Isothermal extrusion energy consumption optimization method based on particle swarm optimization

Abstract

The invention belongs to the technical field of isothermal extrusion parameter optimization, and particularly relates to an isothermal extrusion energy consumption optimization method based on a particle swarm optimization algorithm, which comprises the following steps: S1, constructing an extrusion energy consumption model of a profile to obtain a process parameter combination strongly associated with processing energy consumption; s2, carrying out extrusion forming process simulation on the basis of the strong-correlation process parameter combination to obtain machining simulation data; s3, the machining simulation data are used for training the machining prediction model, and the incidence relation between the machining parameters and the energy consumption and profile quality indexes is obtained; s4, based on the trained prediction model, performing multi-objective optimization solution to obtain optimized processing parameters; the targets of the multi-target optimization comprise the lowest energy consumption and the satisfaction of the quality index standard; and S5, the machining parameters obtained in the step S4 are used for conducting actual isothermal extrusion machining on the profile. According to the method, reasonable process parameters can be obtained, the energy consumption in the isothermal extrusion machining process is effectively optimized, and meanwhile the product quality can be collaboratively optimized.

Description

Technical Field

[0001] The invention belongs to the technical field of isothermal extrusion parameter optimization, and in particular relates to an isothermal extrusion energy consumption optimization method based on a particle swarm algorithm. Background Art

[0002] Isothermal extrusion is an advanced processing method that can significantly improve the quality and dimensional accuracy of profiles (such as aluminum profiles) by maintaining a constant temperature of the metal at the die exit during the extrusion process, and reduce the problem of uneven material properties caused by temperature fluctuations. However, the energy consumption of the isothermal extrusion process has always been one of the key factors restricting its widespread application and cost-effectiveness.

[0003] At present, the setting of isothermal extrusion process parameters mainly relies on experience accumulation and product trial and error, and energy consumption factors are rarely directly considered, resulting in the widespread coexistence of high energy consumption and low efficiency. With the rising energy costs and increasingly stringent environmental protection requirements, the aluminum profile industry urgently needs to solve the energy consumption problem in the isothermal extrusion process and achieve a balanced control of energy consumption costs and product quality. Especially in the context of intensified competition and shrinking profit margins in the aluminum profile industry, optimizing extrusion process parameters, reducing energy consumption, and improving production efficiency and product quality are crucial to enhancing corporate competitiveness.

[0004] In previous studies, although some scholars have conducted preliminary explorations on the energy consumption model of the extrusion process and the influence of process parameters on energy consumption and quality, most studies focus on single-objective optimization, or focus on improving product quality or reducing energy consumption. Few studies can take into account the coordinated optimization of isothermal extrusion energy consumption and product quality. In addition, the construction of energy consumption models in isothermal extrusion processes mostly focuses on theoretical derivation and simplified assumptions, and lacks full consideration of complex factors in actual production, such as heat conduction of equipment and wear of molds, which limits the applicability and accuracy of the model. In terms of algorithm application, although the particle swarm algorithm has shown great potential in the field of multi-objective optimization due to its simplicity and fast convergence speed, its application in isothermal extrusion process parameter optimization is still relatively rare, especially in combining the support vector machine prediction model to achieve accurate prediction and optimization of the complex relationship between process parameters and energy consumption and quality. There is a large research gap.

[0005] Therefore, how to obtain reasonable process parameters that can effectively optimize the energy consumption of the isothermal extrusion process while also synergistically optimizing product quality. Summary of the invention

[0006] In view of the above-mentioned deficiencies of the prior art, the present invention provides an isothermal extrusion energy consumption optimization method based on a particle swarm algorithm, which can obtain reasonable process parameters, effectively optimize the energy consumption of the isothermal extrusion process, and simultaneously optimize the product quality.

