Iron ore sintering ore blending optimization method and system based on improved differential evolution

Abstract

The invention discloses an iron ore sintering ore blending optimization method and system based on improved differential evolution, and the method comprises the steps: building a sintering ore blending optimization model of various raw materials in the sintering process by taking the minimum sintering cost and constraint and proportion penalty as optimization targets and combining with constraint conditions; and an improved differential evolution algorithm combining a simulated annealing algorithm and a local search algorithm is adopted to optimize the target, and a proportioning scheme for iron ore sintering ore blending optimization is obtained. According to the method, a sintering ore blending optimization model is constructed by taking minimum sintering cost and constraint and proportion penalty as optimization targets and combining constraint conditions; on the basis, an improved differential evolution algorithm is provided to optimize the target, and an excellent solution of the sintered ore proportion optimization problem in the iron ore sintering and blending process can be obtained within a short time.

Description

Technical Field

[0001] The present invention relates to an iron ore sintering and ore matching optimization method and system based on improved differential evolution, belonging to the technical field of intelligent optimization of iron ore sintering and ore matching optimization problems. Background Art

[0002] As a key raw material for industrial development, steel production is increasingly demanded. The steelmaking process begins with the sintering of various raw ores. Before blast furnace smelting, powdered iron ore is mixed with a certain amount of fuel and flux and sintered on sintering equipment to produce agglomerated ore before subsequent smelting. The quality of the sinter directly impacts blast furnace stability, operational smoothness, efficiency, and pig iron quality.

[0003] As demand for steel increases, high-grade iron ore resources are becoming increasingly scarce. Steel companies are using large quantities of low-grade ore fines, ore containing harmful elements, and various iron-containing industrial materials and waste. This results in sintering plants using a wide variety of complex raw materials, generating large amounts of waste slag and exhaust gas, which pose a serious environmental threat. Therefore, how to optimally combine multiple raw materials to produce high-quality products in the most economical way has become a critical issue in the steel industry.

[0004] Optimizing the raw material ratios for sintering is a complex continuous combinatorial optimization problem, classified as NP-hard. Currently, most steel companies rely on experience accumulated over long periods of production to determine the sintering ratio during the sintering process. However, due to the constraints of various sintering components and the varying prices of various raw material ores, manual sintering often makes it difficult to achieve an economical and high-quality sintering ratio. Therefore, an efficient solution to optimize the sintering ratio is urgently needed. Conventional heuristic algorithms cannot produce optimal solutions for this NP-hard problem, and mathematical programming methods struggle to achieve a good solution in a short period of time.

[0005] In view of this, the present invention is proposed. Summary of the Invention

[0006] The present invention provides an iron ore sintering and ore proportioning optimization method based on improved differential evolution. By taking minimizing sintering costs and constraints and proportioning penalties as optimization objectives and combining constraint conditions, a sintering and ore proportioning optimization model is constructed. On this basis, an improved differential evolution algorithm is proposed to optimize the objectives, which can obtain an excellent solution to the sintering ore proportioning optimization problem in the iron ore sintering and ore proportioning process in a relatively short time, thereby reducing the production cost of the sintering plant, improving its production efficiency, and enhancing the competitiveness of the plant. It can effectively solve the problems of factory cost waste and low economic benefits caused by improper proportioning of raw materials in the raw material sintering process.

[0007] The technical solution of the present invention is:

[0008] According to the first aspect of the present invention, a method for optimizing the sintering and ore distribution of iron ore based on improved differential evolution is provided. The method takes minimizing the sintering cost and constraints and proportion penalties as the optimization goal and combines the constraints to establish an optimization model for the sintering and ore distribution of various raw materials in the sintering process. The improved differential evolution algorithm combined with the simulated annealing algorithm and the local search algorithm is used to optimize the target to obtain an optimized proportion scheme for the sintering and ore distribution of iron ore.

