A dual-objective optimization decision method for a slag cooling process

Abstract

The application provides a dual-target optimization decision method for a slag cooling process, and belongs to the technical field of industrial production process optimization. The method collects historical slag cooling data and performs preprocessing, and uses a multiple linear regression model to predict the temperature and cooling time after cooling. A dynamic weighted objective function is further constructed, and a hybrid algorithm (a combination of genetic algorithm, particle swarm algorithm and reinforcement learning algorithm) is used to solve the optimal control parameters. Finally, the optimal control parameters obtained are integrated into a real-time control system, and the control parameters are dynamically adjusted according to the real-time monitoring data to realize dual-target optimization of the cooling time and the furnace temperature fluctuation. The application adopts the above-mentioned dual-target optimization decision method for the slag cooling process, which can effectively improve the efficiency and stability of the slag cooling process, reduce energy consumption, and has significant economic and environmental benefits.

Description

Technical Field

[0001] This invention relates to the field of industrial production process optimization technology, and in particular to a dual-objective optimization decision-making method for slag cooling process. Background Technology

[0002] In modern industrial production, slag cooling is an indispensable step in high-temperature processes such as steelmaking. Slag cooling not only affects production efficiency but also directly relates to energy consumption, equipment lifespan, and the stability of subsequent processes. However, existing slag cooling technologies have the following shortcomings:

[0003] Low cooling efficiency: Traditional slag cooling methods typically employ a single cooling strategy, such as fixed airflow or fixed cooling time, which cannot be adjusted in real time according to dynamic factors such as the initial temperature of the slag and the amount of slag discharged. This results in excessively long cooling times, low production efficiency, and high energy consumption.

[0004] Unstable temperature control: During the cooling process, the furnace temperature fluctuates significantly and is difficult to control precisely. This unstable temperature change may lead to uneven slag cooling, affecting product quality and even damaging the cooling equipment.

[0005] Lack of optimized decision-making mechanisms: In existing technologies, the setting of cooling parameters mostly relies on experience or fixed process parameters, lacking scientific optimization decision-making methods. This makes it impossible for the cooling process to achieve an optimal balance between cooling time and temperature stability, making it difficult to meet the requirements of efficient production and energy conservation.

[0006] Insufficient real-time adjustment capability: In actual production, the properties of slag and the cooling environment are constantly changing, but existing technologies cannot monitor and dynamically adjust cooling parameters in real time. This makes the cooling process difficult to adapt to complex production conditions and unable to effectively cope with emergencies.

[0007] Insufficient data utilization: Although a large amount of slag cooling data has been accumulated in modern industrial production, existing technologies have failed to fully utilize this data for in-depth analysis and optimization modeling. The value of the data has not been fully explored, and it cannot provide scientific guidance for the cooling process. Summary of the Invention

[0008] The purpose of this invention is to provide a dual-objective optimization decision-making method for the slag cooling process, which can effectively improve the efficiency and stability of the slag cooling process, reduce energy consumption, and has significant economic and environmental benefits.

[0009] To achieve the above objectives, the present invention provides a dual-objective optimization decision-making method for slag cooling processes, comprising the following steps:

[0010] Step S1: Collect and preprocess historical slag cooling data, and use a multiple linear regression model to predict the temperature T and cooling time t after cooling, respectively.

[0011] Step S2: Construct a dynamically weighted objective function;

[0012] Step S3: Solve for the optimal control parameters using a hybrid algorithm; the hybrid algorithm is a combination of genetic algorithm, particle swarm optimization algorithm, and reinforcement learning algorithm.

[0013] Step S4: Integrate the obtained optimal control parameters into the real-time control system, and dynamically adjust the control parameters based on the real-time monitoring data.

[0014] Preferably, in step S1, the historical slag cooling data includes: steel belt speed v, damper opening d, initial temperature T0, slag discharge rate Q, cooling air volume F, and cooling air temperature T. f .

