Optimal blank weight calculation method and device based on particle swarm optimization
The optimal billet weight was calculated by using the particle swarm optimization algorithm, which solved the problems of low flexibility and insufficient standard length ratio in the heat treatment cross-cutting production of steel enterprises, and achieved higher production efficiency and resource utilization.
Patent Information
- Application Number
- CN202411621819.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-14
- Publication Date
- 2026-01-23
AI Technical Summary
In existing technologies, steel companies face problems such as low flexibility, resource waste, and complex inventory management in heat treatment cross-cutting production, especially under the make-to-order production model, where production efficiency is low and the standard length ratio is insufficient.
The optimal billet weight calculation method based on particle swarm optimization is adopted. By acquiring order information, classifying and grouping it, and combining production conditions and equipment constraints, the optimal billet weight is calculated using the particle swarm optimization algorithm to improve the straight-line rate.
It improved the accuracy of cross-cutting production, enhanced the steel companies' responsiveness to market demands, reduced resource waste, optimized inventory management, and improved production efficiency.
Smart Images
Figure CN121389691A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the technical field of heat treatment, and particularly relates to an optimal billet weight calculation method and device based on particle swarm optimization. BACKGROUND
[0002] The production of heat treatment cross-cutting is a process of heating a billet to a specified temperature in a heating furnace, rolling the billet to a specified thickness, width and length after the furnace, and then shearing the width and length in a shearing line to form the order delivery size. The production process of heat treatment cross-cutting mainly involves three types of products, namely, a slab (a billet heated in a heating furnace), a mother roll (a material formed one-to-one after the slab is rolled by a rolling mill) and a sub-board (a deliverable finished product after the mother roll is sheared).
[0003] Generally, the steelmaking and billet drawing production processes of a steel enterprise will determine the specifications of the slabs according to the maximum steelmaking capacity of the production line. This is because the maximum steelmaking capacity of the production line determines the maximum size and weight of the billet, which in turn affects the production capacity and efficiency of the downstream processes. Therefore, when planning production, the steel enterprise will determine the specifications of the slabs according to the technical parameters and capacity of the production line to ensure smooth production and maximize production efficiency. Moreover, in order to reduce the number of piles, the production of slabs is single, and after entering the slab warehouse, experienced operators will manually select orders with similar specifications for the slabs according to the order situation of the warehouse location. However, this method also has some potential drawbacks: relying on the maximum capacity of the production line to set the specifications of the slabs may limit the enterprise's response to market changes and reduce flexibility; when producing according to the maximum steelmaking capacity, some materials with inappropriate sizes may be produced and treated as waste, resulting in resource waste; relying on experienced operators to select orders with similar specifications may be influenced by subjective factors, leading to less objective or less accurate selection, and increasing the likelihood of errors.
[0004] Solving these problems usually requires enterprises to increase the flexibility of the production process while maintaining production efficiency and maximizing capacity. The cross-cutting accuracy rate refers to the proportion or percentage of the cross-cutting sub-plate size that meets the contract requirements. It is an evaluation index for the accuracy and quality of the steel enterprise in controlling the cross-cutting size during the cross-cutting production process. A higher cross-cutting accuracy rate means that the enterprise has better size control capability and can produce cross-cutting that better meets the design requirements, reducing production costs and resource waste. Therefore, steel enterprises will strive to improve the cross-cutting accuracy rate by improving production processes, strengthening quality management, optimizing material design, and other ways to achieve this goal. To improve the accuracy rate, the most effective solution in theory is to produce according to the order, that is, to calculate the size of the slab according to the order specifications and the maximum production capacity of the production line. However, this single-order production mode will actually lead to frequent changes in slab length or width on the slab continuous casting production line, which may cause a series of problems: frequent changes in slab width or length require downtime for adjustment, which will reduce production efficiency and increase production costs; with a wide variety of orders, steel mills need to store a large amount of materials of different specifications, increasing the complexity and challenge of inventory management, making inventory management difficult; due to the variety of orders, inconsistent specifications, low storage space utilization, and other reasons, in order to arrange slabs that meet the orders, operators may need to frequently move slabs, arrange and adjust, resulting in frequent pile rearrangement. SUMMARY
[0005] The main purpose of the embodiments of the present application is to provide a particle swarm optimization-based optimal slab weight calculation method and device, which can improve the accuracy rate of cross-cutting, improve production efficiency, and help steel enterprises better cope with the diversification of market demand, solving the problems of low efficiency of single-order production and low accuracy rate of maximum capacity steelmaking.
[0006] In a first aspect, a particle swarm optimization-based optimal slab weight calculation method is provided. The method includes: obtaining slab thickness and order information of heat treatment cross-cutting, the order information including: steel grade, sub-plate thickness, sub-plate width, sub-plate length, and total order weight; classifying and grouping the orders according to the steel grade and sub-plate width, and determining the slab width and slab thickness in each group of orders; determining the slab design rules and slab coil sub-plate conversion rules in each group of orders according to the production conditions; determining the objective function and constraint conditions of the optimal slab weight design in each group of orders according to the slab width, slab thickness, slab design rules, and slab coil sub-plate conversion rules; and obtaining the accuracy rate and optimal slab weight in each group of orders that satisfy the constraint conditions and maximize the objective function based on the particle swarm optimization algorithm.
[0007] In another possible implementation, order information of the slab thickness and the heat treatment crosscut is acquired, the order information including: a steel grade, a thickness, a sub-plate width, a sub-plate length and a total weight of an order, and further including: determining a product specification code and a metallurgical specification code according to the steel grade and the delivery specification of the order; determining process requirements and test requirements of each step according to the product specification code and the metallurgical specification code of the order, in combination with a pre-established metallurgical specification system.
[0008] In a possible implementation, the order is classified and grouped according to the steel grade and the sub-plate width, and the slab width and the slab thickness in each group of orders are determined, including: classifying the orders according to the steel grade, and orders of the same steel grade are classified into one category; for each category, the orders are arranged in ascending order according to the sub-plate width, and orders with a sub-plate width variation range not greater than a maximum width difference are classified into one group starting from the smallest sub-plate width; the slab width in each group of orders is determined according to the principle that the maximum width of the order ≤ the slab width ≤ the minimum width of the order + the maximum width difference; and the slab thickness in each group of orders is determined according to the continuous casting production equipment.
