目的 利用复杂截面薄壁铝型材挤压的数值模拟数据集和RBMO-BP神经网络,构建挤压能耗和温度均方差的预测模型,为铝型材挤压多目标优化与结果决策提供应用依据。方法 首先,构建1 mm壁厚复杂截面铝型材挤压有限元模型并进行数值模拟,利用仿真结果构建挤压过程数据集;其次,利用数据集对构建的RBMO-BP神经网络进行训练并预测;再次,建立铝材挤压成形多目标优化模型,利用RBMO-BP-MOPSO进行求解;最后,利用熵权法-灰色关联分析法对铝材挤压成形多目标优化结果进行决策。结果 与BP神经网络相比,MAE、MSE、RMSE、MAPE等指标RBMO-BP在挤压能耗预测方面分别下降了19.91 kJ、679.55 kJ2、23.73 kJ、1.61%,在出口温度均方差预测方面分别下降了1.83 ℃、5.09 ℃2、1.84 ℃、23.45%。利用熵权法-灰色关联分析法对RBMO-BP-MOPSO优化得出的Pareto前沿解进行决策,获得了最佳挤压工艺参数,该工艺参数相比于优化前的工艺参数,挤压能耗减少了14.22%,出口温度均方差减小了37.28%。结论 本研究通过数值模拟、数据预测、多目标优化方法的结合,精准预测了复杂截面薄壁铝型材挤压的能耗与温度均方差,优化了挤压工艺参数并且得以验证。结果表明,本文提出的复杂截面薄壁铝型材挤压多目标优化方法具有可行性和先进性。
Abstract
The work aims to employ numerical simulation dataset and RBMO-BP neural network of extrusion processes for complex cross-sectional thin-walled aluminum profiles to develop predictive models for both extrusion energy consumption and temperature variance, thereby providing a practical foundation for multi-objective optimization and decision-making in aluminum extrusion. Firstly, a finite element model of a complex cross-sectional aluminum profile with 1 mm wall thickness was constructed and numerically simulated to generate a dataset of the extrusion process. Secondly, this dataset was used to train and predict outcomes with the proposed RBMO-BP neural network. Thirdly, a multi-objective optimization model for aluminum extrusion forming was established and solved with the RBMO-BP-MOPSO algorithm. Finally, an integrated entropy weight method and grey relational analysis were applied to make decisions among the obtained Pareto-optimal solutions. Compared with the conventional BP neural network, the RBMO-BP model achieved significant improvements in prediction accuracy. For extrusion energy consumption, the MAE, MSE, RMSE, and MAPE decreased by 19.91 kJ, 679.55 kJ2, 23.73 kJ, and 1.61%, respectively. For exit temperature mean square deviation, these metrics decreased by 1.83 ℃, 5.09 ℃2, 1.84 ℃, and 23.45%, respectively. Bu using entropy weight method and grey relational analysis to select the optimal solution from the Pareto front generated by RBMO-BP-MOPSO, the best extrusion parameters were identified. Under these optimized parameters, extrusion energy consumption was reduced by 14.22% and the exit temperature mean square deviation decreased by 37.28% compared to the original process parameters. By integrating numerical simulation, data-driven prediction, and multi-objective optimization methods, this work accurately predicts both energy consumption and temperature uniformity in the extrusion of complex cross-sectional thin-walled aluminum profiles, successfully optimizes the extrusion process parameters, and validates their effectiveness. The results demonstrate that the proposed multi-objective optimization approach for complex thin-walled aluminum extrusion is both feasible and advanced.
关键词
薄壁铝型材 /
挤压成形 /
工艺参数 /
RBMO-BP-MOPSO /
能耗多目标优化
Key words
thin-walled aluminum profiles /
extrusion forming /
process parameters /
RBMO-BP-MOPSO /
multi-objective energy consumption optimization
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 葛红林. 再创铝工业发展的全球比较优势[N]. 中国有色金属报, 2024-10-29.
GE H L. Re-Creating the Global Comparative Advantage of Aluminum Industry Development[N]. China Nonferrous Metals News, 2024-10-29.
[2] FERRÁS A F, DE Almeida F, Silva E C E, et al. Sustainable and Optimized Production in an Aluminum Extrusion Process[J]. Sustainability, 2025, 17(9): 4179.
[3] ALAFAGHANI A, PULEO R, ADAMS L, et al.A Study on Internal Quenching of Hollow Extrusions to Reduce Distortion and Increase the Energy to Failure of Aluminum Profiles[J]. International Journal of Material Forming, 2025, 18(1): 19.
[4] 刘禹江, 杜文玉, 王延, 等. 铜包铝层状材料的制备技术、应用现状及发展趋势研究[J]. 精密成形工程, 2025, 17(7): 119-137.