[0009] Furthermore, the sintering ore blending optimization model of various raw materials in the sintering process includes an objective function and constraint conditions;

[0010] The objective function is as follows:

[0011]

[0012] in:

[0013]

[0014] The constraints are as follows:

[0015] x i,min ≤x i ≤x i,max ;

[0016] W j,min ≤W j ≤W j,max ;

[0017] Where Z is the fitness of the ratio scheme, cost represents the raw material cost per ton of the sintered ore, P is the ratio constraint value of the sintering raw material ratio, and α is the penalty factor for exceeding the chemical composition constraint; t j is the amount of the jth chemical component content in the sintered ore that exceeds the constraint range, j = 1, ..., m, m is the number of chemical components; θ is the penalty factor for the sintered ore exceeding the basicity constraint, g represents the amount of the sintered ore exceeding the basicity range after sintering the sintered raw materials; c i is the price per ton of the i-th sintering raw material, x i is the ratio of the i-th sintering raw material, i=1,...,n, n is the number of sintering raw materials; H2O i is the moisture content percentage of the i-th sintering raw material, Burn i is the burnout ratio of the i-th sintering raw material, Remaining is the residual ratio of sintered ore produced after sintering all sintering raw materials; β is the sintering raw material ratio and penalty factor; w i,j is the percentage of the jth chemical component in the i-th sintering raw material, W j is the percentage of the jth chemical component in the sintered ore; [Wj,min ,W j,max ] is the content constraint range of the jth chemical component in the sintered ore; W CaO 、 is the content of CaO and SiO2 in sintered ore; [PH min ,PH max ] is the basicity constraint range of sintered ore, [x i,min ,x i,max ] is the ratio range of the i-th raw material in the sintering raw materials.

[0018] Furthermore, the improved differential evolution algorithm combining the simulated annealing algorithm and the local search algorithm is used to optimize the target, specifically:

[0019] Step 1, encoding and decoding method: adopt the real number encoding and decoding strategy based on the ratio of various sintering raw materials, set the ratio of each raw material to [x i,min ,x i,max ] is a real number within the range. When coding, the raw material ratios are arranged in sequence to form a raw material ratio scheme π k,gen =[π k,gen (1),π k,gen (2),...,π k,gen (n)]; when decoding, the fitness value of the corresponding individual is calculated according to each raw material ratio in the individual; where π k,gen =[π k,gen (1),π k,gen (2),...,π k,gen (n)] is the matching scheme corresponding to individual k in the gen generation;

[0020] Step 2, initialization of the main population;

[0021] Step 3: Based on the main population, randomly select a preset number of individuals to perform mutation operations to generate mutant individuals and form a mutant population that is smaller than the main population;

[0022] Step 4. Generate a cross population: for each individual π in the main population k,gen Randomly select an individual from the mutant population Perform a crossover operation to generate a crossover population individual

[0023] Step 5: Update the main population individuals based on the cross population combined with the simulated annealing rule, and update each individual π in the main population. k,gen and the corresponding crossover population individuals For comparison, if The fitness is better than π k,gen The fitness of π k,gen Replace with On the contrary, use the simulated annealing rule to determine whether to accept The main population individuals;

[0024] Step 6. Perform local search on the main population individuals: For each individual π in the main population k,gen The ratio of each raw material in π k,gen (i) Perform local search to obtain the individual π' after local search k,gen ; Compare the fitness of individuals after local search, if π' k,gen The fitness is better than π k,gen The fitness of , then update π k,gen is π' k,gen ; Otherwise, do not update;

[0025] Step 7. Set the termination condition: If the termination condition is met, output "optimal ratio", otherwise repeat Step 3, Step 4, Step 5, and Step 6 until the termination condition is met.

[0026] Furthermore, the Step 3 includes:

[0027] Step 3.1. Randomly select an individual π from the main population k,gen , i=1, s=1;

[0028] Step 3.2, for each selected individual π k,gen The ratio of each raw material π k,gen (i) Implement mutation based on mutation value:

[0029] According to π k,gen (i) Obtain the first variation value and make a judgment on the first variation value: If the first variation value is between [x i,min ,x i,max ], then the ratio of the i-th raw material in the s-th mutant individual of the gen-th generation is Take the first variation value, i=i+1; otherwise, according to π k,gen (i) Obtain the second variation value and make a judgment on the second variation value: If the second variation value is between [x i,min ,x i,max ], then the ratio of the i-th raw material in the s-th mutant individual of the gen-th generation is Take the second variation value, i=i+1; if the first variation value and the second variation value are not [x i,min ,x i,max ] range, repeat this step until all raw material ratios are mutated;

[0030] Step 3.3, repeat Step 3.1 and Step 3.2 until the mutant population reaches the mutant population size; wherein, the mutant population size is smaller than the main population size.

