Aluminum profile extrusion production energy consumption prediction method and system

Abstract

The application discloses an aluminum profile extrusion production energy consumption prediction method and system, and relates to the technical field of aluminum profile production, and in particular to a method for predicting the energy consumption of aluminum profile production. The method comprises the following steps: designing process parameter schemes for various aluminum profile production; performing aluminum profile production experiments according to the process parameter schemes to determine the minimum production energy consumption of various existing aluminum profiles; calculating the similarity value of the cross-sectional shape of a product to be processed and the cross-sectional shape of an experimental production product; calculating the ratio of the extrusion coefficient of the product to be processed and the extrusion coefficient of the most similar experimental production product; and calculating the production energy consumption prediction value of the product to be processed by using the production energy consumption value of the most similar aluminum profile, the similarity value of the product to be processed and the most similar experimental production product, and the extrusion coefficient ratio.

Description

Technical Field

[0001] This invention belongs to the field of energy consumption prediction technology for aluminum profile extrusion production, and relates to a method for predicting energy consumption in aluminum profile extrusion production, as well as a system for predicting energy consumption in aluminum profile extrusion production. Background Technology

[0002] Aluminum profiles are widely used in construction and decoration due to their excellent ductility, ease of processing, and formability. However, the production of aluminum profiles requires not only preheating of aluminum rods and molds, but also the application of very high extrusion pressure to extrude the aluminum rods through corresponding molds to obtain the desired product shape. This results in aluminum profile production being a high-energy-consuming and high-emission process. Currently, energy conservation and emission reduction have become the mainstream trend in social production. To reduce energy consumption and emissions in the aluminum profile production process, and to save production costs, businesses and academia have been searching for methods to reduce and predict energy consumption in aluminum profile production.

[0003] Chinese patent application (application number 201911218560.3) discloses an optimization method for isothermal extrusion process parameters based on particle swarm optimization. Addressing the problem that the relationship between process parameters and metal deformation resistance is complex and unclear, making it difficult to achieve the optimal combination of process parameters constrained by qualified product quality and low production energy consumption, the method proposes using support vector machines to establish a predictive model of process parameters, isothermal extrusion forming energy consumption, and profile exit surface temperature. Using the root mean square error of extrusion forming energy consumption and profile exit surface temperature as optimization objectives, a multi-objective optimization model is constructed. The particle swarm optimization algorithm is then used to solve this multi-objective optimization model, thereby optimizing the isothermal extrusion process parameters to achieve the optimal profile quality and minimum forming energy consumption. This solution requires a large amount of data and a complex model to be implemented. Summary of the Invention

[0004] The technical solution adopted in this invention is: a method and system for predicting energy consumption in aluminum profile extrusion production. This method can determine the production energy consumption of aluminum profile products before production. The predicted values ​​can be used for energy consumption control of aluminum profile extrusion equipment and control of extrusion process parameters, thereby achieving the goals of energy conservation, emission reduction, and cost reduction in aluminum profile production. The required quantity is greatly reduced, and the model is simple.

[0005] To address this problem, the technical solution adopted by the present invention is as follows:

[0006] A method for predicting energy consumption in aluminum profile extrusion production, comprising the following steps:

[0007] 1) Design process parameter schemes for the production of various existing aluminum profile products of the enterprise using orthogonal experimental design;

[0008] 2) Conduct aluminum profile production experiments based on the process parameter scheme to determine the minimum production energy consumption for various existing aluminum profile products;

[0009] 3) Use deep learning to compare the cross-sectional shape of the product to be processed with the cross-sectional shape of the experimental product, find the experimental product that is most similar to the cross-sectional shape of the product to be processed, and calculate the similarity value of the cross-sections of the two products.

[0010] 4) Calculate the extrusion coefficient λ of the product to be processed. d The extrusion coefficient λ of the product produced in the most similar experiment b The ratio;

[0011] 5) Calculate the predicted production energy consumption of the product to be processed using the production energy consumption value of the most similar aluminum profile product, the similarity value between the product to be processed and its most similar experimental production product, and the extrusion coefficient ratio.

[0012] Furthermore, the process parameters in step 1) above include the extrusion speed of the extrusion bar, the temperature of the aluminum rod, and the temperature of the die.

[0013] Furthermore, in step 3) above, the deep learning method is a convolutional neural network or a generative adversarial network.

