Coking coal blending method based on improved WMA algorithm

Abstract

The invention relates to a coking coal blending method based on an improved WMA algorithm, and the method comprises the steps: initializing a whale position through Tent chaotic mapping, carrying out the classification of individuals in a population through a hybrid evolutionary strategy, and carrying out the position updating of different individuals, thereby preventing the population from falling into local optimum, and enhancing the global search. According to the method, a boundary processing mechanism is introduced, and it is ensured that the updated whale position is in a search space. The improved algorithm aims at reducing the cost in the coking process of a steel mill and guaranteeing the coke quality, a new optimization tool is provided for coking coal blending, and high flexibility and adaptability are achieved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of coking coal blending in steel plants, and particularly relates to a coking coal blending method based on an improved WMA algorithm. BACKGROUND

[0002] The coking process contains complex physical and chemical changes, and various complex factors such as temperature and time have a significant impact on the quality of coke. In traditional coking production, the ratio of different coal types is mainly determined by manual experience, and it is difficult to consider complex changes and influencing factors. In addition to manual blending, coking plants also determine the ratio of coal types through small coke oven experiments, but the experimental time of the small coke oven is relatively long, which greatly increases the time cost. At present, many steel plants have poor benefits, and in order to reduce costs, they will choose low-priced coal types as much as possible, which makes the coal types and components change frequently in the coking production process. If the coking production only relies on manual coal blending and small coke oven experiment guidance, it is difficult to meet the production quality requirements in a short time, which increases the production cost. SUMMARY

[0003] (I) Objectives In order to overcome the above shortcomings, the purpose of the present application is to provide a coking coal blending method based on an improved WMA algorithm to solve the above technical problems.

[0004] (II) Technical solutions To achieve the above object, the technical solutions provided by the present application are as follows: A coking coal blending method based on an improved WMA algorithm, comprising the following steps: S1 Tent chaotic mapping-based whale position initialization; S2 classifying individuals in the whale population; S3 updating the position of different whale individuals respectively; S4 After updating the position of the whale, it is necessary to ensure that the updated whale position is still within the search space. If it exceeds the boundary range, it needs to be projected back to the boundary, and the formula of the updated coking coal blending method is: ; Wherein, L is the lower boundary of the search range, and U is the upper boundary of the search range.

[0005] S5 terminate the algorithm iteration when the algorithm reaches the maximum iteration number or meets other stopping conditions, and return the current optimal whale position W Best as the optimal solution.

[0006] Preferably, the Tent mapping iteration process in S1 is as follows: ; wherein z i is the current value of the m-th i th generation chaotic sequence; p is a control parameter, taken as p = 0.5.

[0007] Preferably, the S2 specifically comprises: sorting based on fitness value, selecting the top 10% of whale individuals in the population as leaders; based on the roulette mechanism, the remaining whale individuals are classified, and the selected whale individuals are selected as followers, and the remaining individuals are selected as stragglers, and then different evolution mechanisms are used for position updating of the above different types of whale individuals.

[0008] Preferably, the position updating specifically comprises: Leader position updating: compared with other types of whale individuals, the whale individuals in the leader position have better position information, so their position information is directly entered into the next iteration to fully play the guiding role of the leader; Follower position updating: based on the guiding role of the leader group, the position of the follower is updated, and the update formula is as follows: ; wherein W mean-l represents the mean value of the leader position, W r1 , W r2 , W r3 is a random individual in the follower, and r 1≠ r 2≠ r 3, F is a positive real number, F > 0, and F ∈ [0, 2], control the amplification multiple of the change; Straggler position updating: for the straggler individuals, we select part of the individuals to make their positions close to the population position, and the position updating formula is as follows: ; wherein W mean represents the average value of the entire whale population position, D represents the number of decision variables, so that the straggler can approach the migration pace of the population after position updating.

[0009] Beneficial effects: The application provides a coking blending method based on an improved WMA algorithm. The application has the following advantages: 1) The method improves the standard WMA algorithm from the aspects of whale population initialization, whale population individual classification and whale position updating, improves the population diversity of the algorithm, avoids falling into local optimum, and accelerates convergence.

[0010] 2) The method converts the traditional artificial coal blending and small coke oven experimental coal blending into artificial intelligence coal blending, determines the coal blending scheme that minimizes the cost based on the coal type, composition and constraint conditions of the raw coal, reduces the cost of artificial coal blending and small coke oven experimental coal blending, and meets the quality of each blending coal and coke.

