期刊
出版社
ELSEVIER SCIENCE SA
DOI: 10.1016/j.msea.2016.02.052
关键词
Artificial neural networks; Genetic algorithms; Monte Carlo simulations; Titanium alloys; Phenomenological equation; Yield strength
类别
资金
- US Air Force Research Laboratory, ISES [FA 8650-08-C-5226]
- NSF [1134873]
- Directorate For Engineering
- Div Of Industrial Innovation & Partnersh [1134873, 1641143] Funding Source: National Science Foundation
A constituent-based phenomenological equation to predict yield strength values from quantified measurements of the microstructure and composition of beta processed Ti-6Al-4V alloy was developed via the integration of artificial neural networks and genetic algorithms. It is shown that the solid solution strengthening contributes the most to the yield strength (similar to 80% of the value), while the intrinsic yield strength of the two phases and microstructure have lower effects (similar to 10% for both terms). Similarities and differences between the proposed equation and the previously established phenomenological equation for the yield strength prediction of the alpha+beta processed Ti-6Al-4V alloys are discussed. While the two equations are very similar in terms of the intrinsic yield strength of the two constituent phases, the solid solution strengthening terms and the 'Hall-Petch'-like effect from the alpha lath, there is a pronounced difference in the role of the basketweave factor in strengthening. Finally, Monte Carlo simulations were applied to the proposed phenomenological equation to determine the effect of measurement uncertainties on the estimated yield strength values. (C) 2016 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据