[0014] Furthermore, in step 3) above, the similarity value is the percentage of the number of identical shape features of the cross-section of the product to be processed and the cross-section of its most similar experimental product that are the same as the total number of features of the cross-section of its most similar experimental product.

[0015] Furthermore, the formula for calculating the predicted production energy consumption of the product to be processed in step 5) above is as follows:

[0016]

[0017] Q y —Predicted energy consumption for the production of products to be processed;

[0018] γ—Similarity value;

[0019] λ d —Extrusion coefficient of the product to be processed;

[0020] λ b —The extrusion coefficient of the product produced in the most similar experiment;

[0021] Q s —Production energy consumption value of the product produced in the most similar experiment;

[0022] An energy consumption prediction system for aluminum profile extrusion production includes,

[0023] The process parameter scheme determination module is used to determine the process parameter schemes for the production of various existing aluminum profile products;

[0024] The minimum production energy consumption determination module is used to conduct experiments based on the parameters determined by the process parameter scheme determination module to determine the minimum production energy consumption of various existing aluminum profile products.

[0025] The similarity determination module is used to determine the similarity value between the cross-sectional shape of the product to be processed and the cross-sectional shape of the experimentally produced product, as determined by deep learning.

[0026] The extrusion coefficient ratio determination module is used to determine the extrusion coefficient λ of the product to be processed. d The extrusion coefficient λ of the product produced in the most similar experiment b The ratio;

[0027] The prediction module predicts the production energy consumption of the product to be processed based on the production energy consumption value of the most similar aluminum profile product, the similarity value between the product to be processed and its most similar experimental production product, and the extrusion coefficient ratio.

[0028] The beneficial effects of the present invention are as follows: Compared with the prior art, the present invention has the following advantages:

[0029] 1) This invention constructs an energy consumption database for aluminum profile production through experiments (process generation schemes after orthogonal experiments), then uses a convolutional neural network to find the experimentally produced products most similar to the product to be processed among the existing experimentally produced products and calculates their similarity values. Next, it calculates the extrusion coefficient λ of the product to be processed. d The extrusion coefficient λ of the product produced in the most similar experiment b The energy consumption of the product is calculated using the established predictive energy consumption calculation formula. This invention aims to accurately predict the production energy consumption of new aluminum profile products by combining databases, deep learning, and energy consumption calculation models. This not only reduces the labor intensity of enterprises and lays a good foundation for achieving optimal energy consumption production planning and scheduling, but also can be used for energy consumption control in the aluminum profile production process, achieving energy conservation, emission reduction, and cost reduction for enterprises.

[0030] Moreover, compared with the Chinese patent application (application number 201911218560.3, i.e., the prior art patent) given in the background art, it has the following advantages:

[0031] (1) Regarding the optimization method of process parameters, the present invention adopts the orthogonal experimental method. Compared with the particle swarm algorithm used in the patent, the method of the present invention is simpler, faster and has less computational load.

[0032] (2) In terms of data volume, the present invention only requires a small amount of experimental data, while the comparative patent requires a large amount of data for model training;

[0033] (3) In terms of model, this invention does not need to construct a complex mathematical model, while the comparative patent needs to construct a multi-objective optimization model with the forming energy consumption and the temperature mean square error of the profile exit surface as optimization objectives, and the multi-objective optimization model needs to be solved, which requires relatively large computing resources.

[0034] (4) Regarding the calculation parameters, the present invention utilizes the energy consumption data of existing products and the design parameters of the products (such as cross-sectional shape, extrusion coefficient, etc.), while the comparative patent requires the temperature data and process parameters of the products before and after extrusion.

[0035] (5) In terms of the number of process parameters, the present invention has extrusion speed, aluminum rod temperature and die temperature, while the comparative patent has billet temperature gradient (temperature before and after extrusion is required), billet initial temperature and extrusion speed. However, die temperature is one of the important parameters of the extrusion process.

