current pellet production gas ratio can be input into the thermodynamic calculation model for calculation to obtain the theoretical slag phase index of the current pellet. The theoretical slag phase index of the pellet includes the pellet slag phase amount, the slag phase melting start temperature, the slag phase complete melting temperature, the slag phase viscosity, etc. The calculation process can refer to existing related technologies and will not be repeated here.

[0070] If there is no thermodynamic calculation model, the theoretical component content of the current pellet, the current pellet production temperature, the current pellet production pressure, and the current pellet production gas ratio can be input into a pre-built regression model for calculation to obtain the theoretical slag phase index of the current pellet. The construction process of the regression model includes the following steps:

[0071] a. Set the variation range and variation step of the raw material content of all historical pellets to form a component content variation matrix;

[0072] b. Calculating the theoretical slag phase index corresponding to each row of the component content in the component content change matrix using a thermodynamic calculation model based on the historical pellet production temperature, the historical pellet production pressure, and the historical pellet production gas ratio, thereby forming a component content and theoretical slag phase index data set;

[0073] c. Based on the component content and theoretical slag phase index data set, with the component content as input and the theoretical slag phase index as the target, a machine learning model is trained to obtain the trained model as the regression model.

[0074] In other words, historical pelletizing process data can be collected through big data. If a thermodynamic calculation model is available, the theoretical slag phase indicators of the historical pellets can be calculated using the thermodynamic calculation model. Machine learning can then be used to train a regression model. This regression model can then be used directly in the future, rather than the thermodynamic calculation model. The training method can be a random forest approach, for example.

[0075] 104. Calculate a first similarity between the theoretical slag phase index of the current pellet and the theoretical slag phase index of each sample pellet using multiple similarity algorithms and the weight of each similarity algorithm;

[0076] In a specific implementation process, for the theoretical slag phase index of any sample pellet, different similarity algorithms can be used to calculate the individual similarities between the theoretical slag phase index of the current pellet and the theoretical slag phase index of the sample pellet, and then the product of each individual similarity and the weight of the corresponding similarity algorithm is calculated, and the sum of all products is used as the first similarity between the theoretical slag phase index of the current pellet and the theoretical slag phase index of the sample pellet. After traversing the theoretical slag phase indicators of all sample pellets, the first similarity between the theoretical slag phase index of the current pellet and the theoretical slag phase index of each sample pellet can be obtained.

[0077] 105. Based on the correlation between the theoretical slag phase index of the sample pellets, the quality grade of the sample pellets, and the roasting system of the sample pellets, the roasting system of the sample pellet corresponding to the highest quality grade of the sample pellets under the theoretical slag phase index of the sample pellets with the maximum first similarity is selected as the roasting system of the current pellet.

[0078] In a specific implementation process, a correlation between the theoretical slag phase index of the sample pellet, the quality grade of the sample pellet, and the roasting system of the sample pellet can be pre-constructed. Among them, in this correlation, under the theoretical slag phase index of the sample pellet, the quality grade of the sample pellet and the roasting system of the sample pellet are both relatively good, which is used to guide on-site personnel to produce pellets. Therefore, after obtaining the similarity between the theoretical slag phase index of the current pellet and the theoretical slag phase index of each sample pellet, the theoretical slag phase index of the sample pellet with the maximum similarity can be selected, and by searching the correlation, the quality grade of the highest sample pellet is selected, and further the roasting system of the sample pellet corresponding to the quality of the highest sample pellet is selected as the roasting system of the current pellet. In this way, a relatively good roasting system can be provided for subsequent pellet production, and the roasting system has been used, and its use method is easier to explain and understand, so that on-site engineers can use it flexibly.

[0079] In a specific implementation process, the correlation between the theoretical slag phase index of the sample pellet, the quality grade of the sample pellet, and the roasting system of the sample pellet can be obtained as follows:

[0080] a1. Obtain historical pelletizing process data;

[0081] The historical pelletizing process data includes the raw material composition, raw material ratio, quality grade, and roasting system of the historical pellets. The manufacturer's data can be obtained through big data, thereby improving the comprehensiveness of data collection and increasing data utilization.

