【摘要】：篩選試驗常采用二水平部分因子試驗,對試驗數(shù)據(jù)的處理采用方差分析的方法,其結(jié)論沒有定量的指出各因素對結(jié)果影響的大小;針對這一問題,文章對因素的離差平方和作進一步分解,結(jié)果顯示,因素的離差平方和包含該因素所引起的離差平方和以及一部分誤差引起的離差平方和兩部分內(nèi)容。在因素的離差平方和中扣除誤差引起的離差平方和,其余部分就是該因素引起的純離差平方和,純離差平方和在總離差平方和中所占的百分比即為該因素對試驗結(jié)果的貢獻率。最后,以往復泵的設(shè)計因素分析為例,對這一方法進行了說明。
[Abstract]:Two level partial factor test is often used in screening test, and ANOVA is used to deal with the test data. Its conclusion does not indicate quantitatively the influence of each factor on the result. In this paper, the sum of deviation squared of factors is further decomposed. The result shows that the sum of square of deviation of factors includes the sum of square of deviation caused by this factor and the sum of square of deviation caused by some errors. The sum of square of deviation caused by error is deducted from the sum of square of deviation of factor, and the other part is the sum of square of net deviation caused by this factor. The percentage of sum of square of net deviation in total sum of square of deviation is the contribution rate of this factor to the result of experiment. Finally, the design factor analysis of reciprocating pump is taken as an example to illustrate this method.
【作者單位】： 貴州大學現(xiàn)代制造技術(shù)教育部重點實驗室;貴州大學機械工程學院;
【基金】：國家自然科學基金資助項目(71161006)
【分類號】：O212.6

【相似文獻】

相關(guān)期刊論文 前2條

1 費榮昌;關(guān)于系統(tǒng)聚類法的若干注記[J];無錫輕工業(yè)學院學報;1984年02期

2 李永慈;;離差平方和分解的投影分析[J];大學數(shù)學;2008年04期

，

本文編號：2191938