[0036] 2) The advantages of using orthogonal experimental design to determine process parameters include: 1) Reducing the number of experiments: Orthogonal experimental design, through carefully designed experimental schemes, can obtain sufficient data with fewer experiments, thereby reducing the overall number of production trials. This method can effectively save time and costs, especially in aluminum profile production, where comprehensive testing may require a large amount of raw materials and time. For example, orthogonal experimental design can evaluate the impact of multiple process parameters with fewer experiments without sacrificing the quality of results; 2) Improving experimental efficiency: Orthogonal experimental design, by selecting highly representative experimental points, makes the experimental results more universal and generalizable, thereby improving experimental efficiency. This method can quickly identify key factors affecting product quality, helping producers find the optimal combination of process parameters more quickly; 3) Reducing experimental costs: Due to the reduction in the number of experiments, the corresponding costs of raw materials, energy, and labor will also be reduced. 4) Orthogonal experimental design can clearly reveal the impact of various factors on product performance, making the results easier to interpret and generalize. This clear factor influence analysis helps producers understand how process parameters affect the quality of the final product and make process improvements accordingly. 5) Optimization of process parameters: Through orthogonal experimental design, process parameters such as extrusion speed, aluminum rod temperature, and die temperature can be systematically analyzed and optimized, thereby improving the product quality of aluminum profiles. For example, the optimal extrusion process parameters can be determined through orthogonal experimental design to obtain the required mechanical properties and microstructure. 6) High adaptability: Orthogonal experimental design can adapt to various different test conditions and requirements, including multi-factor and multi-level situations, which makes it highly flexible and adaptable in aluminum profile production.

[0037] 3) Deep learning models can automatically learn and extract features from raw data, reducing manual intervention and improving the efficiency and accuracy of feature extraction; deep learning models, especially deep neural networks, have rich hierarchical structures and can express complex feature mapping relationships, enabling them to process high-dimensional data and extract useful information from it; deep learning models exhibit good robustness in dealing with noise and outliers, which makes them perform better when dealing with data in real-world applications.

[0038] 4) This invention combines orthogonal experimental design to determine process parameters, deep learning to determine similarity values, and extrusion coefficient ratio to determine production energy consumption prediction, resulting in more scientific, reasonable, and accurate predictions. Attached Figure Description

[0039] Figure 1 A flowchart illustrating a method for predicting energy consumption in aluminum profile extrusion production;

[0040] Figure 2 This is a schematic diagram of an energy consumption prediction system for aluminum profile extrusion production. Detailed Implementation

[0041] The present invention will be further described below with reference to specific embodiments.

[0042] Example 1: As Figure 1 As shown, a method for predicting energy consumption in aluminum profile extrusion production includes the following steps:

[0043] 1) Design process parameter schemes for the production of various existing aluminum profile products of the enterprise using orthogonal experimental design; process parameters include extrusion speed of the extrusion bar, aluminum rod temperature, and die temperature.

[0044] A three-factor, ten-level experimental scheme for process parameters such as extrusion speed of the extrusion bar, aluminum rod temperature, and die temperature;

[0045] 2) Conduct aluminum profile production experiments based on the process parameter scheme to determine the minimum production energy consumption for various existing aluminum profile products;

[0046] The method for determining this is as follows: compare the processing energy consumption of aluminum profile products of the same specification and model under all process parameter groups in the experimental scheme, and the one with the lowest energy consumption value is the product with the lowest production energy consumption.

[0047] 3) Use convolutional neural networks or generative adversarial networks to compare the cross-sectional shape of the product to be processed with the cross-sectional shape of the experimental product, find the experimental product that is most similar to the cross-sectional shape of the product to be processed, and calculate the similarity value of the cross-sections of the two products.

[0048] The similarity value is the percentage of the number of identical shape features of the cross-section of the product to be processed and the cross-section of the most similar experimental product to the total number of features of the cross-section of the most similar experimental product (similar to face contour recognition in existing face recognition technology).

[0049] 4) Calculate the ratio of the extrusion coefficient λd of the product to be processed to the extrusion coefficient λb of the most similar experimentally produced product;

[0050] 5) Calculate the predicted production energy consumption of the product to be processed using the production energy consumption value of the most similar aluminum profile product, the similarity value between the product to be processed and its most similar experimental production product, and the extrusion coefficient ratio.

[0051] The formula for calculating the predicted energy consumption of the product to be processed is as follows:

[0052]

[0053] Q y —Predicted energy consumption for the production of products to be processed;

[0054] γ—Similarity value;

[0055] λ d —Extrusion coefficient of the product to be processed;

[0056] λ b —The extrusion coefficient of the product produced in the most similar experiment;

[0057] Qs—Energy consumption value of the product produced in the most similar experiment.

