锻压技术

基于PSO-BP模型的5083铝合金力学性能预测

英文标题：Prediction of mechanical properties for 5083 aluminum alloy based on PSO-BP model

作者：崔鑫张建平张能辉

分类号：TG146.2; TP183

出版年,卷(期):页码：2019,44(6):183-187

摘要：

为准确预测5083铝合金热力学参数与其流变应力之间的关系，基于PSO优化BP神经网络方法提出了PSO-BP铝合金性能预测模型，并对BP模型和PSO-BP模型预测结果进行了对比与分析。结果表明：PSO-BP模型预测值与试验值的吻合度高于BP模型，更能准确地反映铝合金在不同工艺条件下流变应力的变化规律；PSO-BP预测模型具有更快的收敛速度，达到BP预测模型的10倍以上；与传统BP模型相比，PSO-BP模型预测值的平均相对误差不到BP模型的50%，在低温和高温时更明显，且4种应变速率下其预测值与试验值的线性回归决定系数更接近于1，证明了PSO-BP模型对5083铝合金力学性能具有更高的预测精度。

In order to accurately predict the relationship between thermodynamic parameters and rheological stress of 5083 aluminium alloy, based on the method of BP neural network optimized by PSO, the performance prediction model PSO-BP of aluminum alloy was proposed, and the prediction results of BP model and PSO-BP model were compared and analyzed. The results indicate that the coincidence degree between the predicted value of PSO-BP model and the experimental value is higher than that of BP model, and the PSO-BP model can more accurately reflect the rheological stress change of aluminium alloy under different process conditions. Furthermore, the PSO-BP prediction model has the faster convergence speed, which is more than 10 times that of the BP prediction model. Compared with the traditional BP model, the average relative error of the predicted value of PSO-BP model is less than 50% that of BP model, it is more obvious at low temperature and high temperature, and its linear regression determinant coefficients between predicted value and experimental value under four strain rates are closer to 1. Thus, it is proved that the PSO-BP model has higher prediction accuracy for mechanical properties of 5083 aluminum alloy.

基金项目：

国家自然科学基金资助项目（11572187）；上海市科学技术委员会项目（18DZ1202105, 18DZ1202302）

崔鑫（1993-），男，硕士研究生 E-mail：cxin777@163.com 通讯作者：张建平（1972-），男，博士后，教授 E-mail：jpzhanglzu@163.com

参考文献：

［1］杨少华, 张丹城, 田亚斌, 等. 5083铝镁合金在不同浓度NaCl溶液中的腐蚀研究
［J］. 有色金属科学与工程, 2018, 9(2): 1-5.

Yang S H, Zhang D C, Tian Y B, et al. Corrosion of 5083 aluminum magnesium alloy in NaCl solution with different concentrations
［J］. Nonferrous Metals Science and Engineering, 2018, 9(2): 1-5.

［2］陈雨, 杨文静, 高兵, 等. 搅拌摩擦加工5083-O铝合金的组织和性能
［J］. 稀有金属, 2017, 41(9): 955-963.

Chen Y, Yang W J, Gao B, et al. Friction stir processing in 5083-O aluminum alloy: Microstructures and mechanical properties
［J］. Chinese Journal of Rare Metals, 2017, 41(9): 955-963.

［3］Mirjavadi S S, Alipour M, Emamian S, et al. Influence of TiO2 nanoparticles incorporation to friction stir welded 5083 aluminum alloy on the microstructure, mechanical properties and wear resistance
［J］. Journal of Alloys and Compounds, 2017, 712: 795-803.

［4］杨如民. 基于人工神经网络的铝合金宽幅板带材质量控制模型的研究
［D］. 重庆：重庆大学, 2015.

Yang R M. The Research of the Quality Control Model for Wild Aluminum Plate and Strip Based on Artificial Neural Network
［D］.Chongqing：Chongqing University, 2015.

［5］陈宝东, 郭锋, 温静, 等. Mg-Zn-Zr-Y合金高温塑性变形本构模型及流变行为预测
［J］. 稀有金属材料与工程, 2017, 46(11): 3305-3310.

Chen B D, Guo F, Wen J, et al. Constitutive model of hot plastic deformation and flow behavior prediction of Mg-Zn-Zr-Y alloy
［J］. Rare Metal Materials and Engineering, 2017, 46(11): 3305-3310.

［6］潘玲, 张强. 基于BP神经网络的汽车轻合金零部件热处理工艺优化
［J］. 热加工工艺, 2017, 46(18): 221-223，227.

Pan L, Zhang Q. Optimization of heat treatment process for automotive light alloy parts based on BP neural network
［J］. Hot Working Technology, 2017, 46(18): 221-223，227.

［7］李先梦, 湛利华, 申儒林, 等. 2A12硬铝合金热拉伸流变行为及本构建模
［J］. 锻压技术, 2017, 42(4): 159-164，193.

Li X M, Zhan L H, Shen R L, et al. Hot tensile flow behavior and constitutive model of aluminum alloy 2A12
［J］. Forging & Stamping Technology, 2017, 42(4): 159-164，193.

［8］初冠南, 林艳丽, 宋伟宁, 等. 基于二次多项式新本构模型的铝合金搅拌摩擦焊板材成形极限研究
［J］.金属学报, 2017, 53(1): 114-122.

Chu G N, Lin Y L, Song W N, et al. Forming limit of FSW aluminum alloy blank based on a new constitutive model
［J］. Acta Metallurgica Sinica, 2017, 53(1): 114-122.

［9］马立勇, 颜景润, 张永清, 等. 基于灰色系统理论的7075铝合金板材疲劳寿命研究
［J］. 锻压技术, 2017, 42(7): 173-176.

Ma L Y, Yan J R, Zhang Y Q, et al. Study on fatigue life of aluminum alloy 7075 sheet based on grey system theory
［J］. Forging & Stamping Technology, 2017, 42(7): 173-176.

［10］罗豪鑫, 陈传勇, 刘建中, 等. 基于遗传规划算法的不同应力比下不同厚度7050铝合金疲劳裂纹扩展寿命预测
［J］. 材料科学与工程学报, 2017, 35(1): 26-31.

Luo H X, Chen C Y, Liu J Z, et al. Prediction of fatigue crack growth life of 7050 aluminum alloy with different thicknesses under various load ratios by genetic programming
［J］. Journal of Materials Science and Engineering, 2017, 35(1): 26-31.

［11］李尚平, 陈曾雄, 周敬辉, 等. BP神经网络对甘蔗宿根切割质量的预测——基于PSO算法
［J］. 农机化研究, 2018, 40(10): 11-17.

Li S P, Chen Z X, Zhou J H, et al. Prediction model of cutting quality of the sugarcane ratoon by BP neural network-Based on PSO algorithem
［J］. Journal of Agricultural Mechanization Research, 2018, 40(10): 11-17.

［12］潘强, 张继春, 肖清华, 等. 动能弹对混凝土靶侵彻深度的PSO-SVM预测
［J］. 高压物理学报, 2018, 32(2): 108-115.

Pan Q, Zhang J C, Xiao Q H, et al. Prediction of penetration depth of projectiles into concrete targets based on PSO-SVM
［J］. Chinese Journal of High Pressure Physics, 2018, 32(2): 108-115.

服务与反馈：

【文章下载】【加入收藏】