[0015] Preferably, in step S1, the preprocessing includes: data cleaning and data normalization.

[0016] Preferably, in step S1, the multiple linear regression model includes a temperature regression model and a cooling time regression model;

[0017] The temperature regression model is as follows:

[0018] T = β 0T +β 1T v+β 2T d+β 3T T0+β 4T Q+β 5T F+β 6T T f +ε T ;

[0019] The cooldown time regression model is as follows:

[0020] t = β 0t +β 1t v+β 2t d+β 3t T0+β 4t Q+β 5t F+β 6t T f +ε t ;

[0021] Where, β 0T β 1T β 2T β 3T β 4T β 5T β 6T All represent the regression coefficients of the temperature regression model; β0t β 1t β 2t β 3t β 4t β 5t β 6t All represent the regression coefficients of the cooling time regression model; ε T ε t This indicates the error term.

[0022] Preferably, in step S2, the dynamically weighted objective function as follows:

[0023]

[0024] Where α and β represent dynamic weighting coefficients; t(v,d) represents cooling time; and ΔT(v,d) represents furnace temperature fluctuation.

[0025] The preferred formula for calculating the dynamic weighting coefficient is as follows:

[0026]

[0027] Where, α max α min These represent the maximum and minimum values ​​of α, respectively; β max β min T represents the maximum and minimum values ​​of β, respectively; max This indicates the maximum number of iterations in the optimization process.

[0028] Preferably, in step S3, the genetic algorithm is as follows:

[0029] Using a dynamically weighted objective function As a fitness function, the fitness value of each individual is calculated;

[0030] Two individuals are selected for crossover based on an adaptive crossover rate to generate a new individual;

[0031] Individuals are mutated according to an adaptive mutation rate to increase population diversity;

[0032] Using a tournament selection method, individuals with higher fitness are selected to enter the next generation;

[0033] The global optimal position g of the particle swarm optimization algorithm is directly added as an elite individual to the next generation of the genetic algorithm:

[0034] GAPopulation new =GAPopulation selected ∪{g};

[0035] Among them, GAPopulationnew GAPopulation represents the next generation population in a genetic algorithm. selected This represents a subset of individuals with high fitness selected from the current population through a tournament selection method.

[0036] Preferably, in step S3, the particle swarm optimization algorithm is as follows:

[0037] An initial population is randomly generated, with each individual representing a set of control parameters (v, d), and the population size is N.

[0038] Initialize the position and velocity of the particle swarm;

[0039] Adjust the particle velocity and position according to the update formula of the particle swarm algorithm;

[0040] The inertia weight w is dynamically adjusted based on the number of iterations.

[0041]

[0042] Among them, w max w min These represent the maximum and minimum values ​​of the inertia weight, respectively; S max s represents the maximum number of iterations; s represents the current number of iterations.

[0043] Compare the current position of the particle with the individual's optimal position p i The solution of the individual with the highest fitness in the genetic algorithm is fed back to the particle swarm algorithm to update the optimal position of the individual;

[0044] Then compare the global optimal position with the global optimal position and update the global optimal position g:

[0045]

[0046]

[0047] in, This represents the position of the i-th particle in the (s+1)-th iteration; p represents the optimal individual position of the i-th particle. i The corresponding fitness function value; This represents the fitness function value corresponding to the globally optimal position.

[0048] Preferably, in step S3, the reinforcement learning algorithm is as follows:

[0049] Select action a based on the current state c;

[0050] Perform action a, adjust control parameters v and d, and calculate the reward value R:

[0051]

[0052] Update the Q value according to the Bellman equation;

[0053] Update state c to the new state c', repeat the above process, and apply the reinforcement learning optimized control parameters (v) * ,d * The results are fed back to the genetic algorithm and particle swarm optimization algorithm to update the population and particle positions.

[0054]

[0055] Among them, v * d represents the optimal steel belt speed; * This indicates the optimal damper opening.

