Computer simulation driven statistical modeling and quality control.

摘要

Under the intense competition in today's market, manufacturing enterprises must continuously strive to bring innovative designs into their products and improve their production processes. On one hand, manufacturing systems become more flexible and complex, which poses significant challenges for system modeling, monitoring and control. On the other hand, the rapid development of information technology, advanced simulation tools, and computing power enable powerful simulation based techniques for process design, monitoring and control.;The objective of this research is to develop innovative methodologies to achieve variation reduction and productivity improvement for manufacturing enterprises through integrated design and in-line statistical monitoring and control with the aid of computer simulation. Specifically, two research topics have been discussed: (i) novel statistical process control tools for emerging micro/nano manufacturing processes; (ii) modeling of complex engineering systems through computer simulation. In the first topic, we focused on characterization and monitoring of spatial point distribution in micro/nano manufacturing processes. In the second topic, a novel multivariate Gaussian process based surrogate modeling method that can incorporate both quantitative input variables and qualitative input variables is developed. This method provides an efficient way for emulating multiple input and multiple output computer simulation models while improving model prediction accuracy.;The effectiveness of our proposed methods has been demonstrated through computer simulations and case studies. Given the ubiquitous trend of data abundance and computer technology advancement in other areas, the proposed methods have the potential to be extended to fields other than manufacturing.