材料工程实验设计及数据处理方法的优选研究

发布时间：2018-04-03 06:49

本文选题：材料基因工程　切入点：实验设计方法　出处：《北方工业大学》2017年硕士论文

【摘要】：本文面向材料工程领域,采用Visual Basic(VB)编程,建立实验设计方法优选软件系统,从而实现以特定目标为导向的实验设计方法筛选,并通过全因子析因法的实验结果与软件所选设计方法的实验结果进行对比,验证该软件的有效性;本文还采取不同的数据处理方式对材料领域的相关典型案例进行数据挖掘分析。本文工作旨在为材料领域一线的工程技术人员开发出一种便捷、实用的实验设计方法以及数据分析方法优选工具。本文开发的实验设计方法优选软件涵盖了目前最常用的三种实验设计方法,分别是完全析因设计、正交实验设计以及均匀化设计,用户可以根据实验目的的具体要求及评价范围,通过软件选择出适用的实验设计方法。本文分别采用完全析因设计和软件推荐的实验设计方法设计SiC/Al复合材料的应力应变实验,并利用SolidWorks建模和有限元分析软件进行模拟计算,将模拟结果进行对比,初步证实了实验设计优选软件的有效性。在数据处理方法的对比与研究方面,通过实例分析,得出以下结论:回归分析在解决线性问题上表现优异,可以较为准确的表达出实验因子与结果数据的关联性;灰色关联法可以将实验因子与目标因子之间的关联顺序进行排序,反映出不同实验因子对目标因子的影响程度;人工神经网络对于非线性的分析有较强的处理能力,并且非常适合拥有大量实验数据的高通量实验,将成为材料基因工程中最重要的计算预测工具之一。
[Abstract]:In this paper, aiming at the field of material engineering, a software system for the optimal selection of experimental design method is established by using Visual basic programming, which realizes the selection of experimental design methods based on specific objectives.The validity of the software is verified by comparing the experimental results of the full-factor factoring method with the experimental results of the design method selected by the software.This paper also uses different data processing methods to analyze the data mining of typical cases in the field of materials.The purpose of this paper is to develop a convenient and practical experimental design method and data analysis tool for engineering technicians in the field of materials.The experimental design method optimization software developed in this paper covers the three most commonly used experimental design methods, namely, complete factoring design, orthogonal experimental design and homogenization design.According to the specific requirements and evaluation scope of the experiment purpose, the user can choose the suitable experimental design method through the software.In this paper, the stress-strain experiments of SiC/Al composites are designed by means of complete factorial design and experimental design recommended by software, and the simulation results are compared by using SolidWorks modeling and finite element analysis software.The effectiveness of the optimal selection software is preliminarily verified.In the data processing method contrast and research, through the example analysis, draws the following conclusion: regression analysis in solving the linear problem performance is excellent, can more accurately express the experimental factor and the result data correlation;Grey correlation method can sort the correlation order between experimental factors and target factors, which reflects the influence of different experimental factors on target factors, and artificial neural network has a strong ability to deal with nonlinear analysis.It is very suitable for high-throughput experiments with a large amount of experimental data and will become one of the most important computational and predictive tools in material genetic engineering.
【学位授予单位】：北方工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TB302

【参考文献】