[0082] b1. Determine the theoretical component content of the historical pellets according to the raw material composition and raw material ratio of the historical pellets;

[0083] This process can be implemented by referring to the aforementioned related content and will not be repeated here.

[0084] c1. Determining the theoretical slag phase index of the historical pellets based on the theoretical component content of the historical pellets;

[0085] This process can be implemented by referring to the aforementioned related content and will not be repeated here.

[0086] d1. Matching the theoretical slag phase index of the historical pellets, the quality grade of the historical pellets, and the roasting system of the historical pellets, establishing a correlation between the theoretical slag phase index of the historical pellets, the quality grade of the historical pellets, and the roasting system of the historical pellets, and forming the historical pellet database;

[0087] In a specific implementation, the theoretical slag phase index of historical pellets, the quality grade of the historical pellets, and the roasting system of the historical pellets can be matched according to the time node, and a correlation relationship between the theoretical slag phase index of the historical pellets, the quality grade of the historical pellets, and the roasting system of the historical pellets can be established to form the historical pellet database. The theoretical slag phase index of a historical pellet can correspond to multiple quality grades of historical pellets, and each quality grade of a historical pellet corresponds to a roasting system of a historical pellet.

[0088] e1. Divide the historical pellet database into the sample database and the training database according to a set ratio.

[0089] In a specific implementation process, after the sample data and the training database are divided, the correlation between the theoretical slag phase index of the historical pellets, the quality grade of the historical pellets and the roasting system of the historical pellets in the sample database can be called the correlation between the theoretical slag phase index of the sample pellets, the quality grade of the sample pellets and the roasting system of the sample pellets, and the correlation between the theoretical slag phase index of the historical pellets, the quality grade of the historical pellets and the roasting system of the historical pellets in the training database can be called the correlation between the theoretical slag phase index of the training pellets, the quality grade of the training pellets and the roasting system of the training pellets.

[0090] The pellet roasting system optimization method of this embodiment obtains the process data of the current pellet, and uses the raw material composition of the current pellet and the raw material ratio of the current pellet in the process data to determine the theoretical component content of the current pellet, and then determines the theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet, and then uses multiple similarity algorithms and the weight of each similarity algorithm to calculate the first similarity between the theoretical slag phase index of the current pellet and the theoretical slag phase index of each sample pellet. Further, based on the correlation between the theoretical slag phase index of the sample pellet, the quality grade of the sample pellet and the roasting system of the sample pellet, the roasting system of the sample pellet corresponding to the highest sample pellet quality grade under the theoretical slag phase index of the sample pellet with the maximum first similarity is selected as the roasting system of the current pellet, thereby achieving rapid and accurate adjustment of the pellet roasting system and improving the production efficiency and quality of the pellets.

[0091] In a specific implementation process, the weight setting process of each similarity algorithm includes:

[0092] Multiple weight allocation schemes are set for the multiple similarity algorithms, and the weight allocation scheme with the highest accuracy is selected for allocation to obtain weights for each similarity algorithm; wherein each weight allocation scheme includes a configuration weight for each similarity algorithm. In other words, different weights are assigned to each similarity algorithm, and each weight allocation scheme will calculate the theoretical slag phase index of the sample pellet that is closest to the theoretical slag phase index of the target training pellet. At this time, it is necessary to determine which weight allocation scheme has the highest accuracy, and thus select that weight allocation scheme to configure the weights for each similarity algorithm.

[0093] In a specific implementation process, respective initial weights can be set for each similarity algorithm; based on the respective initial weights, the respective initial weights are adjusted according to a preset step size to obtain respective multiple adjusted weights for each similarity algorithm; based on the respective initial weights and the respective multiple adjusted weights, the multiple weight configuration schemes are generated; wherein the sum of the respective configuration weights in each weight configuration scheme is 1.

[0094] In a specific implementation process, Figure 2 It is a flowchart of the method for selecting the weight configuration scheme with the highest accuracy, such as Figure 2 As shown, the method may include the following steps:

[0095] 201. Calculate, using each of the similarity algorithms, a second similarity between a theoretical slag phase index of a target training pellet and a theoretical slag phase index of each of the sample pellets;

[0096] The second similarity here is the individual similarity calculated by each similarity algorithm.