[0031] Furthermore, the local search is specifically as follows:

[0032] According to π k,gen (i) Obtain the first variation value and make a judgment on the first variation value: If the first variation value is between [x i,min ,x i,max ], then the ratio of the i-th raw material in the k-th local search individual of the gen-th generation is π' k,gen (i) Take the first variation value; otherwise, according to π k,gen (i) Obtain the second variation value and make a judgment on the second variation value: If the second variation value is between [x i,min ,x i,max ], then the ratio of the i-th raw material in the k-th local search individual of the gen-th generation is π' k,gen (i) Take the second variation value; if the first variation value and the second variation value are not taken [x i,min ,x i,max ] range, repeat this step until the corresponding raw material ratio is completed in the local search, and obtain π' k,gen .

[0033] Furthermore, the expressions of the first variation value and the second variation value are as follows:

[0034] value1=π k,gen (i)+(x i,max -x i,min )*rand

[0035] value2=π k,gen (i)-(x i,max -x i,min )*rand

[0036] In the formula, value1 is the first mutation value, value2 is the second mutation value, and rand represents the random generation of a random number in the interval [0,1].

[0037] According to the second aspect of the present invention, an iron ore sintering and ore matching optimization system based on improved differential evolution is provided, including: a model establishment module, which is used to establish a sintering and ore matching optimization model for various raw materials in the sintering process with the optimization goal of minimizing sintering costs and constraints and ratio penalties and combining constraints; an acquisition module, which is used to optimize the target using an improved differential evolution algorithm combined with a simulated annealing algorithm and a local search algorithm, and obtain an optimized ratio scheme for iron ore sintering and ore matching.

[0038] According to a third aspect of the present invention, a processor is provided, which is used to execute an operation, wherein the operation includes executing any one of the above-mentioned iron ore sintering and ore blending optimization methods based on improved differential evolution.

[0039] The beneficial effects of the present invention are:

[0040] First, the model established by the present invention takes into account important indicators of sintered ore, such as the alkalinity constraint and chemical composition content constraint of sintered ore. The fitness value of sintered ore, including cost and penalty cost, is minimized according to the model of the present invention, which is conducive to minimizing cost while ensuring the quality of sintered ore and improving the product strength of the enterprise; second, the improved differential evolution method proposed in the present invention combines the simulated annealing rule with the local search. The simulated annealing rule can "accept poor solutions" in the early stage of the search, thereby avoiding local traps; at the same time, simulated annealing tends to search unknown areas of the solution space, while local search focuses on in-depth search of the local solution space. The reasonable combination of the two helps to enhance the global search capability and local search capability of the algorithm, greatly enhancing the algorithm search capability. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0042] To make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. It should be noted that, in the absence of conflict, the embodiments in this application and the features in the embodiments can be combined with each other in any way.

[0043] Example 1: Figure 1 As shown, according to the first aspect of an embodiment of the present invention, an iron ore sintering and ore matching optimization method based on improved differential evolution is provided, which takes minimizing sintering cost and constraints and ratio penalties as the optimization goal and combines the constraints to establish a sintering and ore matching optimization model for various raw materials in the sintering process; an improved differential evolution algorithm combining simulated annealing algorithm and local search algorithm is used to optimize the target to obtain an iron ore sintering and ore matching optimization scheme.

[0044] Furthermore, the sintering ore blending optimization model is as follows:

[0045]

[0046]

[0047] x i,min≤x i ≤x i,max (7)

[0048] W j,min ≤W j ≤W j,max (8)

[0049]