[0056] Preferably, in step S4, the optimal control parameters obtained by solving are integrated into the real-time control system, and the control parameters are dynamically adjusted according to the real-time monitoring data. The specific operation is as follows:

[0057] The temperature and cooling time during the cooling process are acquired in real time via sensors and denoted as X. real-time ;

[0058] Predict the current cooling time and temperature fluctuations based on the temperature regression model and cooling time regression model in step S1:

[0059]

[0060] in, This indicates the predicted temperature after cooling. Indicates the predicted cooldown time; f T f t A function representing the prediction model;

[0061] Based on the predicted cooling time and temperature fluctuations, and combined with the optimal control parameters from step S3, the steel belt speed and damper opening are dynamically adjusted:

[0062] v new =v current +Δv;

[0063] d new =d current +Δd;

[0064] Among them, v current d current Indicates the current control parameter; v new d new This represents the adjusted control parameters; Δv and Δd represent the adjustment amounts calculated based on the prediction model and the optimal control parameters.

[0065] Δv=k v (v* -v current );

[0066] Δd=k d (d * -d current );

[0067] Where, k v k d This represents the adaptive control coefficient.

[0068] Therefore, the present invention employs the above-mentioned dual-objective optimization decision-making method for slag cooling process, and the beneficial technical effects are as follows:

[0069] (1) Significantly improves cooling efficiency:

[0070] This invention dynamically adjusts cooling parameters using a multiple linear regression model and a hybrid optimization algorithm. Compared to traditional fixed-parameter cooling methods, this significantly shortens cooling time, improves production efficiency, and reduces energy consumption, resulting in significant economic and environmental benefits.

[0071] (2) Achieve high-precision and stable control of furnace temperature:

[0072] This invention employs a dynamic weighted objective function, which effectively balances cooling time and furnace temperature fluctuations during the cooling process, significantly improving temperature stability. Compared to existing technologies where temperature fluctuations are significant, this invention ensures a uniform and efficient cooling process, further enhancing product quality.

[0073] (3) Enhance real-time dynamic adjustment capabilities:

[0074] This invention integrates optimized control parameters into a real-time control system, dynamically adjusting cooling parameters based on real-time monitoring data, enabling rapid response to changes in the production process. Compared to the shortcomings of existing technologies that lack real-time adjustment capabilities, this invention significantly improves the adaptability and flexibility of the cooling process.

[0075] (4) Optimize the innovation and intelligence of decision-making mechanisms:

[0076] This invention employs a hybrid optimization strategy combining genetic algorithms, particle swarm optimization, and reinforcement learning algorithms, overcoming the limitations of traditional empirical settings and achieving scientific optimization and dynamic adjustment of cooling parameters. This intelligent decision-making mechanism significantly improves the intelligence level of the cooling process and has broad applicability and promotional value. Attached Figure Description

[0077] Figure 1 This is a flowchart of a dual-objective optimization decision-making method for a slag cooling process according to the present invention;

[0078] Figure 2Optimization flowchart for hybrid algorithm. Detailed Implementation

[0079] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0080] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0081] Example 1

[0082] The slag cooling system of a steel plant suffers from low cooling efficiency, high energy consumption, and large temperature fluctuations under traditional cooling methods. To address these issues, the present invention employs a dual-objective optimization decision-making method for the slag cooling process to optimize the cooling process, improve production efficiency, and reduce energy consumption.

[0083] like Figure 1 The diagram shown is a flowchart of a dual-objective optimization decision-making method for a slag cooling process according to the present invention, comprising:

[0084] Step S1: Collect historical data on the slag cooling process over the past year, including the following parameters:

[0085] Steel belt speed (unit: m / min);

[0086] Damper opening (unit: %);

[0087] Initial temperature (unit: °C);

[0088] Slag discharge rate (unit: t / h);

[0089] Cooling air volume (unit: m) 3 / min);

[0090] Cooling air temperature (unit: °C);

[0091] Cooling time (unit: min);

[0092] Temperature after cooling (unit: °C);

[0093] The data comes from the factory's production records and sensor monitoring system.

