中文摘要：

基于裂解气相色谱．质谱联用技术（PY—GC／MS）对两种类型汽油（93#和97#）的燃烧烟尘进行了定性和定量分析．对分析结果应用主成分分析法（PCA）进行降维预处理，然后利用软独立建模分类法（SIMCA）对93#和97#两种汽油分别建立了类模型．在显著性水平α=1％的条件下，两类模型对训练集的识别率为100％，拒绝率为100％；97样模型对未知样本预测集的识别率、拒绝率均为100％；93#JfN型对未知样本预测集的识别率、拒绝率达到80％．研究表明PY-GC／MS结合软独立建模分类法（SIMCA）可以实现对汽油燃烧娴尘的检测分析和类型判别．

英文摘要：

Qualitative and quantitative analyses of 93# and 97# gasoline soot were carried out based on the pyrolysis gas chromatography mass spectrometry （PY-GC/MS） technique. Principal component analysis （PCA）was applied to the reduction of dimensions of the analytical results. Two predictive models of 93# and 97# gasoline were built separately using SIMCA pattern recognition method. Under the condition of significance levelα=l%, both the identification rate and the rejection rate of the two models for the training set were 100%, while for the predictive set, the identification rate and the rejection rate of 97# model were both 100%, and those of 93# model reached 80%. Results show that PY-GC/MS with SIMCA pattern recognition method can realize the detection and classification of gasoline soot.