[0050] In the formula, the number of sintering raw materials is n, x i is the ratio of the i-th sintering raw material (%), i = 1, ..., n; H2O i is the water content percentage (%) of the i-th sintering raw material, Burn i is the burning loss ratio (%) of the i-th sintering raw material, and Remaining is the residual ratio (%) of the sintered ore produced after sintering all the sintering raw materials; w i,j is the percentage (%) of the jth chemical component in the i-th sintering raw material, W j is the percentage (%) of the jth chemical component in the sintered ore; c i is the price per ton of the i-th sintering raw material, cost represents the raw material cost per ton of the sintered ore produced, and the cost is calculated as the sum of the product of the unit price of each raw material and the dry material consumption of the raw material; β is the sintering raw material ratio and penalty factor, which is used to penalize the sintering raw material ratio exceeding the constraint; P is the ratio constraint value of the obtained sintering raw material ratio. When the obtained ratio sum is not equal to 1, the corresponding penalty value is applied to the obtained sintering raw material ratio. In the model of the present invention, there is no mandatory requirement that the sum of the raw material ratios of the sintered ore must be 1, but rather it is accumulated into the cost through the penalty value. This is a soft constraint processing method; its purpose is to not be overly constrained by the raw material ratio and restrictions during the cost optimization process, thereby reducing the optimization difficulty. At the same time, in the actual production process, a ratio exceeding 1 usually means problems such as over-injection of raw materials and excessive mixing, which will bring actual economic losses or process risks. By "reflecting" these consequences in the cost function, the optimization goal is more realistic. In addition, by adjusting the penalty coefficient, the "urgency" of the total ratio approaching 1 can be controlled, so that the optimization process can find a suitable balance between cost and feasibility. In the actual ore blending optimization process, a small deviation (such as 1±0.01) is allowed. The penalty mechanism allows the algorithm to have a "buffer" for searching, increasing the probability of finding a high-quality, feasible solution. m represents the number of chemical components, [W j,min ,W j,max ] is the content constraint range of the jth chemical component in the sintered ore, j = 1, ..., m, t j W is the amount by which the content of the jth chemical component in the sintered ore exceeds the constraint range; CaO 、 [PH] is the content of CaO and SiO2 in the sintered ore, and their ratio determines the basicity of the sintered ore; min ,PH max ] is the basicity constraint range of the sintered ore, g represents the amount of the sintered ore exceeding the basicity range after sintering the sintering raw materials; [x i,min ,x i,max ] is the ratio range of the i-th raw material in the sintering raw materials; Z is the fitness calculation formula of the ratio scheme, where: α is the penalty factor for exceeding the chemical composition constraint, and θ is the penalty factor for exceeding the alkalinity constraint of the sintered ore. The model of the present invention takes into account important sintered ore indicators such as the alkalinity constraint of the sintered ore, the chemical composition content constraint (such as the alkali metal content, and the sintered ore taste (i.e., the iron content in the sintered ore) constraint). Minimizing the sintered ore fitness value, including cost and penalty cost, according to the model of the present invention is conducive to minimizing costs while ensuring the quality of the sintered ore, thereby improving the product strength of the enterprise.

[0051] The specific steps of the sintering ore matching optimization method using the improved differential evolution algorithm that combines the simulated annealing algorithm and the local search algorithm are as follows:

[0052] Step 1, encoding and decoding method: adopt the real number encoding and decoding strategy based on various sintering raw material ratios, set the raw material ratio to [x i,min ,x i,max ] is a real number within the range. When coding, the raw material ratios are arranged in sequence to form a raw material ratio scheme π k,gen =[π k,gen (1),π k,gen (2),...,π k,gen (n)]; where π k,gen =[π k,gen (1),π k,gen (2),...,π k,gen (n)] is the ratio scheme corresponding to the kth generation individual; when decoding, the fitness value of the corresponding individual is calculated according to each raw material ratio in the individual;

[0053] For example, there are 10 individual π formed by the ratio of metallurgical raw materials. 1,1 =[1.10,1.43,14.63,17.82,5.63,8.39,12.67,12.54,11.87,4.88], which respectively correspond to the percentage values ​​(%) of rebar, hot-rolled coil, PB powder, Ba mixed powder, super special powder, Jinbuba powder, Caragas powder, Newman powder, FMG mixed powder and metallurgical coke raw materials in sintered ore.

[0054] Step 2. Initialization of the main population: Generate popSize initial individuals using random initialization rules, and randomly generate an initial matching scheme for each individual according to formula (10).