[0097] 202. Perform comprehensive calculation based on each second similarity and the respective configuration weight to determine a third similarity;

[0098] The third similarity here is the comprehensive similarity obtained after comprehensive calculation.

[0099] 203. Select the highest quality grade of the sample pellet under the theoretical slag phase index of the sample pellet with the maximum third similarity;

[0100] 204. Calculate the absolute difference between the quality grade of the selected highest sample pellet and the quality grade of the training pellet under the theoretical slag phase index of the target training pellet;

[0101] 205. The weight configuration scheme corresponding to the minimum absolute difference is used as the weight configuration scheme with the highest accuracy.

[0102] In a specific implementation process, the above steps 201 to 205 are described with the following example:

[0103] Assume that three similarity calculation methods are used: cosine similarity, Euclidean distance, and Jaccard coefficient. Assume that the theoretical slag phase index of the target training pellet is labeled A and the theoretical slag phase index of the sample pellet is labeled B, and a set of initial weight assignments: cosine similarity weight is 0.5, Euclidean distance weight is 0.3, and Jaccard coefficient weight is 0.2. Calculate the cosine similarity between A and B, and let the result be cos_sim. Calculate the Euclidean distance between A and B, convert it to a similarity form (such as taking the inverse or negative exponent), and let the result be euc_sim. Calculate the Jaccard coefficient between A and B, and let the result be jac_sim. According to the current weight assignment scheme, the comprehensive similarity similarity_score can be calculated by the following calculation formula (2):

[0104] (2)

[0105] The weights of each similarity algorithm are changed with a certain step size (such as 0.1), while ensuring that the sum of the weights of each method is always 1, and the comprehensive similarity calculation is repeated to obtain the comprehensive similarity corresponding to each weight distribution scheme.

[0106] According to the comprehensive similarity, the theoretical slag phase index Z_sample of the sample pellet that is closest to the theoretical slag phase index Z_train of the target training pellet in the training library is obtained, the pellet mass P_train corresponding to Z_train and the pellet mass P_sample corresponding to Z_sample are obtained, the absolute value of the difference Error between P_train and P_sample is calculated, the Errors corresponding to different comprehensive similarities are compared, and the similarity method weight allocation scheme with the smallest Error is obtained as the weight configuration scheme with the highest accuracy.

[0107] In a specific implementation process, the above method realizes the configuration of a better weight for each similarity algorithm. However, as time goes by, the weight of each similarity algorithm may no longer be suitable. In order to avoid the need to re-adjust the weight according to the Figure 2 The process shown configures weights for each similarity algorithm. In this embodiment, the following technical solutions are also provided:

[0108] In the first period after configuring the weights for each similarity algorithm, each time according to the row Figure 1 After executing the process shown, after selecting the theoretical slag phase index of the sample pellet with the greatest first similarity, a fourth similarity between the theoretical slag phase index of the current pellet calculated by each similarity algorithm and the theoretical slag phase index of the sample pellet with the greatest first similarity can be obtained. The difference between each fourth similarity and the greatest first similarity is calculated, and the difference is compared with a preset difference. The number of times the difference corresponding to each similarity algorithm is less than the preset difference is recorded to determine the contribution value of each similarity algorithm to the theoretical slag phase index of the sample pellet with the greatest first similarity. The greater the number of similarity algorithms, the greater the contribution value. When the upper limit of the first time period is exceeded, the weight of each similarity algorithm can be adjusted based on its contribution value.

[0109] In a specific implementation process, the weight of each similarity algorithm can be adjusted as follows:

[0110] The probability value of each contribution value can be calculated. According to the probability value of each contribution value and the weight of each similarity algorithm, the weighted contribution value of each similarity algorithm can be obtained. According to the weighted contribution value of each similarity algorithm, the overall weighted contribution value can be obtained. The ratio of the weighted contribution value of each similarity algorithm to the overall weighted contribution value is used as the dynamic adjustment weight of each similarity algorithm.