[0009] In a possible implementation, the slab design rules include: a minimum / maximum length limit of the slab, a maximum weight constraint, and a production capacity constraint of the equipment, and the slab design rules and the slab coil sub-plate conversion rules in each group of orders are determined according to the production conditions, including: the starting position and the minimum heating length of the production equipment determine the minimum length limit of the slab, and the maximum length limit of the slab is determined according to the length, the transportation capacity and the storage space of the equipment; the maximum weight constraint of the slab is determined according to the carrying capacity of the production equipment and the transportation tool; the weight range of the slab as the feeding slab of the hot rolling process and the weight range of the coil as the feeding coil of the crosscut process are determined according to the capacity of the production equipment; the weight conversion rules from the slab to the coil are determined according to the process feeding coefficient of the hot rolling process, and the weight conversion rules from the coil to the sub-plate are determined according to the process feeding coefficient of the crosscut process.
[0010] In another possible implementation, the target function and the constraint condition of the optimal slab weight design in each group of orders are determined according to the slab width, the slab thickness, the slab design rules and the slab coil sub-plate conversion rules, including: a plurality of constraint inequalities of the optimal slab weight design in each group of orders are determined according to the slab width, the slab thickness, the slab design rules and the slab coil sub-plate conversion rules; and the plurality of constraint inequalities are combined to form two constraint conditions of the slab weight:
[0011]
[0012] wherein, G slab is the slab weight, is the minimum value of the weight of the slab feeding the hot rolling process, λ1 is, is the maximum value of the weight of the slab feeding the hot rolling process, minimizing the weight of the coil fed to the slitting process, maximizing the weight of the coil fed to the slitting process, p is the material density of the slab and the coil, minimizing the length of the slab, maximizing the length of the slab, W slab W is the width of the slab and the coil, t is the thickness, maximizing the weight of the slab;
[0013] The objective function is determined according to the gauge formula of the slitting:
[0014]
[0015] wherein, is the total of the gauges of all orders fulfilled, is the total weight of the material at the entry of the slitting.
[0016] In another possible implementation, a plurality of constraint inequalities of the optimal slab weight design in each group of orders are determined according to the slab width, the slab thickness, the slab design rules and the slab coil sub-plate conversion rules, including: length constraint inequalities are obtained according to the minimum / maximum length limits of the slab in the slab design rules:
[0017]
[0018] slab weight constraint inequalities are obtained according to the maximum weight constraint in the slab design rules:
[0019]
[0020] weight constraint inequalities of the slab and the coil are obtained according to the equipment production capacity constraint in the slab design rules:
[0021]
[0022] wherein, L slab L slab is the length of the slab, G coil is the weight of the coil.
[0023] In another possible implementation, the positive yield and the optimal billet weight in each group of orders that satisfy the constraint condition and maximize the objective function are obtained based on a particle swarm optimization algorithm, including: initializing parameters of the particle swarm optimization algorithm; initializing randomly a speed and a position of each particle in a particle swarm that satisfies the constraint condition, taking the billet weight as the position and the positive yield as the fitness, and calculating an initial fitness value of each particle and a corresponding initial optimal position; iteratively updating the speed and the position of each particle; calculating a current fitness value of each particle, and if the current fitness value is better than an individual optimal fitness value, taking the current fitness value as a new individual optimal fitness value and updating the corresponding position; finding a global optimal fitness value in the particle swarm and updating the corresponding global optimal fitness value and position; and if a maximum iteration number or a same global optimal fitness value repetition number exceeds a maximum repetition number, terminating the algorithm and outputting the current global optimal fitness value of the particle swarm and the corresponding global optimal position.
[0024] In a second aspect, an optimal billet weight calculation device based on particle swarm optimization is provided, and the device includes: a data acquisition module configured to acquire slab thickness and heat treatment cross order information, the order information including: steel grade, sub-plate thickness, sub-plate width, sub-plate length, and total order weight; a parameter planning module configured to classify and group orders according to the steel grade and the sub-plate width, and determine the slab width and the slab thickness in each group of orders; a rule determination module configured to determine the slab design rule and the slab coil sub-plate conversion rule in each group of orders according to the production condition; a condition constraint module configured to determine the objective function and the constraint condition of the optimal billet weight design in each group of orders according to the slab width, the slab thickness, the slab design rule, and the slab coil sub-plate conversion rule; and an optimization calculation module configured to obtain the positive yield and the optimal billet weight in each group of orders that satisfy the constraint condition and maximize the objective function based on the particle swarm optimization algorithm.
[0025] In a third aspect, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the optimal billet weight calculation method based on particle swarm optimization provided in the first aspect when executing the program.
[0026] In a fourth aspect, a non-transitory computer readable storage medium is provided, and the computer readable storage medium stores a computer program, and the computer program is executable on a processor to implement the optimal billet weight calculation method based on particle swarm optimization provided in the first aspect. BRIEF DESCRIPTION OF DRAWINGS
[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the description of the embodiments of the present application will be briefly introduced as follows.
[0028] Figure 1A flow chart of a particle swarm optimization-based optimal billet weight calculation method provided for an embodiment of the present application is shown in FIG. 1.
[0029] Figure 2 A particle swarm algorithm-based optimal billet weight calculation method provided for an embodiment of the present application is shown in FIG. 2.
[0030] Figure 3 A diagram showing the change of the gauge factor with the weight of the slab for an embodiment of the present application is shown in FIG. 3.
[0031] Figure 4 A diagram showing the change of the optimal billet weight in the iteration process of the algorithm for an embodiment of the present application is shown in FIG. 4.
[0032] Figure 5 A diagram showing the change of the gauge factor in the iteration process of the algorithm for an embodiment of the present application is shown in FIG. 5.
[0033] Figure 6 A structure diagram of a particle swarm optimization-based optimal billet weight calculation device provided for an embodiment of the present application is shown in FIG. 6.