[0055] π k,1 (i) = x i,min +(x i,max -x i,min )*rand (10)

[0056] In the formula, rand means randomly generating a random number in the interval [0,1];

[0057] Step 3: Generate a small-scale mutant population: Based on the main population, randomly select some individuals for mutation operation to generate mutant individuals, and form a mutant population smaller than the main population. This process is completed according to the following steps:

[0058] Step 3.1. Randomly select an individual π from the main population k,gen , i=1, s=1;

[0059] Step 3.2, for each selected individual π k,gen The ratio of each raw material π k,gen (i) Implement mutation based on mutation value:

[0060] According to π k,gen (i) Obtain the first variation value and make a judgment on the first variation value: If the first variation value is between [x i,min ,x i,max ], then the ratio of the i-th raw material in the s-th mutant individual of the gen-th generation is Take the first variation value, i=i+1; otherwise, according to π k,gen (i) Obtain the second variation value and make a judgment on the second variation value: If the second variation value is between [x i,min ,x i,max ], then the ratio of the i-th raw material in the s-th mutant individual of the gen-th generation is Take the second variation value, i=i+1; if the first variation value and the second variation value are not [x i,min ,x i,max ] range, repeat this step until all raw material ratios are mutated;

[0061] The expressions of the first variation value and the second variation value are as follows:

[0062] value1=π k,gen (i)+(x i,max -x i,min )*rand (11)

[0063] value2=π k,gen (i)-(x i,max -x i,min)*rand (12)

[0064] Step 3.3, repeat Step 3.1 and Step 3.2 until the mutant population reaches the mutant population size; wherein, the mutant population size is smaller than the main population size.

[0065] Step 4. Generate a cross population: for each individual π in the main population k,gen Randomly select an individual from the mutant population Perform crossover operation according to formula (13) to generate a crossover population individual

[0066]

[0067] Where cr is the crossover probability.

[0068] Step 5: Update the main population individuals based on the cross population combined with the simulated annealing rule: for each individual π in the main population k,gen and the corresponding crossover population individuals For comparison, if The fitness is better than π k,gen The fitness of π k,gen Replace with On the contrary, use the simulated annealing rule to determine whether to accept is the main population individual; the process is expressed by formula (14), which is as follows:

[0069]

[0070] Where, Indicates the calculated value based on the sintering ore matching model The fitness of Z(π k,gen ) represents π k,gen The fitness of , Temperature is the temperature in the simulated annealing rule, and e is a natural constant;

[0071] When using the simulated annealing rule to accept Then, use formula (15) to update Temperature:

[0072] Temperature=Temperature×ε (15)

[0073] Here, ε is the annealing coefficient, 0 < ε < 1. This method updates the main population individuals using a crossover population combined with simulated annealing, preventing the main population from falling into local optima during the search process. Furthermore, as the program iterates, the temperature gradually decreases, gradually reducing the acceptance of relatively non-optimal individuals and gradually concentrating computing resources on in-depth searches for high-quality individuals.

[0074] Step 6. Perform local search on the main population individuals: For each individual π in the main population k,gen The ratio of each raw material in π k,gen (i) Perform local search to obtain the individual π' after local search k,gen ; Compare the fitness of individuals after local search. If Z(π' k,gen ) <Z(π k,gen ), then update π k,gen is π' k,gen ; Otherwise, do not update;

[0075] The local search is specifically:

[0076] According to π k,gen (i) Obtain the first variation value and make a judgment on the first variation value: If the first variation value is between [x i,min ,x i,max ], then the ratio of the i-th raw material in the k-th local search individual of the gen-th generation is π' k,gen (i) Take the first variation value; otherwise, according to π k,gen (i) Obtain the second variation value and make a judgment on the second variation value: If the second variation value is between [x i,min ,x i,max ], then the ratio of the i-th raw material in the k-th local search individual of the gen-th generation is π' k,gen (i) Take the second variation value; if the first variation value and the second variation value are not taken [x i,min ,x i,max ] range, repeat this step until the corresponding raw material ratio is completed in the local search, and obtain π' k,gen .

[0077] Step 7, termination condition: Set the termination condition to the maximum number of iterations. If the termination condition is met, the "optimal ratio" is output. In the present invention, the sum of the final optimal ratio may not be 1. At this time, it is necessary to divide the ratio of each raw material in the optimal ratio by the sum of the current ratios of each raw material (that is, the raw material ratios are scaled proportionally so that the sum of the optimal ratios is 1). Otherwise, repeat Step 3, Step 4, Step 5, and Step 6 until the termination condition is met.