[0007] In order to solve the above technical problems, the present invention adopts the following technical solutions: An isothermal extrusion energy consumption optimization method based on particle swarm algorithm comprises the following steps: S1. Construct an extrusion energy consumption model of profiles to study the direct relationship between processing energy consumption and various process parameters; and based on the analysis of the extrusion energy consumption model, obtain a combination of process parameters that is strongly correlated with processing energy consumption; S2. Based on the selection of a strongly correlated process parameter combination, an extrusion forming process is simulated to obtain processing simulation data; the processing simulation data includes parameter data of the process parameter combination and corresponding processing result data; the processing result data includes energy consumption and profile quality indicators; S3, using the processing simulation data obtained in S2, training the preset processing prediction model to obtain the correlation between the processing parameters and the energy consumption and profile quality indicators; wherein the input of the processing prediction model is the processing parameters, and the output is the energy consumption and profile quality indicators; S4. Based on the trained prediction model, multi-objective optimization is performed to obtain optimized processing parameters; the objectives of the multi-objective optimization include minimizing energy consumption and meeting quality index standards; S5. Using the processing parameters obtained in S4, actual isothermal extrusion processing of the profile is performed.

[0008] Compared with the prior art, the present invention has the following beneficial effects: 1. Using this method, after selecting a strongly correlated process parameter combination, and using it to perform processing simulation to obtain processing simulation data; then, the preset processing prediction model will be trained using the processing simulation data to obtain the correlation between the processing parameters and the energy consumption and profile quality indicators, and then based on the trained prediction model, multi-objective optimization will be performed to obtain the optimized processing parameters. Among them, the goals of multi-objective optimization include minimizing energy consumption and meeting quality indicator standards.

[0009] The process parameters obtained by the multi-objective optimization solution can significantly reduce the energy consumption in the profile extrusion process, as one of the goals is to minimize energy consumption, which is conducive to achieving green manufacturing and energy conservation and emission reduction. In addition, as another goal is to meet the quality index standard, the optimization process not only focuses on energy consumption, but also considers the quality indicators of the profile (such as dimensional accuracy, surface quality, etc.), which can ensure that product quality is not sacrificed while reducing energy consumption.

[0010] 2. In this method, by constructing a profile extrusion energy consumption model, the direct relationship between processing energy consumption and various process parameters is studied, and based on this, a combination of process parameters that is strongly associated with processing energy consumption is obtained. Afterwards, simulation and prediction models are used to replace some actual experiments. This setting can quickly screen out effective process parameter combinations, thereby shortening the product development and optimization cycle.

[0011] In summary, this method can obtain reasonable process parameters, which can not only effectively optimize the energy consumption of the isothermal extrusion process, but also synergistically optimize the product quality.

[0012] Preferably, the process parameters include extrusion speed, initial temperature of the blank, temperature gradient of the blank, and metal deformation resistance.

[0013] Preferably, in S1, based on the upper limit theory and combined with the isothermal extrusion mechanism, an isothermal extrusion forming energy consumption model is constructed; the expression of the isothermal extrusion forming energy consumption model is: E = f(v, T i , △T, F); In the formula, E represents energy consumption, v is the extrusion speed, T i is the initial temperature of the blank, △T is the temperature gradient of the blank, and F is the metal deformation resistance.

[0014] This setting, by incorporating key process parameters in the isothermal extrusion process (such as extrusion speed v, billet initial temperature Ti, billet temperature gradient △T and metal deformation resistance F) into the energy consumption model, can more accurately reflect the energy consumption in the actual extrusion process. This accuracy helps the subsequent optimization work to be closer to actual production needs.

[0015] The upper limit theory is used as the theoretical basis to make the model more physically reasonable and explainable. The upper limit theory provides a method to estimate the upper limit of energy consumption in the material processing process, thereby ensuring the theoretical reliability and rigor of the model. A clear relationship between energy consumption and key process parameters (E=f(v, Ti, △T, F)) is established, which provides a solid foundation for the subsequent process parameter optimization. By adjusting these parameters, the energy consumption performance under different process conditions can be systematically explored to find the optimal process parameter combination.

[0016] Preferably, in S2, a simulation model of the isothermal extrusion process of the profile is established by a finite element analysis method; and then the energy consumption and profile quality indicators under different parameters are calculated by the simulation model; the profile quality indicators include the profile outlet temperature distribution.

