Sintering fuel proportion optimization method and system based on data mining and memory
Based on data -based methods, the operating mode, breathability state and calorie level in the sintering production process, the construction of the fuel ratio optimization rules library, solve the problem of complexity and difficulty implementation of solid fuel ratio regulation in the existing technology. The convenience of operation and the stability of the production process is achieved.
Patent Information
- Application Number
- CN202411378002.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-30
- Publication Date
- 2025-01-28
AI Technical Summary
In the prior art, there are complexity and difficulty in implementing the regulation of solid fuel ratio during sintering production, and there are insufficient differences and timeliness in manual experience operations.
Adopting data -based mining methods, by obtaining the actual data of the sintering production process, identifying the operating mode, breathability state and calorie level, the fuel ratio optimization of the rule library is constructed, and the optimal fuel ratio control amount is output with fuzzy logic reasoning methods.
Effectively solve the problem of complexity and difficulty implementation of regulatory rules, provides a solution optimization solution with convenient operation and simple processes, and improves the stability and energy efficiency of the production process.
Smart Images
Figure CN119376245A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a sintering fuel ratio optimization method, in particular to a sintering fuel ratio optimization method based on data mining and a system and a memory thereof, and belongs to the field of iron ore sintering production. Background Art
[0002] As the first process in steel production, sintering accounts for about 8% of the total energy consumption of steel production, and carbon emissions account for about 17% of the total emissions of steel production, while solid fuel consumption accounts for 80% of the process energy consumption. Therefore, improving the regulation level of solid fuels and reducing the energy consumption and carbon emissions of sintering processes are of great significance to the intelligent, green, low-carbon and high-quality development of the steel industry.
[0003] At present, the solid fuel ratio in the sintering production process is mainly regulated and decided by the operators based on their production operation experience. The advantage of manual experience operation is that the regulation accuracy is high, but there are also differences in the experience accumulation and operation habits of different operators, as well as the lack of timeliness of regulation. Researchers have also tried to use artificial intelligence technologies such as neural networks and expert systems to predict and regulate fuel ratios. For example, scholars use principal component analysis to select parameters such as the rising point of exhaust gas temperature, the position of the sintering end point, and the machine speed as input variables, and establish a carbon ratio regulation model through the multivariate linear regression method. This method has poor interpretability and cannot be understood by operators. Some scholars have also proposed a relationship model between the optimal fuel ratio and the amount of returned ore to achieve rough adjustment of the fuel ratio. Combined with the expert rule adjustment table, the fuel ratio is fine-tuned according to the current average temperature of the burn-through point. The main difficulty of this type of fuel ratio expert system is the update of rules and the adaptability of expert rules to different production conditions and different production sites.
[0004] Therefore, a method and system for optimizing the sintering fuel ratio based on data mining is established, which can mine the fuel ratio optimization control rules according to the actual production data of different production sites. When the rule base is applied to the actual production site, the appropriate fuel ratio setting can be inferred according to the actual production situation, providing reasonable and highly explainable optimization suggestions for the production process, and providing powerful assistance for optimizing the production process and reducing production energy consumption. Summary of the invention
[0005] In view of the problems existing in the prior art, the first purpose of the present invention is to provide a method for optimizing the sintering fuel ratio based on data mining. The method obtains a rule base for optimizing the fuel ratio by means of data mining, and extracts the fuel ratio optimization control rules according to the actual production data of the sintering process, which effectively solves the problem that the control rules in the prior art are complex and difficult to implement on the production site, and has the advantages of convenient operation and simple process.
[0006] The second object of the present invention is to provide a computer-readable memory, which contains a computer program for implementing the above-mentioned uniformity control method and can be read and executed.
[0007] The third object of the present invention is to provide a sintering fuel ratio optimization system based on data mining. The system uses rule reasoning to give fuel ratio optimization and control suggestions, and can give corresponding explanations for the fuel ratio optimization and control suggestions, avoiding the disadvantage of poor solvability of the black box model. In the actual application process, it can provide operators with rich production condition information and reliable fuel ratio optimization suggestions.