[0078] The application of the above technical solutions shows that the combination of simulated annealing rules and local search in the present invention can improve the global search capability of the algorithm. The simulated annealing rule can "accept poor solutions" in the early stages of the search, thereby avoiding local traps. This is particularly advantageous for multi-modal problems (such as different ore combinations in ore optimization). At the same time, simulated annealing tends to search unknown areas of the solution space, while local search focuses on in-depth searches of the local solution space. The rational combination of the two helps to enhance the global and local search capabilities of the algorithm, greatly enhancing the algorithm's search capability.

[0079] Exemplarily, the sintering raw material ratio and penalty factor β=100, the chemical composition exceeds the constraint penalty factor α=100, and the sintered ore exceeds the alkalinity constraint penalty factor θ=10. The main population size is set to popSize=60, the variant population size is one-tenth of the main population size, the crossover population size is consistent with the main population, the crossover probability cr=0.6, and the annealing coefficient ε=0.99. To verify the performance of the algorithm, Table 1 shows the comparative solution results of the method of the present invention with the genetic algorithm (GA) and the particle swarm algorithm (PSO) in solving the ore matching problem in the iron ore sintering ore matching process under different examples. Table 2 shows the comparative verification of key operations of the method of the present invention with variant method 1 (removing the simulated annealing rule, and the rest remains unchanged) and variant method 2 (removing the local search, and the rest remains unchanged). Each algorithm was implemented in MATLAB 2022b, running on Windows 10, a 12th Gen Intel(R) Core(TM) i5-12400 2.50GHz processor, and 16GB of RAM. Each algorithm was independently run 20 times on each case. In Tables 1 and 2, "Best," "Mean," and "Worst" represent the best, average, and worst values ​​from these 20 runs, respectively. Cases 1-6 were constructed with varying raw material ratio constraints and prices. For ease of visualization, the optimal values ​​for each algorithm in each case are indicated in bold.

[0080] Table 1 Comparison of objective function values ​​between the proposed method and two algorithms under different calculation examples

[0081]

[0082] Table 2 Comparison of objective function values ​​between the proposed method and two variant methods under different calculation examples

[0083]

[0084] As can be seen from Tables 1 and 2, the improved differential evolution algorithm proposed in the present invention has achieved the optimal value in terms of the Best, Mean and Worst performance of all the examples. At the same time, the gap between the Best and Worst of the method of the present invention is not large, which shows that the method of the present invention has good stability during operation and can stably output high-quality solutions. Based on the above, the improved differential evolution algorithm proposed in the present invention that combines the simulated annealing algorithm and the local search algorithm has excellent search performance for the iron ore sintering and ore matching optimization problem. Among them, the differential evolution algorithm provides a strong global search capability, increases the diversity and convergence of the population, the simulated annealing rule helps the algorithm to jump out of the local optimal value, and the local search operation can greatly increase the algorithm's ability to mine local optimal solutions.

[0085] Furthermore, taking the procurement information of a steel company on May 12, 2023 as an example, the market transaction prices of raw materials for rebar, hot-rolled coil, PB powder, Ba mixed powder, super special powder, Jinbuba powder, Caragas powder, Newman powder, FMG mixed powder and metallurgical coke are 3680, 3840, 765, 789, 643, 731, 860, 783, 710 and 2290 yuan per ton respectively (that is, the unit price per ton of different raw materials in the mathematical model c i According to the market transaction price information and the measured chemical composition information of each raw material, the present invention is used to optimize the sintered ore ratio. According to the constraint range of the content of the jth chemical component in the sintered ore in the model [W j,min ,W j,max ], set the sinter ore taste constraint range to [53.2, 54.6], SiO2 constraint range to [5.5, 6], Al2O3 constraint range to [0, 2.5], MgO constraint range to [0, 2.6], P content constraint range to [0, 0.08], alkali metal constraint range to [0, 0.05], total K and Na content constraint range to [0, 0.18], and alkalinity constraint range to [1.75, 1.96] (i.e. [PH min ,PH max ] values). The optimized sintering ore ratio for the above raw materials is [0.71, 0.85, 18.02, 17.92, 5.80, 11.83, 16.10, 12.26, 11.99, 4.52]. This ore mix has a cost per ton of 948.71, a grade of 54.67%, a SiO2 content of 4.48%, a MgO content of 0.39%, an Al2O3 content of 1.98%, a basicity of 1.94, an alkali metal content of approximately 0.07%, a P content of approximately 0.07%, and a total K and Na content of approximately 0.07%. This is a relatively high-quality sintering ore mix.