[0094] Preprocessing includes data cleaning and data normalization.

[0095] Data cleaning: Removing missing and outlier values. For example, removing data points with abnormally short or long cooling times, and records with initial temperatures that clearly do not meet production conditions.

[0096] Data normalization: Normalize all parameters to the [0, 1] interval for subsequent modeling and optimization.

[0097] The construction of multiple linear regression models includes temperature regression models and cooling time regression models;

[0098] The temperature regression model is as follows:

[0099] T = β 0T +β 1T v+β 2T d+β 3T T0+β 4T Q+β 5T F+β 6T T f +ε T ;

[0100] The cooldown time regression model is as follows:

[0101] t = β 0t +β 1t v+β 2t d+β 3t T0+β 4t Q+β 5t F+β 6t T f +ε t ;

[0102] Where, β 0T β 1T β 2T β 3T β 4T β 5T β 6T All represent the regression coefficients of the temperature regression model; β 0t β 1t β 2t β 3t β 4t β 5t β 6t All represent the regression coefficients of the cooling time regression model; ε T ε t This indicates the error term.

[0103] Step S2: Construct a dynamic weighted objective function to balance cooling time and furnace temperature fluctuations;

[0104] Dynamic weighted objective function as follows:

[0105]

[0106] Where α and β represent dynamic weighting coefficients; t(v,d) represents cooling time; and ΔT(v,d) represents furnace temperature fluctuation.

[0107] The formula for calculating the dynamic weighting coefficient is as follows:

[0108]

[0109] Where, α max α min These represent the maximum and minimum values ​​of α, respectively; β max β min T represents the maximum and minimum values ​​of β, respectively; max This indicates the maximum number of iterations in the optimization process.

[0110] Step S3, as follows Figure 2 As shown, a hybrid algorithm is used to solve for the optimal control parameters.

[0111] The hybrid algorithm is a combination of genetic algorithm (GA), particle swarm optimization (PSO), and reinforcement learning algorithm.

[0112] Genetic algorithm.

[0113] The parameter settings are as follows:

[0114] Population size: 100;

[0115] Crossover rate: 0.8;

[0116] Variation rate: 0.05;

[0117] Fitness function: Dynamically weighted objective function;

[0118] Selection method: Tournament selection method;

[0119] Number of iterations: 50.

[0120] Using a dynamically weighted objective function As a fitness function, the fitness value of each individual is calculated;

[0121] Two individuals are selected for crossover based on an adaptive crossover rate to generate a new individual;

[0122] Individuals are mutated according to an adaptive mutation rate to increase population diversity;

[0123] Using a tournament selection method, individuals with higher fitness are selected to enter the next generation;

[0124] The global optimal position g of the particle swarm optimization algorithm is directly added as an elite individual to the next generation of the genetic algorithm:

[0125] GAPopulation new =GAPopulation selected ∪{g};

[0126] Among them, GAPopulation new GAPopulation represents the next generation population in a genetic algorithm. selected This represents a subset of individuals with high fitness selected from the current population through a tournament selection method.

[0127] Particle swarm optimization algorithm.

[0128] The parameter settings are as follows:

[0129] Population size: 50;

[0130] Individual learning factor: 2.0;

[0131] Social learning factor: 2.0;

[0132] Number of iterations: 50.

[0133] An initial population is randomly generated, with each individual representing a set of control parameters (v, d), and the population size is N.

[0134] Initialize the position and velocity of the particle swarm;

[0135] Adjust the particle velocity and position according to the update formula of the particle swarm algorithm;

[0136] The inertia weight w is dynamically adjusted based on the number of iterations.

[0137]

[0138] Among them, w max w min These represent the maximum and minimum values ​​of the inertia weight, respectively; S max s represents the maximum number of iterations; s represents the current number of iterations.