[0111] For example, if the cosine similarity weight is 0.5, the Euclidean distance weight is 0.3, and the Jaccard coefficient weight is 0.2; the probability of the cosine similarity contribution is 0.8, the probability of the Euclidean distance contribution is 0.1, and the probability of the Jaccard coefficient contribution is 0.1. Then, the weighted contribution of cosine similarity is 0.5*0.8=0.4, the weighted contribution of Euclidean distance is 0.3*0.1=0.03, and the weighted contribution of Jaccard coefficient is 0.2*0.1=0.02. The total weighted contribution is 0.4+0.03+0.02=0.45. The dynamically adjusted weights of cosine similarity are 0.4 / 0.45≈0.89, Euclidean distance is 0.03 / 0.45≈0.07, and Jaccard coefficient is 0.02 / 0.45≈0.04.

[0112] For example, if the cosine similarity weight is 0.6, the Euclidean distance weight is 0.3, and the Jaccard coefficient weight is 0.1; the probability value of the cosine similarity contribution value is 0.7, the probability value of the Euclidean distance contribution value is 0.2, and the probability value of the Jaccard coefficient contribution value is 0.1. Then the weighted contribution value of cosine similarity is 0.6*0.7=0.42, the weighted contribution value of Euclidean distance is 0.3*0.2=0.06, and the weighted contribution value of Jaccard coefficient is 0.1*0.1=0.01. The total weighted contribution value is 0.42+0.06+0.01=0.49. The dynamically adjusted weights of cosine similarity are 0.42 / 0.49≈0.86, the dynamically adjusted weights of Euclidean distance are 0.06 / 0.49≈0.12, and the dynamically adjusted weights of Jaccard coefficient are 0.01 / 0.49≈0.02.

[0113] It should be noted that the above examples are merely illustrative, and other methods may also be used, which are not listed one by one here.

[0114] In a specific implementation process, although the above method can dynamically adjust the weight of each similarity algorithm, in order to further ensure that the optimal roasting system can be selected, in this embodiment, after the second time period after dynamically adjusting the weight of each similarity algorithm, the roasting system can be adjusted according to the following method: Figure 2 The process re-determines the weight of each similarity algorithm again. That is, two time windows are set. After the first time window, the contribution value of each similarity algorithm is used to dynamically adjust the weight of each similarity algorithm. After the second time window, the contribution value of each similarity algorithm is used to dynamically adjust the weight of each similarity algorithm. Figure 2 The process adjusts the weight of each similarity algorithm again, so that a better roasting system can be selected.

[0115] Based on the same general inventive concept, the present invention also protects a pellet roasting system optimization method. The pellet roasting system optimization method provided by the present invention is described below. The pellet roasting system optimization method described below and the pellet roasting system optimization method described above can be referenced to each other.

[0116] Figure 3 Schematic diagram of the structure of the pellet roasting system optimization method provided by the embodiment of the present invention. Figure 3 As shown, the pellet roasting schedule optimization system of this embodiment includes an acquisition module 31 , a determination module 32 , a calculation module 33 and an optimization module 34 .

[0117] The acquisition module 31 is used to acquire the process data of the current pelletizing process; wherein the process data of the current pelletizing process includes the raw material composition of the current pelletizing process and the raw material ratio of the current pelletizing process;

[0118] A determination module 32 is configured to determine the theoretical component content of the current pellet according to the raw material composition and the raw material ratio of the current pellet; and determine the theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet;

[0119] a calculation module 33 for calculating a first similarity between the theoretical slag phase index of the current pellet and the theoretical slag phase index of each sample pellet using a plurality of similarity algorithms and a weight of each similarity algorithm;

[0120] The optimization module 34 is used to select, based on the correlation between the theoretical slag phase index of the sample pellets, the quality grade of the sample pellets and the roasting system of the sample pellets, the roasting system of the sample pellets corresponding to the highest sample pellet quality grade under the theoretical slag phase index of the sample pellets with the maximum first similarity as the roasting system of the current pellets.

[0121] Figure 4 4 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The pellet roasting schedule optimization method system may include: a processor (processor) 410, a communication interface (communications interface) 420, a memory (memory) 430 and a communication bus 440. The processor 410, the communication interface 420 and the memory 430 communicate with each other via the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the pellet roasting schedule optimization method.

