Page 140 - 机械工程材料2024年第十一期
P. 140
汲高飞,等:基于响应面法和BP神经网络的7050铝合金腐蚀疲劳寿命预测及对比
图 4 BP 神经网络模型训练、验证、测试和整体回归分析结果
Fig. 4 Training (a), validation (b), testing (c) and overall (d) regression analysis results of BP neural network model
参考文献:
[1] 张新,骆宗安,刘照松,等. 7050铝合金复合板界面处
微观组织和力学性能[J]. 中国有色金属学报,2023,
33(11):3558-3568.
ZHANG X,LUO Z A,LIU Z S,et al. Interfacial
microstructure and mechanical properties of 7050
aluminum alloy clad plates[J]. The Chinese Journal of
Nonferrous Metals,2023,33(11):3558-3568.
图 5 BP 神经网络模型训练、验证和测试的均方误差
[2] 臧金鑫,陈军洲,韩凯,等. 航空铝合金研究进展与发
随迭代次数的变化曲线
Fig. 5 Curves of mean square error vs iteration ordinal number of 展趋势[J]. 中国材料进展,2022,41(10):769-777.
BP neural network model during training, validation and testing ZANG J X,CHEN J Z,HAN K,et al. Research
progress and development tendency of aeronautical
表4 BP神经网络模型的对数疲劳寿命预测结果
aluminum alloys[J]. Materials China,2022,41(10):
与试验结果的对比
Table 4 Comparison of prediction results by BP neutral 769-777.
network model and test results of logarithmic fatigue life [3] ZHANG T Y,ZHANG T,HE Y T,et al. Corrosion
and aging of organic aviation coatings:A review[J].
对数疲劳寿命
样本序号 相对误差/% Chinese Journal of Aeronautics,2023,36(4):1-35.
试验值 预测值 [4] 高冲,董丽虹,王海斗,等. 航天铝合金变极性等离子
1 5.060 8 4.932 5 2.535 2 弧焊焊接接头的腐蚀与疲劳交替研究[J]. 材料导报,
2 4.876 5 4.892 0 0.317 9 2024,38(12):145-151.
3 5.061 1 5.037 0 0.476 2 GAO C,DONG L H,WANG H D,et al. Study on
corrosion and fatigue alternation of aerospace aluminum
4 4.760 4 4.847 7 1.833 9
alloy welded joints by variable polarity plasma arc
5 5.315 1 5.246 0 1.300 1
welding[J]. Materials Reports,2024,38(12):145-151.
6 5.185 4 5.164 3 0.406 9
[5] 罗来正,周堃,周洁,等. 海洋大气环境与拉伸疲劳载
7 5.050 4 5.086 5 0.714 8
荷协同作用下不同加载方式对7050铝合金腐蚀损伤特
8 4.617 0 4.694 0 1.667 7
性的影响[J]. 表面技术,2023,52(11):291-299.
132