[0008] To achieve the above technical purpose, a sintering fuel ratio optimization method based on data mining includes the following steps:
[0009] Step S1, obtaining the main operating parameters of the operating surface and identifying the current operating mode of sintering production;
[0010] Step S2, obtaining the main state parameters of the process surface and identifying the permeability state of the material layer in the current sintering process;
[0011] Step S3, obtaining the main thermal parameters of the equipment surface and classifying the thermal level of the sintering process;
[0012] Step S4, using the operation mode, air permeability state and heat level as rule antecedents, using fuzzy logic reasoning to match the corresponding control rules from the fuel ratio optimization rule library, and outputting the optimal fuel ratio control amount;
[0013] The main operating parameters of the operating surface include material layer thickness, trolley speed and main exhaust fan frequency;
[0014] The main state parameters of the process surface include the moisture content of the mixture, the negative pressure of the No. 1 wind box, the negative pressure of the large flue and the sintering end point position;
[0015] The main thermal parameters of the equipment surface include the exhaust gas temperature of the large flue, the exhaust gas temperature of the 22# wind box and the exhaust gas temperature of the ring cooling section I.
[0016] As a preferred solution, the process of identifying the sintering production operation mode in step S1 is: calculating the Euclidean distance between the main operation parameters of the operation surface and the cluster centers corresponding to different operation modes, and identifying the operation mode of the current sintering production.
[0017] As a preferred solution, the process of identifying the permeability state of the material layer in step S2 is: calculating the Euclidean distance between the main state parameters of the process surface and the cluster centers corresponding to different permeability states, and identifying the permeability state of the current sintering process.
[0018] As a preferred solution, the heat level classification process in step S3 is: calculating the Euclidean distance between the main heat parameters of the equipment surface and the cluster centers corresponding to different heat levels, and identifying the heat level of the current sintering material layer.
[0019] As a preferred solution, the calculation process of the Euclidean distance is:
[0020] Formula 1: In formula 1: x j =[x j1 ,x j2 ,x j3 ] is the normalized parameter vector to be identified, which is represented by the normalized value of the material layer thickness, trolley speed and main exhaust fan frequency in step S1, and so on, dimensionless; d ij is the Euclidean distance between the normalized parameter vector j and the i-th cluster center, dimensionless.
[0021] As a preferred solution, the cluster center is obtained from the analysis of historical production data through a fuzzy clustering algorithm, and the specific steps are: obtaining historical data of corresponding parameters, obtaining a data set after preprocessing, and then dividing it through a fuzzy clustering algorithm.
[0022] As a preferred solution, the fuel ratio optimization rule base described in step S4 is constructed by the following steps:
[0023] i) The fuel ratio in the historical production data is divided into five discrete intervals {NB, NS, ZO, PS, PB} using a trapezoidal fuzzy membership function;
[0024] ii) The association rules of fuel ratio intervals under different operation modes, permeability states and heat levels in historical data are mined through fuzzy association rule algorithm to form a fuel ratio optimization rule base.
[0025] As a preferred solution, the screening process of the optimal fuel ratio control amount is: taking the end point of the control rule output interval as the reference value to be output, normalizing the confidence of the rule as the proportion of the rule, and outputting the final setting value of the fuel ratio after defuzzification. The specific calculation process is as follows:
[0026] Formula 2: out=l1*Result(Rulexxx) mid +l2*Result(Ruleyyy) mid ;
[0027] Formula 3:
[0028] Formula 4:
[0029] In formulas 2 to 4, Rulexxx and Ruleyyy are the two rules with the highest support whose antecedents in the rule base are the same as the calculation results in steps S1 to S3 of requirement 1; the rule support is Support Rulexxx Support Ruleyyy , dimensionless; l1 and l2 are the results of support normalization, dimensionless.
[0030] The present invention also provides a computer-readable memory comprising a computer program, wherein the computer program can implement any of the above-mentioned methods for optimizing sintering fuel ratio based on data mining.
[0031] The present invention also provides a sintering fuel ratio optimization system based on data mining, comprising: a programmable logic controller, a processor and a readable memory as described in claim 9; the programmable logic controller collects production process data, and the processor executes a computer program on the readable memory.
[0032] Compared with the prior art, the beneficial technical effects of the technical solution of the present invention are:
[0033] 1) The optimization method provided by the present invention obtains a rule base for optimizing fuel ratio settings by means of data mining, and extracts fuel ratio optimization control rules based on actual production data of the sintering process, which effectively solves the problem in the prior art that the control rules are complex and difficult to implement on the production site, and has the advantages of convenient operation and simple process.