[0017] With such a setting, the finite element analysis method can accurately simulate the complex phenomena such as material flow, deformation, temperature distribution, etc. during isothermal extrusion, providing a solid foundation for predicting energy consumption and profile quality under different process parameters. Through the simulation model, the influence of each parameter on energy consumption and profile quality can be intuitively seen, so as to more accurately predict the situation in actual production. In addition, the simulation model allows the influence of changes in multiple process parameters (such as extrusion speed, initial temperature of billet, temperature gradient, metal deformation resistance, etc.) on energy consumption and profile quality to be considered simultaneously. This multi-parameter analysis capability helps to fully understand the inherent mechanism of the extrusion process and find the best combination between the parameters. Compared with traditional experimental methods, finite element simulation can greatly reduce experimental costs and time. Through the simulation model, the effects of different parameter combinations can be quickly tested in a virtual environment, so as to quickly screen out the optimal process parameters and reduce the trial and error costs and time in actual production.

[0018] Preferably, S3, use support vector regression SVR or other support vector machine variants to build a prediction model, and select a suitable kernel function to adapt to nonlinear relationships.

[0019] With this setting, SVR and its variants can map input data to a high-dimensional feature space by introducing kernel functions, thereby effectively processing nonlinear relationships. This means that even if the original data presents a complex nonlinear relationship in the input space, SVR can find a linear regression function in the high-dimensional space through the mapping of the kernel function, thereby achieving accurate prediction of nonlinear problems. This capability gives SVR a significant advantage in processing process parameters such as metal deformation resistance in metal extrusion processes with complex nonlinear characteristics. The prediction model constructed using support vector machines provides a scientific basis for the intelligent selection of process parameters, making the production process more stable and controllable, reducing dependence on manual experience, and reducing production risks.

[0020] Preferably, in S4, a particle swarm optimization algorithm is used to perform multi-objective optimization solution.

[0021] In this setting, the particle swarm optimization algorithm is an optimization algorithm based on swarm intelligence. It finds the optimal solution by simulating the behavior of natural groups such as bird flocks or fish schools. In multi-objective optimization problems, the particle swarm optimization algorithm can consider multiple objective functions at the same time, and gradually find a non-inferior solution set (Pareto frontier) that satisfies all objective functions through communication and collaboration between particles. This solution method is not only efficient, but also able to handle complex nonlinear and multimodal optimization problems. Using the particle swarm optimization algorithm for multi-objective optimization has significant effects such as efficient solution ability, powerful global search ability, flexible parameter adjustment, easy expansion and integration, and a wide range of application fields. These effects make the particle swarm optimization algorithm have important application value in multi-objective optimization problems.

[0022] Preferably, the solution process of the particle swarm optimization algorithm includes: 1) Initialization: Generate a group of multiple particles, each particle represents a set of process parameters; 2) Evaluation: Based on the support vector machine prediction model, the energy consumption and profile quality of each particle under the corresponding process parameters are calculated; 3) Update: Update the individual optimal position of the particle swarm and the global optimal position; and adjust the position of the particles according to the speed and position update rules of the particle swarm algorithm; 4) Iteration: Repeat the above process until the preset termination condition is met.

[0023] With this setup, through the initialization step, the particle swarm optimization algorithm generates a group of multiple particles, each of which represents a possible combination of process parameters. This enables the algorithm to search in a wide range of parameter spaces to find the optimal or near-optimal combination of process parameters. The evaluation step uses the trained support vector machine prediction model to quickly calculate the energy consumption and profile quality under each set of process parameters, providing accurate data support for subsequent optimization.

[0024] Moreover, the core of the particle swarm optimization algorithm lies in its intelligent search strategy. By continuously updating the individual optimal position and global optimal position of the particle swarm, and adjusting the position of the particles according to the speed and position update rules of the particle swarm algorithm, the algorithm can gradually approach the optimal solution. This search strategy based on swarm intelligence enables the algorithm to make full use of existing information, avoid blind search, and thus improve search efficiency.

