蠕动
本构方程
材料科学
过度拟合
航程(航空)
石墨
压力(语言学)
功能(生物学)
结构工程
复合材料
计算机科学
工程类
有限元法
生物
语言学
哲学
机器学习
进化生物学
人工神经网络
作者
Guoxi Jing,Tian Ma,Teng Ma,Shubo Li,Guang Chen,Xiuxiu Sun,Shuai Sun
摘要
Abstract In order to accurately predict the minimum creep rate under the actual working conditions of compacted graphite cast iron cylinder head, this paper conducted creep experiments in the temperature range of 723.15–823.15 K and the stress range of 100–300 MPa. The creep damage evolution and damage mechanisms under test conditions were also analyzed. Then the data fitting performances were analyzed and compared from three aspects: the optimization‐seeking algorithm, the form of the objective function and the construction of the creep constitutive model. Finally, the chosen optimization algorithm combined with the proposed objective function and creep constitutive model resulted in 100% of the predicted values within the 3 times error band and 26.67% of the predicted values within the 1.5 times error band. This improves the prediction accuracy while reducing the effect of data overfitting.
科研通智能强力驱动
Strongly Powered by AbleSci AI