[0034] Figure 7 An entity structure diagram of an electronic device provided by the present application is shown in FIG. 7.
[0035] Detailed Implementation
[0036] The embodiments of the present application are described in detail below with reference to the accompanying drawings. The embodiments described below are examples for explaining the present application and should not be construed as limiting the present application.
[0037] Those skilled in the art can understand that the singular forms "a", "an" and "the" used herein include plural forms unless specifically stated otherwise. It should be further understood that the use of the term "include" in the specification of the present application means that the features, integers, steps, operations, modules and / or components described in the specification exist, but does not exclude the presence or addition of one or more other features, integers, steps, operations, modules, components and / or groups thereof. It should be understood that when we say a module is "connected" or "coupled" to another module, it can be directly connected or coupled to the other module, or there can be an intermediate module. In addition, "connected" or "coupled" used herein can include wireless connection or wireless coupling. The term "and / or" used herein includes all or any of the associated listed items and all combinations thereof.
[0038] To make the purpose, technical solutions and advantages of the present application clearer, the implementation of the present application will be described in further detail below with reference to the accompanying drawings.
[0039] The technical solutions of the present application and the solutions to the above technical problems will be described in detail below with specific examples. The following specific examples can be combined with each other, and the same or similar concepts or processes may not be described again in some examples. The embodiments of the present application will be described below with reference to the accompanying drawings.
[0040] The cross-cut positive size rate refers to the proportion or percentage of the cross-cut sub-plate size meeting the contract requirements, i.e. positive size / steel coil feeding weight x 100%, which is an evaluation index for the precision and quality of the cross-cut size control in the cross-cut production process of the steel enterprise. A higher cross-cut positive size rate means that the enterprise has better size control capability and can produce cross-cuts that meet the design requirements more, reducing production costs and resource waste. Therefore, the steel enterprise will strive to improve the cross-cut positive size rate, and achieve this goal through improving production process, strengthening quality management, optimizing material design, etc. However, the range of billet weight should also be considered, as too low billet weight may affect production efficiency. Therefore, in actual operation, the contradiction between the two goals needs to be considered to ensure that the positive size rate is improved and the production efficiency is maintained.
[0041] The optimal billet weight calculation method based on particle swarm optimization of the embodiments of the present application is applicable to multiple orders of the same material and similar width, which can be realized by cross-cutting the steel coils rolled from the same gauge plate blank. The design principle is to calculate the maximum plate blank weight that meets the production process equipment constraints. According to the production process, the thickness of the plate blank is fixed. The mother plate needs to be expanded and compressed after being processed by the hot rolling mill, and uneven deformation will cause irregularity at the edge of the mother plate, which needs to be cut to size, resulting in a slight decrease in width. In production, designers will give the cutting loss related coefficient, i.e. feeding coefficient, according to the above situation and production statistical law, and at the same time, the plate blank width will be determined according to the width range of the order.
[0042] For the optimal billet weight design problem of cross-cut production orders, the following assumptions are made for the modeling problem according to the sales orders and the needs of the production site:
[0043] 1) Each order can be realized by one or more plate blanks;
[0044] 2) There are different types of demands in the order. Before applying the optimal billet weight calculation method based on particle swarm optimization of the present application, the order needs to be classified according to the steel grade first, and then the production personnel need to group them according to the width and determine the width of the plate blank. Each group of orders is input respectively, and then the method is applied for solving. In the input order set, it is assumed that the orders have been grouped, and each order has the same requirements in terms of steel grade, thickness, etc. The change of the width of each order is controlled within the range that can be realized by the same specification of plate blank, and the rest will not be described again.
[0045] As Figure 1 shown is a flow chart of an optimal billet weight calculation method based on particle swarm optimization provided by an embodiment of the application, the method comprising:
[0046] Step S11, obtaining heat treatment cross order information, the order information including: steel grade, sub-plate thickness, sub-plate width, sub-plate length, and total order weight.
[0047] Specifically, the sales department issues orders from the sales system, and the production system receives and saves the order information, which mainly includes the steel grade of the order and the thickness t, width length and total order weight of the sub-plate, where i represents the i-th order, n represents the number of orders, and then 1≤i≤n,i∈N + .
[0048] Step S12, classifying and grouping the orders according to the steel grade and the sub-plate width, and determining the slab width and slab thickness in each group of orders.
[0049] The production personnel first classify the orders according to the steel grade, and orders of the same steel grade are classified into one category. Then, they are grouped according to the width.
[0050] Step S13, determining the slab design rules and slab coil to sub-plate conversion rules in each group of orders according to the production conditions.
[0051] Obtain the design rules determined by the on-site personnel from the designated source, including the slab design rules and the slab coil to sub-plate conversion rules. The slab design rules include: slab minimum and maximum length limits, maximum weight constraints, and equipment production capacity constraints.
[0052] Step S14, determining the objective function and constraint conditions of optimal billet weight design in each group of orders according to the slab width, the slab thickness, the slab design rules, and the slab coil to sub-plate conversion rules.
[0053] Determine the mathematical model of optimal billet weight design according to the slab width, the slab thickness, the slab design rules, and the slab coil to sub-plate conversion rules, including the objective function and the constraint conditions, to facilitate subsequent calculation of the optimal billet weight.
[0054] Step S15, obtaining the gauge ratio and optimal billet weight in each group of orders that satisfy the constraint conditions and maximize the objective function based on the particle swarm optimization algorithm.
[0055] Particle Swarm Optimization (PSO) algorithm is a heuristic algorithm invented by Dr. Eberhart and Dr. Kennedy, which is a random search algorithm based on group cooperation by simulating the foraging behavior of bird swarm, and is generally considered as a kind of swarm intelligence (SI). The idea comes from the mutual cooperation and information sharing among groups, which makes the group behavior evolve from disorder to order, and then achieves the purpose. The PSO algorithm is used to solve the above optimization problem because the most typical advantage of the particle swarm algorithm is fast convergence, especially for high-dimensional optimization problems, which converges to the optimal solution faster than genetic algorithm.
[0056] The embodiment of the present application carries out optimal billet weight calculation based on particle swarm optimization, which can improve the accuracy of cross-section, improve production efficiency, and help steel enterprises better cope with the diversification of market demand.