[0025] In addition, in multi-objective optimization problems, there are often conflicts and contradictions between different objectives. The particle swarm optimization algorithm can find a balance between multiple objectives by considering multiple objective functions (such as energy consumption and profile quality) at the same time, thereby obtaining a set of non-inferior solutions that meet all objective requirements. This is of great significance for improving product quality and reducing production costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to make the purpose, technical solution and advantages of the invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings, in which: Figure 1 A flowchart of the method is shown in FIG. Figure 2 Flow chart of the particle swarm algorithm in the embodiment. DETAILED DESCRIPTION

[0027] The following is a further detailed description through specific implementation methods: Example: like Figure 1As shown, this embodiment discloses an isothermal extrusion energy consumption optimization method based on a particle swarm algorithm, comprising the following steps: S1. Construct an extrusion energy consumption model for profiles to study the direct relationship between processing energy consumption and various process parameters; and based on the analysis of the extrusion energy consumption model, obtain a combination of process parameters that are strongly associated with processing energy consumption. Among them, the process parameters include extrusion speed, initial billet temperature, billet temperature gradient, and metal deformation resistance.

[0028] In the specific implementation, based on the upper limit theory and combined with the isothermal extrusion mechanism, an isothermal extrusion forming energy consumption model is constructed; the expression of the isothermal extrusion forming energy consumption model is: E=f(v,T i , △T, F); In the formula, E represents energy consumption, v is the extrusion speed, T i is the initial temperature of the blank, △T is the temperature gradient of the blank, and F is the metal deformation resistance.

[0029] In this way, by incorporating the key process parameters in the isothermal extrusion process (such as extrusion speed v, initial billet temperature Ti, billet temperature gradient △T and metal deformation resistance F) into the energy consumption model, the model can more accurately reflect the energy consumption in the actual extrusion process. This accuracy helps the subsequent optimization work to be closer to the actual production needs. The upper limit theory is used as the theoretical basis, making the model more physically reasonable and explainable. The upper limit theory provides a method to estimate the upper limit of energy consumption in the material processing process, thereby ensuring the theoretical reliability and rigor of the model. A clear relationship between energy consumption and key process parameters (E=f(v, Ti, △T, F)) is established, which provides a solid foundation for the subsequent optimization of process parameters. By adjusting these parameters, the energy consumption performance under different process conditions can be systematically explored to find the optimal combination of process parameters.

[0030] S2. Based on the selection of strongly correlated process parameter combinations, the extrusion forming process is simulated to obtain processing simulation data; the processing simulation data includes parameter data of the process parameter combination and corresponding processing result data; the processing result data includes energy consumption and profile quality indicators.

[0031] In specific implementation, a simulation model of the isothermal extrusion process of the profile is established through the finite element analysis method; then the energy consumption and profile quality indicators under different parameters are calculated through the simulation model; the profile quality indicators include the profile outlet temperature distribution.

[0032] With such a setting, the finite element analysis method can accurately simulate the complex phenomena such as material flow, deformation, temperature distribution, etc. during isothermal extrusion, providing a solid foundation for predicting energy consumption and profile quality under different process parameters. Through the simulation model, the influence of each parameter on energy consumption and profile quality can be intuitively seen, so as to more accurately predict the situation in actual production. In addition, the simulation model allows the influence of changes in multiple process parameters (such as extrusion speed, initial temperature of billet, temperature gradient, metal deformation resistance, etc.) on energy consumption and profile quality to be considered simultaneously. This multi-parameter analysis capability helps to fully understand the inherent mechanism of the extrusion process and find the best combination between the parameters. Compared with traditional experimental methods, finite element simulation can greatly reduce experimental costs and time. Through the simulation model, the effects of different parameter combinations can be quickly tested in a virtual environment, so as to quickly screen out the optimal process parameters and reduce the trial and error costs and time in actual production.

[0033] S3. Use the processing simulation data obtained in S2 to train the preset processing prediction model to obtain the correlation between the processing parameters and the energy consumption and profile quality indicators; wherein the input of the processing prediction model is the processing parameters, and the output is the energy consumption and profile quality indicators.

[0034] In specific implementation, support vector regression SVR or other support vector machine variants are used to build a prediction model, and a suitable kernel function (such as Gaussian radial basis function RBF) is selected to adapt to nonlinear relationships.