[0122] Furthermore, the logic instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0123] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the pellet roasting system optimization method provided by the above methods.

[0124] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the pellet roasting schedule optimization method provided by the above methods.

[0125] It should be noted that the relevant information that may be involved in the various embodiments of this application are all strictly in accordance with the requirements of laws and regulations, follow the principles of legality, legitimacy and necessity, and are based on the reasonable purposes of business scenarios to process information that users actively provide during the use of products / services or generated due to the use of products / services, as well as information obtained with user authorization.

[0126] The information processed by this application will vary depending on the specific product / service scenario and should be based on the specific scenario in which the user uses the product / service. This may involve the user's account information, device information, or other related information. This application will treat the user's personal information and its processing with a high degree of diligence.

[0127] This application attaches great importance to the security of relevant information and has taken reasonable and feasible security protection measures that comply with industry standards to protect relevant information and prevent unauthorized access, public disclosure, use, modification, damage or loss of relevant information.

[0128] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0129] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for optimizing pellet roasting system, characterized in that: include: Acquire process data of the current pelletizing process; wherein the process data of the current pelletizing process includes the raw material composition of the current pelletizing process and the raw material ratio of the current pelletizing process; Determining the theoretical component content of the current pellets according to the raw material components of the current pellets and the raw material ratio of the current pellets; Determining a theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet; Calculating a first similarity between the theoretical slag phase index of the current pellet and the theoretical slag phase index of each sample pellet using a plurality of similarity algorithms and a weight of each similarity algorithm; Based on the correlation between the theoretical slag phase index of the sample pellets, the quality grade of the sample pellets, and the roasting system of the sample pellets, the roasting system of the sample pellet corresponding to the highest quality grade of the sample pellets under the theoretical slag phase index of the sample pellets with the greatest first similarity is selected as the roasting system of the current pellet; The weight setting process of each similarity algorithm includes: Setting multiple weight configuration schemes for the multiple similarity algorithms, selecting the weight configuration scheme with the highest accuracy for configuration, and obtaining the weight of each similarity algorithm; wherein each weight configuration scheme includes the configuration weight of each similarity algorithm; The process of selecting the weight configuration scheme with the highest accuracy includes: Using each of the similarity algorithms, respectively calculating a second similarity between the theoretical slag phase index of the target training pellet and the theoretical slag phase index of each of the sample pellets; Performing a comprehensive calculation based on each second similarity and the respective configuration weight to determine a third similarity; Select the highest quality grade of the sample pellet under the theoretical slag phase index of the sample pellet with the greatest third similarity; Calculating the absolute difference between the quality grade of the selected highest sample pellet and the quality grade of the training pellet under the theoretical slag phase index of the target training pellet; The weight configuration scheme corresponding to the minimum absolute difference is used as the weight configuration scheme with the highest accuracy.

2. The method for optimizing the pellet roasting system according to claim 1, characterized in that: Multiple weight configuration schemes are set for the multiple similarity algorithms, including: Setting respective initial weights for each similarity algorithm; Taking the respective initial weights as a reference and adjusting the respective initial weights according to a preset step size, the respective initial weights are obtained to obtain a plurality of adjusted weights for each similarity algorithm; generating the plurality of weight configuration schemes based on the respective initial weights and the respective plurality of adjusted weights; Wherein, the sum of the respective configuration weights in each weight configuration scheme is 1.

3. The method for optimizing the pellet roasting system according to claim 1, characterized in that: Also includes: Acquire historical pelletizing process data; wherein the historical pelletizing process data includes raw material composition of the historical pelletizing, raw material ratio of the historical pelletizing, quality grade of the historical pelletizing, and roasting system of the historical pelletizing; Determining the theoretical component content of the historical pellets according to the raw material composition of the historical pellets and the raw material ratio of the historical pellets; Determining a theoretical slag phase index of the historical pellets based on the theoretical component content of the historical pellets; Matching the theoretical slag phase index of the historical pellets, the quality grade of the historical pellets, and the roasting system of the historical pellets, establishing a correlation between the theoretical slag phase index of the historical pellets, the quality grade of the historical pellets, and the roasting system of the historical pellets, and forming the historical pellet database; According to a set ratio, the historical pellet database is divided into a sample database and a training database, wherein the sample database includes the correlation between the theoretical slag phase index of the sample pellets, the quality grade of the sample pellets, and the roasting system of the sample pellets, and the training database includes the correlation between the theoretical slag phase index of the training pellets, the quality grade of the training pellets, and the roasting system of the training pellets.