[0034] 2) In the technical solution provided by the present invention, rule reasoning is used to give fuel ratio optimization and control suggestions, and corresponding explanations can be given for the fuel ratio optimization and control suggestions, avoiding the shortcoming of poor solvability of the black box model. In actual application, it can provide operators with rich production condition information and reliable fuel ratio optimization suggestions.
[0035] 3) In the technical solution provided by the present invention, based on the synergy between the programmable logic controller, the processor and the readable memory, timely and accurate optimization settings are made for the sintering fuel ratio, which is beneficial to improving the stability of the sintering process and reducing production energy consumption and carbon emissions. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] Figure 1 A flow chart of a method for optimizing sintering fuel ratio based on data mining;
[0037] Figure 2 A flowchart for mining fuel ratio optimization rules using fuzzy association rules;
[0038] Figure 3A visual interface for the sintering fuel ratio optimization setting system provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0039] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.
[0040] A method for optimizing sintering fuel ratio based on data mining, the workflow is as follows: Figure 1 As shown, including:
[0041] Step 1: Identify the current sintering production operation mode based on the three main operating parameters: material layer thickness, trolley speed and main exhaust fan frequency;
[0042] Step 2: Identify the permeability status of the material layer in the current sintering process based on the four status parameters of permeable mixed material moisture, No. 1 wind box negative pressure, large flue negative pressure and sintering end point position;
[0043] Step 3: Use three main temperature state parameters, namely, the flue gas temperature, the 22# wind box gas temperature, and the ring cooling I section gas temperature, to classify the heat level of the sintering process;
[0044] Step 4: Using the three groups of state values of operation mode, permeability state and fuel ratio optimization as rule antecedents, fuzzy logic reasoning is used to match the fuel ratio optimization control rules from the rule base, and the optimal fuel ratio control amount is output.
[0045] Step 1 specifically includes:
[0046] The Euclidean distance between the three-dimensional key operating parameters of the current material layer thickness, trolley speed and main exhaust fan frequency and the cluster centers corresponding to different types of operating modes is calculated to identify the current operating mode.
[0047] The cluster center of the operation mode is obtained from the historical production data analysis through the fuzzy clustering algorithm. First, the historical production data of material layer thickness, trolley speed and main exhaust fan frequency are collected; then, the characteristic variable data set is obtained after preprocessing the data; finally, the fuzzy clustering algorithm is used to divide the production operation mode into four categories: operation mode I, operation mode II, operation mode III, and operation mode IV.
[0048] Step 2 specifically includes: calculating the Euclidean distance between the four-dimensional characteristic states of the current mixture moisture, 1# wind box negative pressure, large flue negative pressure and sintering end point position and the cluster centers corresponding to different permeability states, and identifying the permeability state of the current sintering process.
[0049] The cluster center of the permeability state is obtained from the historical production data analysis through the fuzzy clustering algorithm. First, the historical production data of the mixture moisture, 1# wind box negative pressure, large flue negative pressure and sintering end position are collected; then, the characteristic variable data set is obtained after preprocessing; finally, the fuzzy clustering algorithm is used to divide the permeability state into: permeability state I, permeability state II, and permeability state III.
[0050] Step 3 specifically includes:
[0051] The four-dimensional temperature characteristics of the current flue exhaust temperature, 22# wind box exhaust temperature and ring cooling section I exhaust temperature are calculated, and the Euclidean distances between the cluster centers corresponding to different heat levels are respectively obtained to identify the heat level of the current sintering material layer.
[0052] The cluster center of the heat level is obtained from the historical production data analysis through the fuzzy clustering algorithm. First, the historical production data of the exhaust gas temperature of the large flue, the exhaust gas temperature of the 22# wind box and the exhaust gas temperature of the ring cooling section I are collected; then the characteristic variable data set is obtained after preprocessing; finally, the fuzzy clustering algorithm is used to divide the permeability state into: heat level I, heat level II, and heat level III.
[0053] Step 4 specifically includes:
[0054] i) Collect the operating parameters such as the thickness of the material layer during the sintering process, the parameters related to the air permeability such as the moisture content of the mixed material, and the heat and moisture related parameters such as the temperature of the flue gas, and identify the operating mode, air permeability and heat level of the historical production data according to the above steps;
[0055] ii) The trapezoidal fuzzy membership function is used to divide the fuel ratio in the historical production data into five discrete intervals: {NB, NS, ZO, PS, PB};
[0056] iii) The association rules of fuel ratio intervals under different operation modes, permeability states and heat levels in historical production data are mined through fuzzy association rule algorithm to form a fuel ratio optimization rule base.