LIU Y J, DU W Y, WANG Y, et al.Preparation Technology, Application Status, and Development Trends of Copper-Clad Aluminum Laminated Materials[J]. Journal of Netshape Forming Engineering, 2025, 17(7): 119-137.
[5] 张琪, 王伟东. 2024年中国铝材及制品进出口贸易分析[J]. 资源再生, 2025(3): 14-18.
ZHANG Q, WANG W D. Analysis of China's Aluminum Materials and Products Import and Export Trade in2024[J]. Resource Recycling, 2025(03): 14-18.
[6] YANG L Q, WANG H Y, SHAN Q C, et al.Optimization of the Hot Extrusion Process for 6082 Aluminum Alloy W-Profile[J]. Journal of Materials Engineering and Performance, 2025, 34(22): 26625-26637.
[7] JIN S X, LI W F, XU X, et al.Extremely Improved the Yield Strength of 6061 Aluminum Alloy by Melt Spinning and Hot Extrusion[J]. Journal of Alloys and Compounds, 2025, 1020: 179599.
[8] HUANG Y C, GAO X B, LIU Y, et al.Investigation of an L-Shaped Cross-Sectioned AA6061 Aluminum Alloy Ring via Extrusion and Self-Bending Integral Processing Technology[J]. JOM, 2025, 77(3): 1133-1147.
[9] RAHIM S N A, LAJIS M A, ARIFFIN S. Effect of Extrusion Speed and Temperature on Hot Extrusion Process of 6061 Aluminum Alloy Chip[J]. ARPN Journal of Engineering and Applied Sciences, 2016, 11(4): 2272-2277.
[10] 王尧, 周照耀, 潘健怡, 等. 基于ALE有限元法的铝型材挤压成形的数值模拟[J]. 锻压技术, 2010, 35(1): 149-153.
WANG Y, ZHOU Z Y, PAN J Y, et al.Numerical Simulation on Aluminum Profile Extrusion Based on ALE Method[J]. Forging & Stamping Technology, 2010, 35(1): 149-153.
[11] FARJAD BASTANI A, AUKRUST T, BRANDAL S.Study of Isothermal Extrusion of Aluminum Using Finite Element Simulations[J]. International Journal of Material Forming, 2010, 3(1): 367-370.
[12] 周晓远, 陈文琳, 潘鹏林, 等. 空心薄壁铝型材挤压数值模拟与模具优化设计[J]. 模具工业, 2018, 44(12): 58-65.
ZHOU X Y, CHEN W L, PAN P L, et al.Numerical Simulation and the Die Optimization Design of Hollow Thin-Wall Aluminum Extrusion[J]. Die & Mould Industry, 2018, 44(12): 58-65.
[13] 曾文浩, 魏刚, 邓小亮, 等. 翻窗上扇框薄壁空心铝型材挤压过程数值模拟与挤压参数优化[J]. 西华大学学报(自然科学版), 2017, 36(4): 70-77.
ZENG W H, WEI G, DENG X L, et al.Numerical Simulation and Extrusion Parameters Optimization for an Upper Window-Fan Frame Aluminum Profile with Thin-Walled Hollow Section[J]. Journal of Xihua University (Natural Science Edition), 2017, 36(4): 70-77.
[14] 阮祥明, 吉宏选, 陈文琳. 薄壁空心铝型材挤压过程数值模拟及模具结构改进[J]. 模具工业, 2017, 43(7): 55-58.
RUAN X M, JI H X, CHEN W L.Numerical Simulation of Extrusion Process for Thin-Walled Hollow Aluminum Profile and the Die Structure Improvement[J]. Die & Mould Industry, 2017, 43(7): 55-58.
[15] 刘国勇, 高士泽, 朱冬梅. 轨道用中空薄壁大小幅铝型材的挤压规律分析[J]. 华南理工大学学报(自然科学版), 2025, 53(5): 45-55.
LIU G Y, GAO S Z, ZHU D M.Analysis of Extrusion Law of Large-Scale and Small-Scale Aluminum Profiles with Hollow Thin Wall for Rails[J]. Journal of South China University of Technology (Natural Science Edition), 2025, 53(5): 45-55.
[16] 陈飞, 魏科, 党利, 等. 薄壁矩形管铝型材分流挤压过程挤压力影响研究[J]. 锻压技术, 2023, 48(12): 129-137.
CHEN F, WEI K, DANG L, et al.Study on Influence of Extrusion Force during Split Extrusion Process for Thin-Walled Rectangular Tube Aluminum Profile[J]. Forging & Stamping Technology, 2023, 48(12): 129-137.
[17] 陈浩, 赵国群, 张存生, 等. 薄壁空心铝型材挤压过程数值模拟及模具优化[J]. 机械工程学报, 2010, 46(24): 34-39.