4. The method for optimizing the pellet roasting system according to claim 1, characterized in that: Determining the theoretical component content of the current pellets according to the raw material components and the raw material ratio of the current pellets includes: Obtaining the content of the raw material components of the current pellet in each raw material; Calculate the product of the content of each current pellet's raw material component and the corresponding raw material ratio; The sum of all products is calculated to obtain the theoretical component content of the current pellet.

5. The method for optimizing the pellet roasting system according to any one of claims 1 to 4, characterized in that: Determining a theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet includes: The theoretical component content of the current pellet, the production temperature of the current pellet, the production pressure of the current pellet and the production gas ratio of the current pellet are input into a thermodynamic calculation model for calculation to obtain the theoretical slag phase index of the current pellet.

6. The method for optimizing the pellet roasting system according to any one of claims 1 to 4, characterized in that: Determining a theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet includes: The theoretical component content of the current pellet, the production temperature of the current pellet, the production pressure of the current pellet and the production gas ratio of the current pellet are input into a pre-built regression model for calculation to obtain the theoretical slag phase index of the current pellet.

7. The method for optimizing the pellet roasting system according to claim 6, characterized in that: The process of building the regression model includes: Set the variation range and variation step of the raw material component content of all historical pellets to form a component content variation matrix; Calculate the theoretical slag phase index corresponding to each row of the component content in the component content change matrix using a thermodynamic calculation model based on the historical pellet production temperature, the historical pellet production pressure, and the historical pellet production gas ratio, to form a component content and theoretical slag phase index data set; According to the component content and theoretical slag phase index data set, with the component content as input and the theoretical slag phase index as the target, a machine learning model is trained to obtain the trained model as the regression model.

8. A pellet roasting system optimization system, characterized in that: include: An acquisition module, configured to acquire process data of the current pelletizing process; wherein the process data of the current pelletizing process includes the raw material composition and the raw material ratio of the current pelletizing process; a determination module, configured to determine the theoretical component content of the current pellet according to the raw material composition of the current pellet and the raw material ratio of the current pellet; and determine the theoretical slag phase index of the current pellet based on the theoretical component content of the current pellet; A calculation module is used to calculate the first similarity between the theoretical slag phase index of the current pellet and the theoretical slag phase index of each sample pellet using multiple similarity algorithms and the weight of each similarity algorithm; wherein the setting process of the weight of each similarity algorithm includes: setting multiple weight configuration schemes for the multiple similarity algorithms, selecting the weight configuration scheme with the highest accuracy for configuration, and obtaining the weight of each similarity algorithm; wherein each weight configuration scheme includes the respective configuration weight of each similarity algorithm; wherein the process of selecting the weight configuration scheme with the highest accuracy includes: using each similarity algorithm to respectively calculate the second similarity between the theoretical slag phase index of the target training pellet and the theoretical slag phase index of each sample pellet; performing a comprehensive calculation based on each second similarity and the respective configuration weight to determine the third similarity; selecting the quality grade of the highest sample pellet under the theoretical slag phase index of the sample pellet with the maximum third similarity; calculating the absolute difference between the quality grade of the selected highest sample pellet and the quality grade of the training pellet under the theoretical slag phase index of the target training pellet; and using the weight configuration scheme corresponding to the minimum absolute difference as the weight configuration scheme with the highest accuracy; The optimization module is used to select, based on the correlation between the theoretical slag phase index of the sample pellets, the quality grade of the sample pellets and the roasting system of the sample pellets, the roasting system of the sample pellets corresponding to the highest sample pellet quality grade under the theoretical slag phase index of the sample pellets with the maximum first similarity as the roasting system of the current pellets.

9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and operable on the processor, wherein when the processor executes the program, the method for optimizing the pellet roasting schedule according to any one of claims 1 to 7 is implemented.

Images