[0139] Compare the current position of the particle with the individual's optimal position p i The solution of the individual with the highest fitness in the genetic algorithm is fed back to the particle swarm algorithm to update the optimal position of the individual;

[0140] Then compare the global optimal position with the global optimal position and update the global optimal position g:

[0141]

[0142] in, This represents the position of the i-th particle in the (s+1)-th iteration; p represents the optimal individual position of the i-th particle. i The corresponding fitness function value; This represents the fitness function value corresponding to the globally optimal position.

[0143] Reinforcement learning algorithms.

[0144] The settings are as follows:

[0145] State space: belt speed and damper opening;

[0146] Action range: The adjustment range of steel belt speed and damper opening;

[0147] Reward function: Calculated based on cooldown time and temperature fluctuations;

[0148] Learning rate: 0.1;

[0149] Discount factor: 0.9;

[0150] Number of iterations: 100.

[0151] Select action a based on the current state c;

[0152] Perform action a, adjust control parameters v and d, and calculate the reward value R:

[0153]

[0154] Update the Q value according to the Bellman equation;

[0155] Update state c to the new state c', repeat the above process, and apply the reinforcement learning optimized control parameters (v) * ,d * The results are fed back to the genetic algorithm and particle swarm optimization algorithm to update the population and particle positions.

[0156] GAPopulation new =GAPopulation new ∪{(v * ,d * )};

[0157]

[0158] Among them, v * d represents the optimal steel belt speed; * This indicates the optimal damper opening.

[0159] Step S4: Integrate the obtained optimal control parameters into the real-time control system, and dynamically adjust the control parameters based on real-time monitoring data. The specific operation is as follows:

[0160] The temperature and cooling time during the cooling process are acquired in real time via sensors and denoted as X. real-time ;

[0161] Predict the current cooling time and temperature fluctuations based on the temperature regression model and cooling time regression model in step S1:

[0162]

[0163] in, This indicates the predicted temperature after cooling. Indicates the predicted cooldown time; f T f t The function representing the prediction model has the same specific form as in step S1;

[0164] Based on the predicted cooling time and temperature fluctuations, and combined with the optimal control parameters from step S3, the steel belt speed and damper opening are dynamically adjusted:

[0165] v new =v current +Δv;

[0166] d new =d current +Δd;

[0167] Among them, v current d current Indicates the current control parameter; v new d new This represents the adjusted control parameters; Δv and Δd represent the adjustment amounts calculated based on the prediction model and the optimal control parameters.

[0168] Δv=k v (v * -v current );

[0169] Δd=k d (d * -d current );

[0170] Where, k v k d This represents the adaptive control coefficient.

[0171] The performance improvement results obtained by using the method proposed in this invention are shown in Table 1.

[0172] Table 1 Performance Improvement Results

[0173]

[0174]

[0175] It is worth noting that all contents not described in detail in this invention are existing technologies and are well known to those skilled in the art.

[0176] Therefore, the present invention employs the above-mentioned dual-objective optimization decision-making method for slag cooling process, which can effectively improve the efficiency and stability of slag cooling process, reduce energy consumption, and has significant economic and environmental benefits.

[0177] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A dual objective optimization decision method for a slag cooling process, characterized in that, The method comprises the following steps: Step S1, collect historical slag cooling data and pre-process, use multiple linear regression model to predict the temperature after cooling and cooling time respectively Step S2, constructing a dynamic weighted objective function; Step S3, solving the optimal control parameters by using a hybrid algorithm; the hybrid algorithm is a combination of a genetic algorithm, a particle swarm algorithm and a reinforcement learning algorithm; Step S4, integrating the solved optimal control parameters into a real-time control system, and dynamically adjusting the control parameters according to real-time monitoring data; In step S1, the historical slag cooling data includes: a steel strip speed , a damper opening , an initial temperature , a slag discharge amount , a cooling air amount , and a cooling air temperature ; In step S1, the multiple linear regression model comprises a temperature regression model and a cooling time regression model; The temperature regression model is as follows: ； The cooling time regression model is as follows: ； wherein, , , , , , , all represent regression coefficients of the temperature regression model; , , , , , , all represent regression coefficients of the cooling time regression model; , represent error terms; In step S2, the dynamic weighted objective function is as follows: ； wherein, , represents a dynamic weight coefficient; represents a cooling time; represents a hearth temperature fluctuation; The calculation formula of the dynamic weight coefficient is as follows: ； ； wherein , respectively denote the maximum and minimum values of ; , respectively denote the maximum and minimum values of ; denotes the maximum number of iterations of the optimization process.