[0011] 3) The method has guiding significance for the purchase of raw coal by the coking plant, can track the price and composition of raw coal in real time, and can obtain the lowest coal blending cost that meets the production demand in combination with market conditions, thereby reducing the cost of the steel plant from the procurement end. BRIEF DESCRIPTION OF DRAWINGS

[0012] Figure 1 is a flow structure schematic diagram of the application; Figure 2 is a comparison chart of the cost ratio of the application and the artificial coal blending scheme. DETAILED DESCRIPTION

[0013] To make the purpose, technical scheme and advantages of the application clearer and more explicit, the following will combine specific embodiments and refer to the accompanying drawings Figures 1-2 , and further explain the application. It should be understood that these descriptions are only exemplary and are not intended to limit the scope of the application. In addition, in the following description, the description of known structures and technologies is omitted to avoid unnecessary confusion of the concept of the application.

[0014] The application provides a coking coal blending method based on an improved WMA algorithm, which comprises the following steps: S1 whale position initialization based on Tent chaotic mapping; In the standard WMA algorithm, the initial position of the whale population is generated by a pseudo-random number, which is generated according to the following formula. Although the generation method of the pseudo-random number can make the whale population unevenly distributed in space, resulting in low efficiency and slow convergence in the optimization process, the optimization efficiency of the algorithm is too dependent on the initial population. The formula is: ; In the improved algorithm, we introduce the Tent chaotic map for initializing the whale population. Compared to pseudo-random numbers, the chaotic map has less repetition and higher randomness, increasing the diversity of the initial whale population and greatly improving the algorithm's global search performance. Furthermore, for the complex problem of coking coal blending with multiple local solutions, the improved WMA algorithm is more likely to find the global optimum, avoiding getting trapped in local optima. The Tent map is introduced because it has a faster search speed and better traversal uniformity; therefore, the Tent map is chosen instead of pseudo-random numbers. The iterative process of the Tent map in S1 is as follows: ; in z i It is the first i The current value of the chaotic sequence; p For control parameters, take p =0.5.

[0015] S2 determines the individual types in whale populations based on a hybrid evolutionary strategy; In the standard WMA algorithm, direct selection N L The individual with the best fitness is selected as the whale leader, among whom... N L Parameter settings are configured. A leader explores new areas, and other individuals gravitate towards the leader. This pseudo-dynamic evolutionary approach makes the population more prone to local searches, failing to balance exploration and development effectively. Therefore, to address the shortcomings of standard algorithms, this invention employs a hybrid evolutionary strategy for individual classification.

[0016] The specific process for determining the leader is as follows: the top 10% of whales are selected as leaders based on their fitness; the remaining whales are selected by roulette, and the selected individuals become followers of the population, while the remaining individuals become laggards of the population. Subsequently, the positions of different types of individuals are updated.

[0017] S3 updates based on the location of each whale; 5. The location update specifically includes: Leader position update: Compared to other types of whale individuals, whale individuals in the leader position have better position information, so their position information is retained and they directly enter the next iteration to give full play to the leader's guiding role. Follower position update: Based on the guiding role of the leader group, the position of followers is updated, and the update formula is as follows: ; in W mean-l This represents the average number of leader positions. Wr1 、 W r2 、 W r3 is a random individual in the followers, and r 1≠ r 2≠ r 3, F is a positive real number, F > 0, and F ∈ [0, 2], control the amplification of the change; Straggler position update: for the straggler individuals, we select a part of individuals from them to make their positions closer to the population position, and the position update formula is as follows: ; where W mean represents the average value of the whole whale population position, D represents the number of decision variables, so that the stragglers can approach the migration pace of the population after position update.

[0018] S4 boundary processing, after updating the position of the whale, it is necessary to ensure that the updated whale position is still within the search space, if it exceeds the boundary range, it needs to be projected back to the boundary, and the formula for updating the coking method is: ; where L is the lower boundary of the search range, and U is the upper boundary of the search range.

[0019] S5 iteration termination condition, when the algorithm reaches the maximum number of iterations or meets other stopping conditions, the algorithm iteration is terminated, and the current optimal whale position W Best is returned as the optimal solution.

[0020] Example 1 In the coking process, selecting the appropriate coal ratio is crucial to reduce costs and ensure coke quality. The traditional manual coal blending method relies on experience and is difficult to adapt to frequent changes in raw coal types and prices. This example aims to achieve lower cost and qualified coke production through improved WMA algorithm.

[0021] Experimental object: 15 times of coking coal blending process in actual production are selected.