[0057] In the embodiment of the present application, in step S11, the slab thickness and the order information of the heat treatment cross-section are obtained, and the product specification code, the final use code and the metallurgical specification code are determined according to the steel grade and the delivery specification of the order; the process requirements and test requirements of each step are determined according to the product specification code and the metallurgical specification code of the order, combined with the pre-established metallurgical specification system. That is, when the sales department issues an order, the product specification code, the final use code and the metallurgical specification code are determined according to the steel grade, the delivery specification and other information of the order, and the production system can determine the process requirements and test requirements of each step according to the specification and the metallurgical specification code of the order from the pre-established metallurgical specification system maintenance.
[0058] In step S12, the order can be classified according to the steel grade, and orders of the same steel grade are classified into one category; for each category, the orders are arranged in ascending order according to the sub-plate width, and orders with sub-plate width variation range greater than the maximum width difference are grouped together from the smallest sub-plate width. Generally, in order to consider factors such as material loss, processing allowance and production efficiency in the production process, avoid operation difficulties or production interruption caused by insufficient width, and ensure that the final product size precision and quality meet the requirements, the slab width is generally not less than the width of the rolled steel coil. After the slab is processed by rolling, the slab width can be gradually reduced to the target steel coil width. However, if the width of the two exceeds a certain difference range, it may affect the quality of the final product. Therefore, the width difference between the slab and the steel coil must be kept within an acceptable range. That is, a batch of orders with similar width can be grouped together, and this group of orders can be realized by rolling the steel coil with the same specification slab.
[0059] The production personnel can design the width of the slab when grouping the orders, generally following the design principle of "order maximum width ≤ slab width ≤ order minimum width + maximum width difference". Thus the width of the slab in each group of orders can be determined according to the principle of "order maximum width ≤ slab width ≤ order minimum width + maximum width difference"; and the thickness of the slab in each group of orders can be determined according to the continuous casting production equipment.
[0060] For example, in the following input order set, it is assumed that the orders have been grouped, each order has the same requirements in terms of steel grade, thickness, etc., and the difference in width between the orders is kept within an acceptable range. The following input order set is a certain group of orders after grouping, and the number of orders is n. In this group of orders, the thickness of the sub-plate is t, the density is p, the width of the sub-plate of the i-th (1≤i≤n, i∈N + ) order is The length is The total weight of the order is and satisfies Δw max i.e. the maximum width difference is a fixed positive number.
[0061] In the embodiment of the present application, the slab design rules include: slab minimum / maximum length limit, maximum weight constraint, and equipment production capacity constraint. In step S13, the starting position of the production equipment and the minimum heating length determine the slab minimum length limit The maximum length limit of the slab is determined according to the length, transportation capacity and storage space of the equipment
[0062] The maximum weight constraint of the slab is determined according to the carrying capacity of the production equipment and the transportation tool. The maximum weight of the slab is Excessively heavy slabs can cause equipment damage or transportation difficulties.
[0063] The weight range of the slab as the input of the hot rolling process and the weight range of the coil as the input of the cross-cutting process are determined according to the capacity of the production equipment. The capacity of the production equipment determines the processing size range of the slab, including cutting, heating, rolling and other processes. The embodiment of the present application only considers the production capacity of the equipment for the hot rolling and cross-cutting processes, and takes the minimum and maximum values of the input weight of the two processes The input material of the hot rolling process is the slab, so the weight of the slab is The input material of the cross-cutting process is the coil, so the weight of the coil is
[0064] The slab to coil weight conversion rule and the coil to subplate weight conversion rule are mainly the weight conversion from slab to coil and the weight conversion from coil to subplate. The weight conversion rule from slab to coil is determined according to the process feed coefficient of the hot rolling process, and the weight conversion rule from coil to subplate is determined according to the process feed coefficient of the cross-cutting process. In the case of not considering external factors such as quality unqualified, it is usually converted by a coefficient, which is called feed coefficient. Let the process feed coefficient of hot rolling be λ1, and the process feed coefficient of cross-cutting be λ2, λ1 and λ2 are generally a real number not less than 1, which can be obtained according to historical experience. The total weight of the cross-cutting plate is set as G crosscut Then the slab to coil to subplate conversion rule is G slab = λ1G coil , G coil = λ2G crosscut
[0065] The design and execution of the above design rules and conversion rules usually require the production manufacturer to have a deep understanding of the production process, equipment capacity and material characteristics.
[0066] In step S14, optionally, a plurality of constraint inequalities of the optimal slab weight design in each group of orders are determined according to the slab width, the slab thickness, the slab design rule and the slab to coil to subplate conversion rule. Wherein, the length constraint inequality is obtained according to the minimum / maximum length limit of the slab in the slab design rule:
[0067]
[0068] The slab weight constraint inequality is obtained according to the maximum weight constraint in the slab design rule:
[0069]
[0070] The weight constraint inequality of slab and coil is obtained according to the equipment production capacity constraint in the slab design rule:
[0071]
[0072] Wherein, L slab is the slab length, and G coil is the weight of the cross-cutting inlet material.
[0073] The plurality of constraint inequalities are combined to form two constraint conditions of the slab weight:
[0074]
[0075] Wherein, G slab is the slab weight, is the minimum value of the weight of the hot rolling process feed slab, and λ1 is is the maximum value of the weight of the hot rolling process feed slab, This represents the minimum weight of the steel coil fed into the cross-cutting process. ρ represents the maximum weight of the steel coil fed into the cross-cutting process, where ρ is the material density of the slab and the steel coil. The minimum length of the slab. w is the maximum length of the slab. slab t represents the width of the slab and the steel coil, and t represents the thickness. This is the maximum weight of the slab;
[0076] The objective function is determined based on the formula for the positive scale of the cross section:
[0077]
[0078] in,
[0079]
[0080] in, This is the sum of the actual dimensions after all orders have been fulfilled. This represents the total weight of the cross-section entry material. The length of the sub-plate, The width of the sub-board. The thickness is the sub-plate.