2. A bi-objective optimization decision method for a slag cooling process according to claim 1, characterized in that, In step S1, the preprocessing comprises data cleaning and data normalization.

3. A bi-objective optimization decision method for a slag cooling process according to claim 1, characterized in that, In step S3, the genetic algorithm is as follows: Using a dynamic weighted objective function As a fitness function, the fitness value of each individual is calculated; According to the adaptive crossover rate, two individuals are selected for crossover to generate new individuals; According to the adaptive mutation rate, the individuals are mutated to increase the population diversity; The tournament selection method is used to select individuals with higher fitness to enter the next generation; the global optimum position of the particle swarm algorithm as elite individuals directly into the next generation population of the genetic algorithm: ； wherein represents a new generation of the population of the genetic algorithm; represents a subset of individuals with higher fitness selected from the current population by tournament selection.

4. A bi-objective optimization decision method for a slag cooling process according to claim 3, characterized in that, In step S3, the particle swarm algorithm is as follows: randomly generate an initial population, each individual representing a set of control parameters , population size is ; The position and speed of the particle swarm are initialized; According to the update formula of the particle swarm algorithm, the speed and position of the particles are adjusted; Adjusting dynamic inertia weight according to iteration number : ； wherein, , respectively denote the maximum and minimum values of the inertial weight; denotes the maximum number of iterations; denotes the current iteration number; Comparing the position of the current particle and the individual optimal position The individual solution with the highest fitness in the genetic algorithm is fed back to the particle swarm algorithm, and the individual optimal position is updated; Then compare the global optimal position, update the global optimal position : ； ； in, Indicates the first The particle in the first The position of the next iteration; Indicates the first The optimal position of each particle The corresponding fitness function value; This represents the fitness function value corresponding to the globally optimal position.

5. A bi-objective optimization decision method for a slag cooling process according to claim 4, characterized in that, In step S3, the reinforcement learning algorithm is as follows: According to the current state Selecting an action ; performing an action , adjusting a control parameter and , calculating a reward value : ； updating according to bellman equation value; state is updated to a new state , the above process is repeated, and the control parameters optimized by reinforcement learning are fed back to the genetic algorithm and particle swarm algorithm, and the population and particle positions are updated: ； ； wherein, represents the optimal steel strip speed; represents the optimal damper opening.

6. A bi-objective optimization decision method for a slag cooling process according to claim 5, characterized in that, In step S4, the solved optimal control parameters are integrated into the real-time control system, and the control parameters are dynamically adjusted according to the real-time monitoring data, and the specific operation is as follows: The temperature and cooling time during the cooling process are recorded in real time by the sensor as Tc(t) and tc(t), respectively. ; According to the temperature regression model and the cooling time regression model in step S1, the current cooling time and temperature fluctuation are predicted: ； ； wherein, represents the predicted temperature after cooling; represents the predicted cooling time; , represents a function of the prediction model; According to the predicted cooling time and temperature fluctuation, the optimal control parameters in step S3 are combined to dynamically adjust the steel strip speed and the damper opening degree: ； ； wherein , represents the current control parameter; , represents the adjusted control parameter; , represents the adjustment calculated from the prediction model and the optimal control parameter; ； ； wherein , denotes an adaptive control coefficient.

Images