[0086] Example 2: An iron ore sintering and ore matching optimization system based on improved differential evolution, comprising: a model building module for establishing a sintering and ore matching optimization model for various raw materials during the sintering process with the optimization goal of minimizing sintering costs and constraints and proportioning penalties, combined with constraints; an acquisition module for optimizing the target using an improved differential evolution algorithm that combines a simulated annealing algorithm and a local search algorithm to obtain an optimized proportioning scheme for iron ore sintering and ore matching. The term "module" as used above may be a combination of software and / or hardware that implements a predetermined function. For parts not described in detail in each module, please refer to the relevant descriptions of other embodiments.

[0087] Embodiment 3: A processor configured to execute an operation, the operation comprising executing any one of the above-described methods for optimizing iron ore sintering and ore proportioning based on improved differential evolution. When executing the operation, the processor implements the following steps: establishing an optimization model for sintering and ore proportioning of various raw materials during the sintering process, taking minimization of sintering cost and constraints and proportioning penalties as the optimization objective and in combination with the constraints; and optimizing the objective using an improved differential evolution algorithm that combines a simulated annealing algorithm with a local search algorithm to obtain an optimized proportioning scheme for iron ore sintering and ore proportioning.

[0088] The specific embodiments of the present invention are described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Various changes can be made within the knowledge of ordinary technicians in this field without departing from the scope of the present invention.

Claims

1. A method for optimizing iron ore sintering and ore blending based on improved differential evolution, characterized in that: With the minimization of sintering cost and constraints and proportion penalties as the optimization goal and combined with the constraints, an optimization model for sintering ore proportioning of various raw materials in the sintering process is established; the improved differential evolution algorithm combining simulated annealing algorithm and local search algorithm is used to optimize the goal and obtain the optimized proportion scheme for iron ore sintering.

2. The iron ore sintering and ore blending optimization method based on improved differential evolution according to claim 1, characterized in that: The sintering ore blending optimization model of various raw materials in the sintering process includes an objective function and constraint conditions; The objective function is as follows: in: The constraints are as follows: x i,min ≤x i ≤x i,max ; IN j,min ≤W j ≤W j,max ; Where Z is the fitness of the ratio scheme, cost represents the raw material cost per ton of the sintered ore, P is the ratio constraint value of the sintering raw material ratio, and α is the penalty factor for exceeding the chemical composition constraint; t j is the amount of the jth chemical component content in the sintered ore that exceeds the constraint range, j = 1, ..., m, m is the number of chemical components; θ is the penalty factor for the sintered ore exceeding the basicity constraint, g represents the amount of the sintered ore exceeding the basicity range after sintering the sintered raw materials; c i is the price per ton of the i-th sintering raw material, x i is the ratio of the i-th sintering raw material, i=1,...,n, n is the number of sintering raw materials; H2O i is the moisture content percentage of the i-th sintering raw material, Burn i is the burnout ratio of the i-th sintering raw material, Remaining is the residual ratio of sintered ore produced after sintering all sintering raw materials; β is the sintering raw material ratio and penalty factor; w i,j is the percentage of the jth chemical component in the i-th sintering raw material, W j is the percentage of the jth chemical component in the sintered ore; [W j,min ,W j,max ] is the content constraint range of the jth chemical component in the sintered ore; W CaO 、 is the content of CaO and SiO2 in sintered ore; [PH min ,PH max ] is the basicity constraint range of sintered ore, [x i,min ,x i,max ] is the ratio range of the i-th raw material in the sintering raw materials.