[0035] With this setting, SVR and its variants can map input data to a high-dimensional feature space by introducing kernel functions, thereby effectively processing nonlinear relationships. This means that even if the original data presents a complex nonlinear relationship in the input space, SVR can find a linear regression function in the high-dimensional space through the mapping of the kernel function, thereby achieving accurate prediction of nonlinear problems. This capability gives SVR a significant advantage in processing process parameters such as metal deformation resistance in metal extrusion processes with complex nonlinear characteristics. The prediction model constructed using support vector machines provides a scientific basis for the intelligent selection of process parameters, making the production process more stable and controllable, reducing dependence on manual experience, and reducing production risks.

[0036] S4. Based on the trained prediction model, multi-objective optimization is performed to obtain optimized processing parameters; the objectives of the multi-objective optimization include minimizing energy consumption and meeting quality index standards.

[0037] In the specific implementation, the particle swarm optimization algorithm is used to solve multi-objective optimization. With such a setting, the particle swarm optimization algorithm is an optimization algorithm based on swarm intelligence. It finds the optimal solution by simulating the behavior of natural groups such as bird flocks or fish schools. In multi-objective optimization problems, the particle swarm optimization algorithm can consider multiple objective functions at the same time, and gradually find a non-inferior solution set (Pareto frontier) that satisfies all objective functions through communication and collaboration between particles. This solution method is not only efficient, but also able to handle complex nonlinear and multi-modal optimization problems. Using the particle swarm optimization algorithm for multi-objective optimization has significant effects such as efficient solution ability, powerful global search ability, flexible parameter adjustment, easy expansion and integration, and a wide range of application fields. These effects make the particle swarm optimization algorithm have important application value in multi-objective optimization problems.

[0038] Among them, Figure 2 As shown in Figure 2, the solution process of the particle swarm optimization algorithm includes: 1) Initialization: Generate a group of multiple particles, each particle represents a set of process parameters; 2) Evaluation: Based on the support vector machine prediction model, the energy consumption and profile quality of each particle under the corresponding process parameters are calculated; 3) Update: Update the individual optimal position of the particle swarm and the global optimal position; and adjust the position of the particles according to the speed and position update rules of the particle swarm algorithm; 4) Iteration: Repeat the above process until the preset termination condition is met.

[0039] With such a setting, through the initialization step, the particle swarm optimization algorithm generates a group of multiple particles, each particle represents a set of possible process parameter combinations. This enables the algorithm to search in a wide range of parameter spaces to find the optimal or near-optimal process parameter combination. The evaluation step uses the trained support vector machine prediction model to quickly calculate the energy consumption and profile quality under each set of process parameters, providing accurate data support for subsequent optimization. In addition, the core of the particle swarm optimization algorithm lies in its intelligent search strategy. By continuously updating the individual optimal position and global optimal position of the particle swarm, and adjusting the position of the particles according to the speed and position update rules of the particle swarm algorithm, the algorithm can gradually approach the optimal solution. This search strategy based on swarm intelligence enables the algorithm to make full use of existing information, avoid blind search, and thus improve search efficiency. In addition, in multi-objective optimization problems, there are often conflicts and contradictions between different objectives. The particle swarm optimization algorithm can find a balance point between multiple objectives by considering multiple objective functions (such as energy consumption and profile quality) at the same time, thereby obtaining a set of non-inferior solutions that meet all objective requirements. This is of great significance for improving product quality and reducing production costs.

[0040] S5. Using the processing parameters obtained in S4, actual isothermal extrusion processing of the profile is performed.

[0041] Using this method, a strongly correlated process parameter combination is selected, and the processing simulation is performed to obtain the processing simulation data; then, the preset processing prediction model is trained using the processing simulation data, and after obtaining the correlation between the processing parameters and the energy consumption and the profile quality index, a multi-objective optimization solution is performed based on the trained prediction model to obtain the optimized processing parameters. Among them, the goals of the multi-objective optimization include the lowest energy consumption and meeting the quality index standard. The process parameters obtained by the multi-objective optimization solution can significantly reduce the energy consumption in the profile extrusion process, since one of the goals is the lowest energy consumption, which is conducive to the realization of green manufacturing and energy conservation and emission reduction. In addition, since another goal is to meet the quality index standard, the optimization process not only focuses on energy consumption, but also considers the quality index of the profile (such as dimensional accuracy, surface quality, etc.), which can ensure that the product quality is not sacrificed while reducing energy consumption. In addition, in this method, by constructing a profile extrusion energy consumption model, the direct relationship between the processing energy consumption and each process parameter is studied, and based on this, a process parameter combination strongly associated with the processing energy consumption is obtained. Afterwards, the simulation and prediction model is used to replace some actual experiments. Such a setting can quickly screen out effective process parameter combinations, thereby shortening the product development and optimization cycle. In summary, this method can obtain reasonable process parameters, effectively optimize the energy consumption of the isothermal extrusion process, and also coordinately optimize the product quality.