[0081] Formula (8) represents the value range given the slab weight G within the constraints (5) and (6). slab Under the premise that the number of slabs required for each order is determined by the total weight of the order. The quantity q of slabs is determined by dividing the slab weight by the weight and rounding down. i There may be some contractual shortfall, but this shortfall is generally small compared to the contract weight and can be settled through other means during actual production, so it is not considered for now. Formula (9) represents the weight of a single sub-plate. This formula also applies to the weight calculation of steel coils and slabs, all of which have the same density. The thickness of the slab is determined by the production line, while the width and thickness of the steel coil are the same as those of the slab. Formula (10) represents the calculation of the weight of a steel coil with a weight of G, without considering other quality issues. slab The standard length of slab fulfillment order i. Formula (11) represents the sum of the standard lengths after all orders are fulfilled. Formula (12) represents the total weight of the steel coil.
[0082] For the optimization problem of billet weight, it is a little difficult to solve directly, and therefore an optimization algorithm needs to be used. The simplest and most direct method is to use the exhaustive method, that is, all possible solutions are compared, but this will cause waste of resources. Therefore, the embodiment of the application adopts a heuristic search algorithm. If a simulated annealing algorithm is used to solve, the calculation process is simple, and the universality, robustness and optimization effect are good, but there are disadvantages such as slow convergence speed and long execution time. In view of this shortcoming, the embodiment of the application adopts a PSO algorithm to solve the above optimization problem, because the most typical advantage of the particle swarm optimization algorithm is fast convergence speed, especially for high-dimensional optimization problems, it converges to the optimal solution faster than the genetic algorithm.
[0083] The PSO algorithm first initializes a group of particles in the feasible solution space, that is, a batch of initial billet weights are randomly determined within the value range determined by the constraint conditions (5) and (6). Each particle can be represented by three indicators of position, speed and fitness value to represent the particle characteristics. The speed represents the direction and distance of the next iteration movement of the particle, the position, that is, the billet weight, is a solution to the problem to be solved, and the fitness, that is, the gauge, represents the value of the optimization objective function.
[0084] Suppose there are N particles in the solution space, then:
[0085] (1) The position (billet weight) of the i th particle is G i , and G i satisfies the constraint conditions (5) and (6);
[0086] (2) The speed of the i th particle is V i ;
[0087] (3) The optimal position (billet weight) searched by the i th particle is G i,pbest , that is, the individual optimal solution;
[0088] (4) The optimal position (billet weight) searched by the group is G gbest , that is, the group optimal solution;
[0089] (5) The fitness value (gauge) of the optimal billet weight searched by the i th particle is f i , that is, the individual historical optimal fitness value;
[0090] (6) The fitness value of the optimal billet weight searched by the group is f g , that is, the group historical optimal fitness value;
[0091] (7) The speed update formula is
[0092] (8) The position (billet weight) update formula is
[0093] ω is an inertia weight, indicating the influence of the velocity of the last generation of particles on the velocity of the current generation of particles, c1 and c2 are an individual learning factor and a social learning factor respectively, the individual learning factor indicates the weight of the part of the next action of the particle derived from its own experience, and the social learning factor indicates the weight of the part of the next action of the particle derived from the experience of other particles, and r1 and r2 are random numbers in the range of [0, 1].
[0094] In the embodiment of the present application, in step S15, parameters of the particle swarm algorithm are initialized, the slab weight is taken as the position, the gauge accuracy is taken as the fitness, each particle in the particle swarm satisfying the constraint condition is randomly initialized in speed and position, the initial fitness of each particle and the corresponding initial optimal position are calculated, the speed and position of each particle are iteratively updated, the current fitness of each particle is calculated, if the current fitness is better than the individual optimal fitness, the current fitness is taken as the new individual optimal fitness, and the corresponding position is updated, the global optimal fitness is searched in the particle swarm, and the corresponding global optimal fitness and position are updated, if the maximum iteration number or the repetition number of the same global optimal fitness exceeds the maximum repetition number, the algorithm is terminated, and the current global optimal fitness of the particle swarm and the corresponding global optimal position are output.
[0095] The steps of solving the optimal slab weight problem based on the particle swarm algorithm are as follows:
[0096] (1) Initialize the particle swarm. A group of particles are randomly generated, and each particle is assigned an initial position (slab weight) and speed. Set the inertia factor, learning factor, maximum iteration number and maximum repetition number of the solution.
[0097] (2) Evaluate the initial fitness value of each particle. The initial fitness value (gauge accuracy) of each particle is taken as its local optimal value, and the corresponding position is taken as the position where the local optimal value is located.
[0098] (3) Find the best initial fitness value, and take it as the global optimal value, and record the position reaching the fitness value.
[0099] (4) Update the speed of each particle. According to the speed update formula and parameters such as inertia factor and learning factor, the speed of each particle is updated.
[0100] (5) Limit the amplitude of the speed. Ensure that the speed of the particle does not exceed the set maximum speed.
[0101] (6) Update the position of each particle. According to the displacement update formula and the updated speed, the position of each particle is updated. At the same time, ensure that the position of the particle does not exceed the set range.
[0102] (7) Compare the fitness value of each particle. If the current fitness value of a particle is better than its historical optimal value, it is taken as the new individual historical optimal fitness value and the corresponding position is updated.
[0103] (8) Compare the global optimal value. Find the best global fitness value in the particle swarm as the swarm optimal fitness value and update the corresponding position.
[0104] (9) Check the termination condition. If the maximum number of iterations or the number of repetitions of the same swarm optimal fitness value exceeds the maximum number of repetitions of the solution, go to step (10) and the algorithm terminates. If the termination condition is not met, go to step (4).
[0105] (10) Output the result. After the algorithm ends, output the global optimal value of the optimal particle and the corresponding position
[0106] More specifically, referring to Figure 2 , the method for solving the optimal blank weight by the particle swarm algorithm comprises:
[0107] Step 200: Start.
[0108] Step 201: Initialize the particle swarm parameters.
[0109] The particle swarm parameters include the number of iterations, the discount factor, etc.
[0110] Step 202: Randomly initialize the position and velocity of each particle.