3. The iron ore sintering and ore blending optimization method based on improved differential evolution according to claim 1, characterized in that: The improved differential evolution algorithm combining the simulated annealing algorithm and the local search algorithm is used to optimize the target, specifically: Step 1, encoding and decoding method: adopt the real number encoding and decoding strategy based on various sintering raw material ratios, set the raw material ratio to [x i,min ,x i,max ] is a real number within the range. When coding, the raw material ratios are arranged in sequence to form a raw material ratio scheme π k,gen =[π k,gen (1),π k,gen (2),...,π k,gen (n)]; when decoding, the fitness value of the corresponding individual is calculated according to each raw material ratio in the individual; where π k,gen =[π k,gen (1),π k,gen (2),...,π k,gen (n)] is the matching scheme corresponding to individual k in the gen generation; Step 2, initialization of the main population; Step 3: Based on the main population, randomly select a preset number of individuals to perform mutation operations to generate mutant individuals and form a mutant population that is smaller than the main population; Step 4. Generate a cross population: for each individual π in the main population k,gen Randomly select an individual from the mutant population Perform a crossover operation to generate a crossover population individual Step 5: Update the main population individuals based on the cross population combined with the simulated annealing rule, and update each individual π in the main population. k,gen and the corresponding crossover population individuals For comparison, if The fitness is better than π k,gen The fitness of π k,gen Replace with On the contrary, use the simulated annealing rule to determine whether to accept The main population individuals; Step 6. Perform local search on the main population individuals: For each individual π in the main population k,gen The ratio of each raw material in π k,gen (i) Perform local search to obtain the individual π' after local search k,gen ; Compare the fitness of individuals after local search, if π' k,gen The fitness is better than π k,gen The fitness of , then update π k,gen is π' k,gen ; Otherwise, do not update; Step 7. Set the termination condition: If the termination condition is met, output "optimal ratio", otherwise repeat Step 3, Step 4, Step 5, and Step 6 until the termination condition is met.

4. The iron ore sintering and ore blending optimization method based on improved differential evolution according to claim 3 is characterized in that: Step 3 includes: Step 3.

1. Randomly select an individual π from the main population k,gen , i=1, s=1; Step 3.2, for each selected individual π k,gen The ratio of each raw material π k,gen (i) Implement mutation based on mutation value: According to π k,gen (i) Obtain the first variation value and make a judgment on the first variation value: If the first variation value is between [x i,min ,x i,max ], then the ratio of the i-th raw material in the s-th mutant individual of the gen-th generation is Take the first variation value, i=i+1; otherwise, according to π k,gen (i) Obtain the second variation value and make a judgment on the second variation value: If the second variation value is between [x i,min ,x i,max ], then the ratio of the i-th raw material in the s-th mutant individual of the gen-th generation is Take the second variation value, i=i+1; if the first variation value and the second variation value are not [x i,min ,x i,max ] range, repeat this step until all raw material ratios are mutated; Step 3.3, repeat Step 3.1 and Step 3.2 until the mutant population reaches the mutant population size; wherein, the mutant population size is smaller than the main population size.

5. The iron ore sintering and ore blending optimization method based on improved differential evolution according to claim 3, characterized in that: The local search is specifically: According to π k,gen (i) Obtain the first variation value and make a judgment on the first variation value: If the first variation value is between [x i,min ,x i,max ], then the ratio of the i-th raw material in the k-th local search individual of the gen-th generation is π' k,gen (i) Take the first variation value; otherwise, according to π k,gen (i) Obtain the second variation value and make a judgment on the second variation value: If the second variation value is between [x i,min ,x i,max ], then the ratio of the i-th raw material in the k-th local search individual of the gen-th generation is π' k,gen (i) Take the second variation value; if the first variation value and the second variation value are not taken [x i,min ,x i,max ] range, repeat this step until the corresponding raw material ratio is completed in the local search, and obtain π' k,gen .

6. The iron ore sintering and ore blending optimization method based on improved differential evolution according to claim 4 or 5, characterized in that: The expressions of the first variation value and the second variation value are as follows: value1=π k,gen (i)+(x i,max -x i,min )*rand value2=π k,gen (i)-(x i,max -x i,min )*rand In the formula, value1 is the first mutation value, value2 is the second mutation value, and rand represents the random generation of a random number in the interval [0,1].

7. An iron ore sintering and ore matching optimization system based on improved differential evolution, characterized in that: include: The model building module is used to establish a sintering ore blending optimization model for various raw materials in the sintering process, taking minimization of sintering cost and constraints and proportion penalties as the optimization goal and combining constraints; The acquisition module is used to optimize the target by using an improved differential evolution algorithm that combines a simulated annealing algorithm and a local search algorithm to obtain an optimized ratio scheme for iron ore sintering.

8. A processor, characterized in that: The processor is used to execute an operation, which includes executing the iron ore sintering and ore matching optimization method based on improved differential evolution as described in any one of claims 1 to 6.

Images