[0022] Comparison scheme: the red line is the cost of manual coal blending scheme. The blue curve is the cost of improved WMA algorithm coal blending scheme.

[0023] Implementation steps (1) Data collection: collect the types, prices, ratios of coal used in 15 production processes, and the corresponding coke quality data.

[0024] (2) Algorithm initialization: use Tent chaotic mapping to initialize the whale position, i.e. the initial coal ratio. Calculate the cost and coke quality of each scheme according to the initial coal ratio, and evaluate the fitness. Set algorithm parameters, including population size, iteration number, magnification factor, etc.

[0025] (3) Determine individual type: select the top 10% of the schemes with the best fitness as leaders, the rest are selected as followers by roulette, and the rest are stragglers.

[0026] (4) Position update: Leader position update: do not update the position, and directly enter the next iteration.

[0027] Follower position update: update the distance according to the distance update formula.

[0028] Straggler position update: update the position by approaching the average position of the population.

[0029] (5) Boundary processing: ensure that the updated coal ratio is within the search boundary.

[0030] (6) Population update: first calculate the cost of all blending schemes, then update the population individuals based on greedy selection, and finally update the optimal blending scheme based on greedy selection.

[0031] (7) Iterative optimization: repeat steps 3-6 until the maximum iteration number is reached or other stopping conditions are met.

[0032] (8) Result evaluation: compare the cost of the improved WMA algorithm and the artificial coal blending scheme.

[0033] Calculate the cost savings for each production, and record the maximum and minimum cost savings.

[0034] (9) Result analysis The results of the present application show that Figure 2 As shown in the line chart, the horizontal axis is the coal blending number, the vertical axis is the scheme cost, the red line is the artificial coal blending scheme cost, and the blue curve is the improved WMA algorithm coal blending scheme cost. Select 15 times of production in actual production for comparison, it can be found that the WMA algorithm coal blending scheme cost is generally lower than the artificial coal blending scheme cost. The maximum cost saving is 8.19 yuan / ton, and the minimum cost saving is 0.732 yuan / ton.

[0035] It is to be noted that, in the present text, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0036] The above examples are merely used to illustrate the technical solutions of the present application, but not to limit it; although the present application has been described in detail with reference to the foregoing examples, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing examples, or make equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A coking coal blending method based on an improved WMA algorithm, characterized in that, Includes the following steps: S1 initializes the whale's position based on the Tent chaotic map; S2 classifies individuals within a whale population; S3 updates the location of each individual whale separately. After updating the whale's position, S4 needs to ensure that the updated whale position is still within the search space. If it exceeds the boundary, it needs to be reprojected back onto the boundary. The updated formula for the coking coal blending method is: ； Where L is the lower boundary of the search range and U is the upper boundary of the search range.

2. S5 terminates the algorithm iteration when it reaches the maximum number of iterations or meets other stopping conditions, and returns the current optimal whale position. W Best As the optimal solution.

3. The coking coal blending method based on the improved WMA algorithm according to claim 1, characterized in that, The process of Tent mapping iteration in S1 is as follows: ； in z i It is the first i The current value of the chaotic sequence; ρ For control parameters, take ρ =0.

5.

4. The coking coal blending method based on the improved WMA algorithm according to claim 1, characterized in that, S2 specifically includes: sorting based on fitness values, selecting the top 10% of whale individuals in the population as leaders; classifying the remaining whale individuals based on a roulette wheel mechanism, selecting the selected whale individuals as followers, and the remaining individuals as laggards, and then using different evolutionary mechanisms to update the positions of the above different types of whale individuals.

5. A coking coal blending method based on an improved WMA algorithm according to claim 3, characterized in that, The location update specifically includes: Leader position update: Compared to other types of whale individuals, whale individuals in the leader position have better position information, so their position information is retained and they directly enter the next iteration to give full play to the leader's guiding role. Follower position update: Based on the guiding role of the leader group, the position of followers is updated, and the update formula is as follows: ； in W mean-l This represents the average number of leader positions. W r1 , W r2 , W r3 It is a random individual among the followers, and r 1≠ r 2≠ r 3, F It is a positive real number. F >0, and F ∈[0,2], controls the amplification factor of the change; Lag member position update: For slack individuals, we select some individuals and move their positions closer to the population position. The position update formula is as follows: ； in W mean This represents the average location of the entire whale population. D This represents the number of decision variables, allowing laggards to move closer to the population's migration pace after their position is updated.

Images