[0111] Step 203: Determine whether the end condition is met. If yes, go to step 208, otherwise, go to step 204.
[0112] The end condition is that the maximum number of iterations is reached or the number of repetitions of the same swarm optimal fitness value exceeds the maximum number of repetitions of the solution.
[0113] Step 204: Update the velocity and position of each particle.
[0114] Step 205: Calculate the fitness value of each particle.
[0115] Step 206: Calculate the individual historical optimal fitness value and position of each particle.
[0116] Step 207: Update other parameters such as the discount factor, the number of iterations, etc. Then return to step 203.
[0117] Step 208: Output the optimal solution. Then go to step 209.
[0118] The output optimal solution includes a global optimal fitness value of the optimal particle and a corresponding global optimal position, and the global optimal fitness value is an optimal yield rate, and the corresponding global optimal position is an optimal billet weight corresponding to the optimal yield rate.
[0119] Step 209: end.
[0120] Subsequent production according to the obtained optimal billet weight can ensure that the yield rate is improved and the production efficiency is maintained.
[0121] The following takes the heat treatment cross production of an enterprise as an experimental object to verify the results of the optimal billet weight calculation method based on particle swarm optimization of the embodiment of the present application. Taking four orders in Table 1 as an example.
[0122] Table 1 Order information table
[0123] Number Length Width Thickness Cutting plate weight Contract weight 0 10100 mm 1600 mm 8 mm 1.015t 100t 1 9000 mm 1610 mm 8 mm 0.910t 120t 2 12000 mm 1620 mm 8 mm 1.221t 140t 3 14000 mm 1630 mm 8 mm 1.433t 110t
[0124] An optimal billet weight is obtained, so that the yield rate of the above orders is maximized after the billet is redeemed by the weight. The width of the billet is 1650mm, the thickness is 240mm, the minimum length is 4000mm, the maximum length is 10980mm, the feeding coefficients of the billet to the steel coil and the steel coil to the cut plate are 1.013 and 1.055 respectively, and it is stipulated that the designed steel coil weight cannot be less than 25t. The algorithm uses Python programming and runs on a computer with Windows 10 operating system.
[0125] The calculated value range of the billet is [25.325t, 34.132t]. In order to better compare the performance of the algorithm, the optimal billet weight is first calculated, which makes the yield rate of the above orders redeemed maximum. First, the exhaustive method is used, and it is observed that the yield rate changes with the solution as shown in Figure 3 The maximum yield rate is 94.231% when the billet weight is 26.258t.
[0126] Observing Figure 3 It can be found that in the feasible solution region, there are still some local optimal solutions, and these local optimal solutions and the corresponding yield rates are shown in Table 2, and the data is arranged in descending order of yield rate. It can be observed that when the billet weight is 34.038t, the obtained yield rate is close to the maximum yield rate. Although the yield rate obtained by this solution is not the maximum, the billet weight is close to the maximum value, and the production efficiency can be greatly improved according to the billet weight for steelmaking.
[0127] Table 2 Local optimal solution and corresponding yield rate
[0128]
[0129]
[0130] By changing the width and thickness of the slab, the optimal slab weight was determined using the same exhaustive method, and the results are shown in Table 3.
[0131] Table 3 Optimal billet weight and optimal length ratio under different slab width and thickness settings.
[0132]
[0133] Observing the above experiments, it can be seen that changes in width and thickness did not affect the results of the other experiments. This is because changes in slab width and thickness only affect the range of slab weight values. If the optimal slab weight is not generated in the neighborhood of the maximum and minimum solutions, then the limitations on slab thickness and width have almost no impact on the results.
[0134] In this embodiment of the invention, the initialization of the particle swarm is randomly selected within the value range. For ease of comparison, the results of the simulated annealing algorithm are also shown, and the relevant parameters are set as shown in Table 4.
[0135] Table 4 Parameter Settings
[0136]
[0137]
[0138] First, observe the convergence of the algorithm, and then observe the current optimal billet weight and optimal length ratio obtained after each iteration. The results are as follows: Figure 4 to Figure 5 As shown, where Figure 4 This represents the change in the optimal billet weight during the algorithm iteration process. Figure 5 The value represents the change in the optimal positive scale during the algorithm iteration process. It can be seen that the optimal solutions obtained by both algorithms during the iteration process are close to the actual optimal solutions. In this experiment, after approximately 20 iterations, the optimal solution sequence of the PSO algorithm converged to 26.258, which is the same as the result shown in Table 3. The simulated annealing algorithm found its optimal solution after approximately 170 iterations. Therefore, the convergence speed of the PSO algorithm is superior to that of the simulated annealing algorithm.
[0139] When the slab thicknesses are 230mm and 240mm respectively, and the slab widths are 1630mm, 1640mm and 1650mm respectively, the running algorithm obtains the results shown in Table 5, and the corresponding real global optimal slab weight and optimal gauge are shown in Table 3. In order to facilitate comparison, the running results of the simulated annealing (SA) algorithm are also shown in Table 5. Through observation and comparison of Table 3 and Table 5, it is found that, except for Experiment 2, the optimal solution obtained by the PSO algorithm of the present application has smaller error with the real value, faster convergence speed, and very stable results. Although the gauge of Experiment 2 is slightly lower than that of SA, the slab weight is close to the maximum value, which is more conducive to improving production efficiency. That is, for the problem described in the present application, if the PSO algorithm is used for solving, not only better optimization can be achieved than the simulated annealing algorithm (SA), but also obvious advantages in the speed of optimization.
[0140] Table 5 optimal solution under different slab width and thickness settings
[0141]
[0142]
[0143] The embodiment of the present application comprehensively considers the problems of low efficiency of single production and low gauge of maximum capacity steelmaking, and proposes a batch order production mode. Taking hot treatment cross-cut plate as the research object, an effective algorithm for coexistence of hot slab weight maximization and gauge and weight design is studied. For a batch of orders with the same material, the same thickness and the width varying within a certain range, a slab weight is designed based on the particle swarm optimization algorithm, so that the overall gauge is maximized after the batch order is realized by the designed slab, the gauge of the cross-cut is improved, the quality design calculation model of the steel is optimized, and the diversified market demand of the steel enterprise is better met. Experiments show that the particle swarm optimization algorithm has better optimization and convergence for solving the optimal slab weight problem proposed in the present application.