[0057] Example 1
[0058] Please refer to Figure 1 The present invention provides a method for optimizing the sintering fuel ratio based on data mining, which includes the following steps:
[0059] (1) According to the method of step 1 above, the currently input material layer thickness, trolley speed and main exhaust fan frequency (the average of the dual main exhaust fan frequencies) are shown in Table 1.
[0060] Table 1 Key operating parameters for input
[0061]
[0062] According to the historical production data, the operation modes are divided into four categories. The corresponding cluster centers are shown in Table 2. The Euclidean distance between the current operation parameters and the cluster centers is calculated (the calculation formula is shown in Formula 1). It can be seen from Table 2 that the Euclidean distance between the current production operation mode and operation mode I is the shortest, and it is judged that the current operation mode belongs to operation mode I.
[0063] Formula 1:
[0064] Table 2 Cluster centers of four types of operation modes
[0065]
[0066] (2) According to the method in step 2 above, the four-dimensional characteristic state parameters of the current mixture moisture, No. 1 wind box negative pressure, large flue negative pressure and sintering end point position are collected, as shown in Table 3.
[0067] Table 3 Input characteristic state parameters
[0068]
[0069] The cluster center of the permeability state is also obtained through fuzzy clustering training of historical data, as shown in Table 4. The Euclidean distance between the four characteristic parameters of the current mixture moisture, 1# wind box negative pressure, large flue negative pressure and sintering end position and each cluster center is calculated. It can be seen that the Euclidean distance between the current characteristic state and the permeability state II is the shortest, and the current permeability state is judged to belong to Class II.
[0070] Table 4 Cluster centers of air permeability status
[0071]
[0072] (3) According to the method in step 3 above, the three-dimensional temperature characteristics of the current flue exhaust temperature, the 22# wind box exhaust temperature and the ring cooling section I exhaust temperature are collected, as shown in Table 5.
[0073] Table 5 Input temperature characteristic parameters
[0074]
[0075] The cluster centers of heat levels are trained from historical production data, as shown in Table 6. At the same time, the Euclidean distance between the current input temperature characteristic parameters and each cluster center is calculated, and it is identified that the heat level of the current production process belongs to Class II.
[0076] Table 6 Cluster centers of heat levels
[0077]
[0078]
[0079] (4) According to step 4, the three groups of state values, namely, operation mode, air permeability state, and fuel ratio optimization, are used as rule antecedents, and the fuel ratio optimization control rules are matched from the rule base using fuzzy logic reasoning. The matching rules are shown in Table 7. The optimal fuel ratio control amount is output.
[0080] Table 7 Fuel ratio optimization setting rules matched under current production conditions
[0081]
[0082] The end point of the rule output interval is used as the reference value to be output, the confidence of the rule is normalized as the proportion of the rule, and the final setting value of the fuel ratio is output after defuzzification. The specific calculation process is as follows:
[0083] Formula 2: out=l1*Result(Rulexxx) mid +l2*Result(Ruleyyy) mid ;
[0084] Formula 3:
[0085] Formula 4:
[0086] Result is the result of the acquisition rule, cost is the setting interval of the fuel ratio in the acquisition rule, and the mid operation is to take the median of the interval.
[0087] Based on the above operations, the fuel ratio should be set to:
[0088] out=0.52*3.2+0.48*3.1=3.152
[0089] The results of this example show that the sintering fuel ratio setting given by the present invention is relatively consistent with the actual adjustment value of the on-site operator when the sintering conditions are good. It can replace the operator to a certain extent and can adjust the sintering fuel ratio setting value more quickly, thereby achieving the purpose of stabilizing the sintering conditions. It has important application value and economic value for sintering production and is practical.
[0090] As a rule-based optimization setting system for the amount of water added to sintered mixture, it is composed of: a programmable logic controller, a sintering fuel bin gate controller, a memory, a processor, and a computer program stored in the memory and running on the processor. It is characterized in that the programmable logic controller collects production process data, and the processor implements the above control method when executing the computer program, and the fuel cut-out gate opening setting amount is finally output to the sintering fuel bin gate controller. The computer program in the memory is compiled in Java language, and the program visualization interface is as follows Figure 3 shown.
[0091] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which all belong to the protection of the present invention.