CHEN H, ZHAO G Q, ZHANG C S, et al.Numerical Simulation of Extrusion Process and Die Structure Optimization for a Hollow Aluminum Profile with Thin Wall[J]. Journal of Mechanical Engineering, 2010, 46(24): 34-39.
[18] 刘鹏程, 彭炳锋, 刘寒龙, 等. 基于神经网络的6063铝型材挤压工艺多目标优化[J]. 中南大学学报(自然科学版), 2025, 56(3): 881-890.
LIU P C, PENG B F, LIU H L, et al.Multi-objective Optimization of Extrusion Process for 6063 Aluminum Profile Based on Neural Network[J]. Journal of Central South University (Science and Technology), 2025, 56(3): 881-890.
[19] 张明杰, 杨柳, 肖云. 基于神经网络的铝型材挤压过程能耗工艺参数优化研究[J]. 装备制造技术, 2018(6): 258-259.
ZHANG M J, YANG L, XIAO Y.Optimization of Extrusion Energy Consumption Parameters of Aluminum Profiles Based on Neural Network[J]. Equipment Manufacturing Technology, 2018(6): 258-259.
[20] 莫建虎, 李落星, 周佳, 等. 采用BP神经网络和挤压形状因子预测挤压力与挤压出口温度[J]. 矿冶工程, 2009, 29(02): 90-94.
MO J H, LI L X, ZHOU J, et al.Application of BP Neural Network and Shape Factor in Predicting Extrusion Force and Exit Temperature[J]. Mining and Metallurgical Engineering, 2009, 29(2): 90-94.
[21] 杨海东, 江海昌, 方华, 等. 基于GA-SVR的挤压机能耗异常检测模型研究[J]. 机床与液压, 2019, 47(05): 163-168.
YANG H D, JIANG H C, FANG H, et al.Study on Abnormal Energy Consumption Model of Extruder Based on GA-SVR[J]. Machine Tool & Hydraulics, 2019, 47(05): 163-168.
[22] FU S, LI K, HUANG H, et al.Red-billed Blue Magpie Optimizer: a Novel Metaheuristic Algorithm for 2D/3D UAV Path Planning and Engineering Design Problems[J]. Artificial Intelligence Review, 2024, 57(6): 134.
[23] LI S, KOU L.An Enhanced Red-Billed Blue Magpie Optimizer Based on Superior Data Driven for Numerical Optimization Problems[J]. Biomimetics, 2025, 10(11): 780.
[24] 鲁钧豪, 孙先锋, 王致桦, 等. 响应面法与GA-BP神经网络联合优化细菌降解石油烃参数研究[J]. 现代化工, 2025, 45(4): 102-109.
LU J H, SUN X F, WANG Z H, et al.Parameter Optimization for Bacterial Degradation of Petroleum Hydrocarbons by Response Surface Methodology and GA- BP Neural Network Jointly[J]. Modern Chemical Industry, 2025, 45(4): 102-109.
[25] 兰智, 吴江江. 基于改进粒子群优化BP神经网络的短期光伏发电功率预测[J]. 科技与创新, 2025(4): 5-8.
LAN Z, WU J J.Short-Term Photovoltaic Power Prediction Based on Improved PSO-BP Neural Network[J]. Science and Technology & Innovation, 2025(4): 5-8.
[26] SEXTON R S, GUPTA J N D. Comparative Evaluation of Genetic Algorithm and Backpropagation for training Neural Networks[J]. Information Sciences, 2000, 129(1/2/3/4): 45-59.
[27] 徐敏, 林卫星, 石磊, 等. 基于I-GWO-BP神经网络的矿区爆破振动预测[J]. 矿业研究与开发, 2025, 45(10): 121-128.
XU M, LIN W X, SHI L, et al.Prediction of Mining Area Blasting Vibration Based on I-GWO-BP Neural Network[J]. Mining Research and Development, 2025, 45(10): 121-128.
[28] 方华. 基于粒子群算法的等温挤压能耗优化[D]. 广州: 广东工业大学, 2019.
FANG H.Optimization of Isothermal Extrusion Energy Consumption Based on PSO[D]. Guangzhou: Guangdong University of Technology, 2019.
[29] 姜兴宇, 刘傲, 杨国哲, 等. 激光增材制造过程低碳建模与工艺参数优化[J]. 机械工程学报, 2022, 58(5): 223-238.
JIANG X Y, LIU A, YANG G Z, et al.Low-Carbon Modeling and Process Parameter Optimization in Laser Additive Manufacturing Process[J]. Journal of Mechanical Engineering, 2022, 58(5): 223-238.
基金
贵州省科技支撑项目（黔科合支撑[2023]一般302）
PDF(2968 KB)
渝公网安备 50010702501719号 渝ICP备15012534号-4
