| 摘要: |
| 以2 种原油和3 种溢油分散剂的混合物为对象, 借助SPSS 统计分析软件, 采用主成分分析法研究基于诊断比值的原油分类, 旨在判别分散剂对原油的影响。KOM 值近似等于0.6、Bartlett球形度检验对应概率P 值为0, 表明原始数据适合做主成分分析; 前两个主成分F1和F2的特征值均大于1, 累积贡献率为81.433%, 多变量可降至2 维。根据F1和F2表达式得到主成分分类图可知: 两种原油中加入不同分散剂后还可在图中大致区分, 富肯-2 号分散剂对两种原油几乎没有影响, GM-2 分散剂对渤海原油分类程度有一定影响, 但对华北原油却有较大影响; 海鸥4 号分散剂对两种原油分类程度都有影响。结果表明: 主成分分析法可用于添加分散剂的溢油分类辨别, 研究发现一些分散剂对原油影响较小, 而某些分散剂对原油有复杂的影响。因此, 溢油指纹鉴定需考虑分散剂的影响。 |
| 关键词: 主成分分析 分散剂 影响 |
| DOI：10.11759/hykx20141109002 |
| 分类号: |
| 基金项目:国家自然科学基金项目(41206095, 21207010);国家教育部博士点基金(20122125120010) |
|
| Analysis of principal component of crude oil influenced by dispersant |
|
|
| Abstract: |
| In this paper, the mixture samples were obtained by 2 kinds of crude oils and 3 kinds of oil spill dispersants (referred to as dispersant). The classification of crude oil based on the diagnostic ratios was analyzed and calculated by Principal Component Analysis (PCA) with the assistance of SPSS statistical analysis software, aiming at judging whether the dispersant influenced the crude oil or not. The value of KMO was approximately 0.6 and the probability P-value of Bartlett’s test of sphericity was 0, which was suitable for PCA. The eigenvalues of the first two principal components were both greater than 1, and cumulative contribution rate was 81.433%, so more variables can be dropped to 2. Finally, according to mathematical expressions of principal component F1 and F2, the principal component classification figure with 20 samples were obtained. It was found that 2 kinds of original crude oils were still clearly distinguished and little affected by Fuken-2 dispersant, but the classification of Huabei crude oil was affected more significantly than Bohai crude oil by GM-2, and the two were both influenced significantly by Haiou 4 #. Results showed the PCA is an effective analysis method to oil spill research, which can be used to classify different kinds of crude oils. Studies have also found that some of dispersants have less effect on crude oils, and some others were complex. Therefore, the fingerprint identification of oil spill needs to consider the effect of dispersants. |
| Key words: Principal Component Analysis dispersant influence |