[0042] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit the technical solution. Those skilled in the art should understand that those modifications or equivalent substitutions of the technical solution of the present invention that do not depart from the purpose and scope of the technical solution should be included in the scope of the claims of the present invention.

Claims

1. An isothermal extrusion energy consumption optimization method based on particle swarm algorithm, characterized in that: The following steps are involved: S1. Construct an extrusion energy consumption model of profiles to study the direct relationship between processing energy consumption and various process parameters; and based on the analysis of the extrusion energy consumption model, obtain a combination of process parameters that is strongly correlated with processing energy consumption; S2. Based on the selection of a strongly correlated process parameter combination, an extrusion forming process is simulated to obtain processing simulation data; the processing simulation data includes parameter data of the process parameter combination and corresponding processing result data; the processing result data includes energy consumption and profile quality indicators; S3, using the processing simulation data obtained in S2, training the preset processing prediction model to obtain the correlation between the processing parameters and the energy consumption and profile quality indicators; wherein the input of the processing prediction model is the processing parameters, and the output is the energy consumption and profile quality indicators; S4. Based on the trained prediction model, multi-objective optimization is performed to obtain optimized processing parameters; the objectives of the multi-objective optimization include minimizing energy consumption and meeting quality index standards; S5. Using the processing parameters obtained in S4, actual isothermal extrusion processing of the profile is performed.

2. The isothermal extrusion energy consumption optimization method based on particle swarm algorithm according to claim 1, characterized in that: Process parameters include extrusion speed, initial temperature of billet, temperature gradient of billet, and metal deformation resistance.

3. The isothermal extrusion energy consumption optimization method based on particle swarm algorithm according to claim 2, characterized in that: In S1, based on the upper limit theory and combined with the isothermal extrusion mechanism, an isothermal extrusion forming energy consumption model is constructed; the expression of the isothermal extrusion forming energy consumption model is: E=f(v,T i ,△T,F); In the formula, E represents energy consumption, v is the extrusion speed, T i is the initial temperature of the blank, △T is the temperature gradient of the blank, and F is the metal deformation resistance.

4. The isothermal extrusion energy consumption optimization method based on particle swarm algorithm according to claim 1, characterized in that: In S2, a simulation model of the isothermal extrusion process of the profile is established through the finite element analysis method; then the energy consumption and profile quality indicators under different parameters are calculated through the simulation model; the profile quality indicators include the profile outlet temperature distribution.

5. The isothermal extrusion energy consumption optimization method based on particle swarm algorithm according to claim 1, characterized in that: S3. Use support vector regression SVR or other support vector machine variants to build a prediction model and select a suitable kernel function to adapt to nonlinear relationships.

6. The isothermal extrusion energy consumption optimization method based on particle swarm algorithm according to claim 1, characterized in that: In S4, the particle swarm optimization algorithm is used to solve the multi-objective optimization problem.

7. The isothermal extrusion energy consumption optimization method based on particle swarm algorithm according to claim 6, characterized in that: The solution process of the particle swarm optimization algorithm includes: 1) Initialization: Generate a group of multiple particles, each particle represents a set of process parameters; 2) Evaluation: Based on the support vector machine prediction model, the energy consumption and profile quality of each particle under the corresponding process parameters are calculated; 3) Update: Update the individual optimal position of the particle swarm and the global optimal position; and adjust the position of the particles according to the speed and position update rules of the particle swarm algorithm; 4) Iteration: Repeat the above process until the preset termination condition is met.

Images