Font Size: a A A

Nano-materials Preparationprocess Modeling Based On Physical And Computer Experiments

Posted on:2015-05-23Degree:MasterType:Thesis
Country:ChinaCandidate:J M LouFull Text:PDF
GTID:2181330452969636Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
Nanomaterial hasgreat significance in many applications, although so far most ofthe preparation of nanomaterial is still unable to realize scale, low cost of manufacturingbecause of the lack of basis of the process modeling.Basic modeling method is based on physical experiment and statistical DOEmethod. Set CVD process of new materialgrapheme film as an example. Coppernetwork is treated as substrates andCH4as carbon source with transmissivity andresistivity as response variables. Pre-experiment is designed to narrow the extent oftemperature and flux. Experiment I is22full factorial design in order to get a knowledgeof the relationship between the factors. Then, experiment II (a response surfacedesign)is designed to refine the statistical model. Two separated statistical model basedon DOE data are built up including the analysis of relationship between the tworesponse.Because physical experiment of nanomaterial has high cost and low replication,computer experiment based on theoretical modelcan be useful in nano-area.Wellaccepted Kriging surrogate model is used as predicting method. In this paper,considering the deviation between physical and computer experiment, two kinds ofcorrection models are built up. First, when two kinds of samples are consistent,thecomputer results are treated as an input parameter. Second,enlarge the computerexperimental samples, and separatethe computer resultand the difference of the two intotwo parts to make modeling more accurate.Under the background ofgraphene growth, the dynamic numerical solution ofdifferential equations is the computer experiment. Kriging model is established, withtemperature, CH4and hydrogen pressure as control variables, the amount of methanecarbon atoms, adhesion molecules, coverage,and the time needed for equilibrium asresponse variables. The relationship between them is described.But predicted results arenot entirely consistent with the experimental results.With the actual physical experimentresult, two methods of correction model are applied.Through the interactive testverification, both of the two correction model perform well. And the second modeladding more computer experiments can obtain more accurate prediction.
Keywords/Search Tags:nanomaterial, graphene, DOE, Computer experiment, Kriging
PDF Full Text Request
Related items