[0058] Example 2: As Figure 2 As shown, an energy consumption prediction system for aluminum profile extrusion production includes,

[0059] The process parameter scheme determination module is used to determine the process parameter schemes for the production of various existing aluminum profile products;

[0060] The minimum production energy consumption determination module is used to conduct experiments based on the parameters determined by the process parameter scheme determination module to determine the minimum production energy consumption of various existing aluminum profile products.

[0061] The similarity determination module is used to determine the similarity value between the cross-sectional shape of the product to be processed and the cross-sectional shape of the experimentally produced product, as determined by deep learning.

[0062] The extrusion coefficient ratio determination module is used to determine the extrusion coefficient λ of the product to be processed. d The extrusion coefficient λ of the product produced in the most similar experiment b The ratio;

[0063] The prediction module predicts the production energy consumption of the product to be processed based on the production energy consumption value of the most similar aluminum profile product, the similarity value between the product to be processed and its most similar experimental production product, and the extrusion coefficient ratio.

Claims

1. A method for predicting energy consumption in aluminum profile extrusion production, characterized in that: The method includes the following steps: 1) Design process parameter schemes for the production of various existing aluminum profile products using orthogonal experimental design; 2) Conduct aluminum profile production experiments based on the process parameter scheme to determine the minimum production energy consumption for various existing aluminum profile products; 3) Deep learning is used to compare the cross-sectional shape of the product to be processed with the cross-sectional shape of the experimental product, find the experimental product that is most similar to the cross-sectional shape of the product to be processed, and calculate the similarity value of the cross-sections of the two products. 4) Calculate the extrusion coefficient λ of the product to be processed. d The extrusion coefficient λ of the product produced in the most similar experiment b The ratio; 5) Calculate the predicted production energy consumption of the product to be processed using the production energy consumption value of the most similar aluminum profile product, the similarity value between the product to be processed and its most similar experimental production product, and the extrusion coefficient ratio. In step 3), the similarity value is the percentage of the number of shape features of the cross-section of the product to be processed that are identical to the cross-section features of the most similar experimental product. The formula for calculating the predicted production energy consumption of the product to be processed in step 5) is as follows: （1） —Predicted energy consumption for the production of products to be processed; —Similarity value; —Extrusion coefficient of the product to be processed; —The extrusion coefficient of the product produced in the most similar experiment; —The energy consumption value of the production product of the most similar experiment.

2. The method for predicting energy consumption in aluminum profile extrusion production according to claim 1, characterized in that: The process parameters in step 1) include the extrusion speed of the extrusion bar, the temperature of the aluminum rod, and the temperature of the die.

3. The method for predicting energy consumption in aluminum profile extrusion production according to claim 2, characterized in that: In step 1), the orthogonal experimental design is a 3-factor, 10-level experimental design.

4. The method for predicting energy consumption in aluminum profile extrusion production according to claim 2, characterized in that: In step 3), the deep learning method is either a convolutional neural network or a generative adversarial network.

5. An energy consumption prediction system for aluminum profile extrusion production, characterized in that: include, The process parameter scheme determination module is used to determine the process parameter schemes for the production of various existing aluminum profile products; The minimum production energy consumption determination module is used to conduct experiments based on the parameters determined by the process parameter scheme determination module to determine the minimum production energy consumption of various existing aluminum profile products. The similarity determination module is used to determine the similarity value between the cross-sectional shape of the product to be processed and the cross-sectional shape of the experimentally produced product, as determined by deep learning. The similarity value is the percentage of the number of shape features of the cross-section of the product to be processed that are identical to the cross-section features of the most similar experimental product. The extrusion coefficient ratio determination module is used to determine the extrusion coefficient of the product to be processed. The extrusion coefficient of the product produced in the most similar experiment The ratio; The prediction module predicts the production energy consumption of the product to be processed based on the production energy consumption value of the most similar aluminum profile product, the similarity value between the product to be processed and its most similar experimental production product, and the extrusion coefficient ratio. The formula for calculating the predicted energy consumption of the product to be processed is as follows: （1） —Predicted energy consumption for the production of products to be processed; —Similarity value; —Extrusion coefficient of the product to be processed; —The extrusion coefficient of the product produced in the most similar experiment; —The energy consumption value of the production product of the most similar experiment.

Images