基于筛选试验的设计因素贡献率分析  

Analysis on Design Factor Contribution Rate Based on Screening Experiment

作　　者：吴扬东[1,2] Wu Yangdong(a. Key Laboratory of Modem Manufacturing Technology （Ministry of Education b. School of Mechanical Engineering, Guizhou University, Guiyang 550003, Chin)

机构地区：[1]贵州大学现代制造技术教育部重点实验室,贵阳550003 [2]贵州大学机械工程学院,贵阳550003

出　　处：《统计与决策》2017年第21期175-178,共4页Statistics & Decision

基　　金：国家自然科学基金资助项目(71161006)

摘　　要：筛选试验常采用二水平部分因子试验,对试验数据的处理采用方差分析的方法,其结论没有定量的指出各因素对结果影响的大小;针对这一问题,文章对因素的离差平方和作进一步分解,结果显示,因素的离差平方和包含该因素所引起的离差平方和以及一部分误差引起的离差平方和两部分内容。在因素的离差平方和中扣除误差引起的离差平方和,其余部分就是该因素引起的纯离差平方和,纯离差平方和在总离差平方和中所占的百分比即为该因素对试验结果的贡献率。最后,以往复泵的设计因素分析为例,对这一方法进行了说明。Two-level fractional factorial experiment is usually used for screening experiment; variance analysis method is of- ten applied to experimental data processing, hut its result does not quantitatively point out the degree to which each factor affects the outcome. To solve this problem, this paper decomposes factor＇ s sum of squares of deviations, and the result shows that total sum of squares of deviations includes two parts： one resulted from the factor and the other from the fractional error; after the sum of squares of deviations resulted from error is deducted from the sum of squares of deviations resulted from the factor, the remaining part is pure sum of squares of deviations resulted from the factor. The percentage of the sum of squares of deviations to the total sum of squares of deviations is the factor＇ s contribution rate to the experimental result. Finally, the paper takes a reciprocating pump design as an example to illustrate this method.

关 键 词：筛选试验 离差平方和分解 产品设计 设计因素分析 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]