Premium
A Bayesian approach for integration of physical and computer experiments for quality improvement in nano‐composite manufacturing
Author(s) -
Vanli O. Arda,
Zhang Chuck,
Chen LiJen,
Wang Kan Kevin,
Wang Ben
Publication year - 2010
Publication title -
quality and reliability engineering international
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.913
H-Index - 62
eISSN - 1099-1638
pISSN - 0748-8017
DOI - 10.1002/qre.1130
Subject(s) - computer science , bayesian probability , process (computing) , quality (philosophy) , experimental data , composite number , data mining , machine learning , reliability engineering , industrial engineering , artificial intelligence , algorithm , engineering , mathematics , philosophy , statistics , epistemology , operating system
This paper presents a new Bayesian predictive approach for quality improvement in nano‐composite manufacturing that aims to improve the accuracy of computer model simulations by combining physical process data and computer model predictions. Such a data augmentation method is of crucial importance for capital‐intensive nano‐composite processes for which large experimental studies are usually very costly or impractical. A simulation example and a case study on a real vacuum‐assisted resin transfer molding (VARTM) composite manufacturing process are used to illustrate the proposed approach. The results of these studies show that the proposed approach can achieve more accurate predictions of the process variability than pure empirically based or pure simulation‐based methods for a given experimental data size. An important practical implication of the results of this work is that, by using the proposed Bayesian method that leverages computational methods with the experimental data, one can accomplish significant reduction in product development costs by reducing the number of physical experiments and improving the prediction accuracy of computer models. Copyright © 2010 John Wiley & Sons, Ltd.