Calibration and adjustment of mechanical property prediction model for poly(vinyl alcohol)-enhanced carbon nanotube buckypaper manufacturing

详细信息    查看全文

• 作者：Kan Wang ; Arda Vanli ; Chuck Zhang ; Ben Wang
• 关键词：Nanostructures ; Mechanical properties ; Finite element analysis ; Bayesian additive model ; Buckypaper
• 刊名：The International Journal of Advanced Manufacturing Technology
• 出版年：2017
• 出版时间：February 2017
• 年：2017
• 卷：88
• 期：5-8
• 页码：1889-1901
• 全文大小：
• 刊物类别：Engineering
• 刊物主题：Industrial and Production Engineering; Media Management; Mechanical Engineering; Computer-Aided Engineering (CAD, CAE) and Design;
• 出版者：Springer London
• ISSN：1433-3015
• 卷排序：88

文摘

Buckypaper is a free-standing carbon nanotube (CNT) sheet used to improve handling and manufacturability of CNT-based nanocomposites. To enhance the mechanical properties and manufacturing efficiency of buckypaper, polymeric binders, such as poly(vinyl alcohol) (PVA), are often added into the CNT network during the manufacturing process. This paper describes a physics-based model to predict the elastic modulus of PVA-enhanced buckypaper and a statistical approach to calibrate and adjust the model based on physical experiments. Compared to the physics-based model alone, the hybrid model can provide more accurate predictions for the Young’s modulus of PVA-enhanced buckypaper and give a 95 % confidence interval on the prediction. One of the inputs for this model, the average length of carbon nanotubes, was calibrated using maximum likelihood estimation (MLE). The bias of this model was adjusted by estimating a bias function. Both the calibration parameter and adjustment function were estimated from a set of experimental measurements. The improvement in making a prediction was validated by comparing the performance of the physics-based model, statistical model, and statistics-enhanced physics model at a new experiment point. The hybrid model provides a more accurate prediction than either the physics-based model or statistical model does. This model calibration technique provides an effective tool for nanomanufacturing process design and material property prediction.