蠕动
投影(关系代数)
一致性(知识库)
残余物
材料科学
过程(计算)
残余应力
计算机科学
算法
人工智能
复合材料
操作系统
作者
Lei Cheng,Qing Guo,Wei Yu,Ying Han,Qingwu Cai
标识
DOI:10.1002/srin.202200270
摘要
The phenomenological characteristics of θ projection method constrain its prediction accuracy, especially at the primary creep stage. Herein, the mathematical features of the classical θ projection method and its modified forms based on a group of creep curves of a newly designed low‐cost heat‐resistant steel are analyzed. A detailed data processing method is introduced, and the prediction accuracy is comparatively studied with respect to the relative error on different creep stages. Results reveal that the simplified 3‐parameter method brings a balanced improvement on the residual errors by sacrificing the prediction accuracy on primary creep. However, the 5‐parameter method can precisely predict the characteristics of primary creep, and the tolerance fitting process can further reduce the deviation at the tertiary creep stage. Furthermore, the 5‐parameter method shows a larger reliable prediction region than the 3‐parameter method, and the two methods present good consistency when the applied stress is larger than 134 MPa. Herein, theoretical guidance for the selection of different θ projection methods is provided.
科研通智能强力驱动
Strongly Powered by AbleSci AI