[0144] In summary, the embodiment of the present application obtains the slab thickness and the order information of the hot treatment cross-cut, the order information including: steel grade, sub-plate thickness, sub-plate width, sub-plate length and total order weight; classifies and groups the orders according to the steel grade and sub-plate width, and determines the slab width and slab thickness in each group of orders; determines the slab design rules and slab coil sub-plate conversion rules in each group of orders according to the production conditions; determines the objective function and constraint condition of the optimal slab weight design in each group of orders according to the slab width, slab thickness, slab design rules and slab coil sub-plate conversion rules; and obtains the gauge and optimal slab weight in each group of orders that meet the constraint condition and maximize the objective function based on the particle swarm optimization algorithm, which can improve the cross-cut gauge, improve the production efficiency, and promote the diversified market demand of the steel enterprise.
[0145] As Figure 6 Fig. 1 shows a structural diagram of a device for calculating optimal slab weight based on particle swarm optimization according to an embodiment of the present application, which comprises:
[0146] A data acquisition module 601 is configured to acquire slab thickness and order information of heat treatment cross-section, and the order information comprises: steel grade, sub-plate thickness, sub-plate width, sub-plate length, and total weight of order;
[0147] A parameter planning module 602 is configured to classify and group orders according to the steel grade and the sub-plate width, and determine slab width and slab thickness in each group of orders;
[0148] A rule determination module 603 is configured to determine slab design rules and slab coil sub-plate conversion rules in each group of orders according to production conditions;
[0149] A condition constraint module 604 is configured to determine target function and constraint conditions of optimal slab weight design in each group of orders according to the slab width, the slab thickness, the slab design rules and the slab coil sub-plate conversion rules;
[0150] An optimization calculation module 605 is configured to acquire gauge ratio and optimal slab weight in each group of orders that satisfy the constraint conditions and maximize the target function based on a particle swarm optimization algorithm.
[0151] The device of the above embodiment is applied to the corresponding method in the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be described here.
[0152] Figure 7 Fig. 1 shows a structural diagram of a device for calculating optimal slab weight based on particle swarm optimization according to an embodiment of the present application, which comprises: Figure 7As shown, the electronic device can include a processor 701, a communications interface 702, a memory 703, and a communications bus 704, wherein the processor, the communications interface, and the memory complete mutual communication through the communications bus. The processor can invoke a logic instruction in the memory to execute an optimal billet weight calculation method based on a particle swarm optimization, which includes: obtaining slab thickness and heat treatment cross-order information, the order information including: steel grade, sub-plate thickness, sub-plate width, sub-plate length, and total order weight; classifying and grouping the order according to the steel grade and the sub-plate width, and determining the slab width and the slab thickness in each group of orders; determining the slab design rule and the slab coil sub-plate conversion rule in each group of orders according to the production conditions; determining the objective function and the constraint condition of the optimal billet weight design in each group of orders according to the slab width, the slab thickness, the slab design rule, and the slab coil sub-plate conversion rule; and obtaining the gauge and the optimal billet weight in each group of orders that satisfy the constraint condition and maximize the objective function based on a particle swarm optimization algorithm.
[0153] In addition, the logic instruction in the memory described above can be implemented in the form of a software functional unit and sold or used as an independent product, which can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various program code storage media.
[0154] In another aspect, the embodiments of the present application also provide a computer program product, which comprises a computer program stored on a non-transitory computer-readable storage medium, and the computer program comprises program instructions, and when the program instructions are executed by a computer, the computer is capable of executing the optimal slab weight calculation method based on particle swarm optimization provided by the above-mentioned method embodiments. The method comprises the following steps: obtaining order information of slab thickness and heat treatment cross-cutting, and the order information comprises the following: steel grade, sub-plate thickness, sub-plate width, sub-plate length, and total order weight; classifying and grouping the orders according to the steel grade and the sub-plate width, and determining the slab width and the slab thickness in each group of orders; determining the slab design rules and the slab coil sub-plate conversion rules in each group of orders according to the production conditions; determining the objective function and the constraint condition of the optimal slab weight design in each group of orders according to the slab width, the slab thickness, the slab design rules and the slab coil sub-plate conversion rules; and obtaining the gauge ratio and the optimal slab weight in each group of orders that satisfy the constraint condition and make the objective function maximum based on the particle swarm optimization algorithm.
[0155] In another aspect, the embodiments of the present application also provide a non-transitory computer-readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the optimal slab weight calculation method based on particle swarm optimization provided by the above-mentioned embodiments. The method comprises the following steps: obtaining order information of slab thickness and heat treatment cross-cutting, and the order information comprises the following: steel grade, sub-plate thickness, sub-plate width, sub-plate length, and total order weight; classifying and grouping the orders according to the steel grade and the sub-plate width, and determining the slab width and the slab thickness in each group of orders; determining the slab design rules and the slab coil sub-plate conversion rules in each group of orders according to the production conditions; determining the objective function and the constraint condition of the optimal slab weight design in each group of orders according to the slab width, the slab thickness, the slab design rules and the slab coil sub-plate conversion rules; and obtaining the gauge ratio and the optimal slab weight in each group of orders that satisfy the constraint condition and make the objective function maximum based on the particle swarm optimization algorithm.
[0156] It should be understood that, although each step in the flowchart of the accompanying drawings is displayed in sequence according to the direction of the arrow, these steps are not necessarily executed in sequence according to the direction of the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and they can be executed in other sequences. Moreover, at least part of the steps in the flowchart of the accompanying drawings can comprise multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution sequence is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or sub-steps or stages of other steps.
[0157] The above merely describes some implementation manners of the present application, and it should be noted that, for those skilled in the art, some improvements and refinements can be made without departing from the principles of the present application, and these improvements and refinements should also be considered as the protection scope of the present application.