Claims
1. A method for optimizing sintering fuel ratio based on data mining, characterized in that: The following steps are involved: Step S1, obtaining the main operating parameters of the operating surface and identifying the current operating mode of sintering production; Step S2, obtaining the main state parameters of the process surface and identifying the permeability state of the material layer in the current sintering process; Step S3, obtaining the main thermal parameters of the equipment surface and classifying the thermal level of the sintering process; Step S4, using the operation mode, air permeability state and heat level as rule antecedents, using fuzzy logic reasoning to match the corresponding control rules from the fuel ratio optimization rule library, and outputting the optimal fuel ratio control amount; the main operating parameters of the operation surface include material layer thickness, trolley speed and main exhaust fan frequency; The main state parameters of the process surface include the moisture content of the mixture, the negative pressure of the No. 1 wind box, the negative pressure of the large flue and the sintering end point position; The main thermal parameters of the equipment surface include the exhaust gas temperature of the large flue, the exhaust gas temperature of the 22# wind box and the exhaust gas temperature of the ring cooling section I.
2. The method for optimizing sintering fuel ratio based on data mining according to claim 1, characterized in that: The identification process of the sintering production operation mode in step S1 is: calculating the Euclidean distance between the main operation parameters of the operation surface and the cluster centers corresponding to different operation modes, and identifying the operation mode of the current sintering production.
3. The method for optimizing sintering fuel ratio based on data mining according to claim 1, characterized in that: The process of identifying the permeability state of the material layer in step S2 is: calculating the Euclidean distance between the main state parameters of the process surface and the cluster centers corresponding to different permeability states, and identifying the permeability state of the current sintering process.
4. The method for optimizing sintering fuel ratio based on data mining according to claim 1, characterized in that: The heat level classification process in step S3 is: calculating the Euclidean distance between the main heat parameters of the equipment surface and the cluster centers corresponding to different heat levels, and identifying the heat level of the current sintering material layer.
5. A method for optimizing sintering fuel ratio based on data mining according to any one of claims 2 to 4, characterized in that: The calculation process of the Euclidean distance is: Formula 1: In formula 1, n is the number of cluster centers, which is numerically equal to the categories of each classification in steps S1 to S3, and x j =[x j1 ,x j2 ,x j3 ] is the normalized parameter vector to be identified, which is represented by the normalized value of the material layer thickness, trolley speed and main exhaust fan frequency in step S1, and so on, dimensionless; d ij is the Euclidean distance between the normalized parameter vector j and the i-th cluster center, dimensionless.
6. A method for optimizing sintering fuel ratio based on data mining according to any one of claims 2 to 4, characterized in that: The cluster center is obtained by analyzing the historical production data through a fuzzy clustering algorithm, and the specific steps are: obtaining the historical data of the corresponding parameters, obtaining the data set after preprocessing, and then dividing it through a fuzzy clustering algorithm.
7. The method for optimizing sintering fuel ratio based on data mining according to claim 1, characterized in that: The fuel ratio optimization rule base described in step S4 is constructed in the following steps: i) The fuel ratio in the historical production data is divided into five discrete intervals {NB, NS, ZO, PS, PB} using a trapezoidal fuzzy membership function; ii) The association rules of fuel ratio intervals under different operation modes, permeability states and heat levels in historical data are mined through fuzzy association rule algorithm to form a fuel ratio optimization rule base.
8. The method for optimizing sintering fuel ratio based on data mining according to claim 1, characterized in that: The screening process of the optimal fuel ratio control amount is as follows: the end point of the control rule output interval is used as the reference value to be output, the confidence of the rule is normalized as the proportion of the rule, and the final setting value of the fuel ratio is output after defuzzification. The specific calculation process is as follows: Formula 2: out = l1 * Result(Rulexxx) mid + l2 * Result(Ruleyyy) mid ; Formula 3: Formula 4: In formulas 2 to 4, Rulexxx and Ruleyyy are the two rules with the highest support whose antecedents in the rule base are the same as the calculation results in steps S1 to S3 of requirement 1; the rule support is Support Rulexxx Support Ruleyyy , dimensionless; l1 and l2 are the results of support normalization, dimensionless.
9. A computer readable memory comprising a computer program, characterized in that: The computer program can implement the sintering fuel ratio optimization method based on data mining as described in any one of claims 1 to 8.
10. A sintering fuel ratio optimization system based on data mining, characterized in that: include: A programmable logic controller, a processor and a readable memory as described in claim 9; the programmable logic controller collects production process data, and the processor executes the computer program on the readable memory.