Artificial Neural Network Modeling and Mechanism Study for Relaxation of Deformed Rubber

详细信息    查看全文

• 作者：Xiu-Juan Wang ; Xiu-Ying Zhao ; Qiang-Guo Li ; Tung W. Chan ; Si-Zhu Wu
• 刊名：Industrial & Engineering Chemistry Research
• 出版年：2016
• 出版时间：April 13, 2016
• 年：2016
• 卷：55
• 期：14
• 页码：4059-4070
• 全文大小：695K
• 年卷期：0
• ISSN：1520-5045

文摘

An artificial neural network (ANN) was developed to estimate the relaxation property of diene rubber. Regularization was introduced into the ANN and the average prediction accuracy was 98.72%. The sensitivity analysis shows that compressive strain is the crucial influence on the relaxation property. Diene rubber shows a higher relaxation at low compressive strain than at high compressive strain. Molecular simulations show that rubber at low compressive strain possesses high fractional free volume, molecular chain movement, and ozone permeability. The chemical characterization of cross-link density and 2D-FTIR correlation analysis show that the rubber network at low compressive strain is seriously degraded by random scission and the generation rate of the carbonyl group is faster than that of ozonide group, indicating chain scission predominates in ozonation. These fundamental studies are expected to provide a comprehensive understanding of the relaxation property of deformed rubber and guidance for the design of antiaging materials.