Claims
1. A method for calculating an optimal billet weight based on particle swarm optimization, characterized by, The method comprises: Obtaining slab thickness and heat treatment cross-section order information, the order information comprising: steel grade, sub-plate thickness, sub-plate width, sub-plate length, and total order weight; Classifying and grouping orders according to the steel grade and the sub-plate width, and determining slab width and slab thickness in each group of orders; Determining slab design rules and slab coil sub-plate conversion rules in each group of orders according to production conditions; Determining the target function and constraint conditions of the optimal slab weight design in each group of orders according to the slab width, the slab thickness, the slab design rules and the slab coil sub-plate conversion rules; Based on the particle swarm optimization algorithm, the gauge and the optimal slab weight in each group of orders that satisfy the constraint conditions and maximize the target function are obtained.
2. The method of claim 1, wherein, The order information of the slab thickness and the heat treatment cross-section also includes: Determining the product specification code and the metallurgical specification code according to the steel grade and the delivery specification of the order; According to the product specification code and the metallurgical specification code of the order, and combining the pre-established metallurgical specification system, the process requirements and test requirements of each step are determined.
3. The method of claim 1, wherein, The classification and grouping of orders according to the steel grade and the sub-plate width, and the determination of slab width and slab thickness in each group of orders, comprise: Classify orders according to steel grade, orders of the same steel grade are classified into one category; For each category, arrange the orders in ascending order according to the sub-plate width, starting with the smallest sub-plate width, and group orders with sub-plate width variation range not greater than the maximum width difference; Determine the slab width in each group of orders according to the principle that the order maximum width ≤ slab width ≤ order minimum width + maximum width difference; Determine the slab thickness in each group of orders according to continuous casting production equipment.
4. The method of claim 1, wherein, The slab design rules include: slab minimum / maximum length limit, maximum weight constraint, and equipment production capacity constraint. The determination of slab design rules and slab coil sub-plate conversion rules in each group of orders according to production conditions comprises: The starting position and minimum heating length of the production equipment determine the minimum length limit of the slab, and the length, transportation capacity and storage space of the equipment determine the maximum length limit of the slab; Determine the maximum weight constraint of the slab according to the carrying capacity of the production equipment and the transportation tool; Determine the weight range of the slab as the input of the hot rolling process and the weight range of the coil as the input of the cross-section process according to the capacity of the production equipment; Determine the weight conversion rule from slab to coil according to the process feeding coefficient of the hot rolling process, and determine the weight conversion rule from coil to sub-plate according to the process feeding coefficient of the cross-section process.
5. The method of claim 1, wherein, The determination of the target function and constraint conditions of the optimal slab weight design in each group of orders according to the slab width, the slab thickness, the slab design rules and the slab coil sub-plate conversion rules comprises: Determine multiple constraint inequalities of the optimal slab weight design in each group of orders according to the slab width, the slab thickness, the slab design rules and the slab coil sub-plate conversion rules; Combine multiple constraint inequalities to form two constraint conditions of the slab weight: wherein G slab is the slab weight, is the minimum value of the weight of the slab fed to the hot rolling process, λ1 is, is the maximum value of the weight of the slab fed to the hot rolling process, is the minimum value of the weight of the coil fed to the cross-cutting process, λ2 is, is the maximum value of the weight of the coil fed to the cross-cutting process, ρ is the material density of the slab and the coil, is the minimum length of the slab, is the maximum length of the slab, W slab is the width of the slab and the coil, t is the thickness, is the maximum weight of the slab; Determine the target function according to the gauge formula of the cross-section: wherein, is the sum of the positive yardage for all orders filled, is the total weight of the crosscut entry material.
6. The method of claim 5, wherein, The multiple constraint inequalities for determining the optimal billet weight design in each group of orders according to the slab width, the slab thickness, the slab design rule and the slab steel coil sub-plate conversion rule, comprise: The length constraint inequality is obtained according to the minimum / maximum length limit of the slab in the slab design rule: The slab weight constraint inequality is obtained according to the maximum weight constraint in the slab design rule: The weight constraint inequality of the slab and the steel coil is obtained according to the equipment production capacity constraint in the slab design rule: where L slab is the slab length, G coil is the weight of the coil.
7. The method of claim 1, wherein, The maximum size ratio and the optimal billet weight in each group of orders that satisfy the constraint conditions and maximize the target function are obtained based on the particle swarm optimization algorithm, comprising: Initializing the parameters of the particle swarm algorithm; Randomly initializing the speed and position of each particle in the particle swarm that satisfies the constraint conditions with the slab weight as the position and the maximum size ratio as the fitness, and calculating the initial fitness value of each particle and the corresponding initial optimal position; Iteratively updating the speed and position of each particle; The current fitness value of each particle is calculated, and if the current fitness value is better than the individual optimal fitness value, the current fitness value is taken as the new individual optimal fitness value, and the corresponding position is updated; The global optimal fitness value is found in the particle swarm, and the corresponding global optimal fitness value and position are updated; If the maximum iteration number or the same global optimal fitness value repetition number exceeds the maximum repetition number, the algorithm is terminated, and the current global optimal fitness value and the corresponding global optimal position of the particle swarm are output.
8. An optimal billet weight calculation device based on particle swarm optimization, characterized in that, The device comprises: A data acquisition module for acquiring slab thickness and heat treatment cross-sectional order information, the order information comprising: steel grade, sub-plate thickness, sub-plate width, sub-plate length, and total order weight; A parameter planning module for classifying and grouping orders according to the steel grade and the sub-plate width, and determining the slab width and the slab thickness in each group of orders; A rule determination module for determining the slab design rule and the slab steel coil sub-plate conversion rule in each group of orders according to production conditions; A condition constraint module for determining the target function and the constraint conditions of the optimal billet weight design in each group of orders according to the slab width, the slab thickness, the slab design rule and the slab steel coil sub-plate conversion rule; An optimization calculation module for obtaining the maximum size ratio and the optimal billet weight in each group of orders that satisfy the constraint conditions and maximize the target function based on the particle swarm optimization algorithm.
9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the particle swarm optimization-based optimal billet weight calculation method of any one of claims 1-7 when executing the program.
10. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program implements the particle swarm optimization-based optimal billet weight calculation method of any one of claims 1-7 when executed by the processor.