关键词: 混源油, 多元数理统计学, 交替最小二乘算法, 塔河油田, 塔里木盆地

Abstract:

         Quantificational discrimination of mixed oils is a hot concept and difficult point in petroleum geology and geochemistry research. This article shows a case study on Tahe oilfield, Tarim Basin, NW China. Comparative study of geochemistry and multivariate data analysis  indicate  that oils accumulated  in Ordovician  reservoir  in Tahe oilfield are characterized by multi-stage hydrocarbon generation and mixed sources. The early-stage generated oils were biodegraded,  thus  the  light hydrocarbons, alkanes, and regular steranes in present oils show features of later-stage generated oils. Quantitative data of tricyclic terpanes and hopanes with  little  influence of biodegradation were calculated  through Alternating Least Squares algorithm. The  results show that there were four end-member oils contributed to the mixed oils in this region. The end-members 1 and 2 indicate middle-upper Ordovician source  rocks, and  the end-members 3 and 4  indicate Cambrian source  rocks.  In  the main area of Tahe oilfield, oilsgenerated from the Cambrian source rocks were more than those from the middle-upper Ordovician source rocks. But in the area around the main area, the middle-upper Ordovician source rocks were the major contributors. For the whole oilfield, the Cambrian source rocks were more  important. This comparative research shows  that  the multivariate data analysis plays significant role  in calculating proportions of mixed oils and  identifying end-member oils. As an effective complement  to  traditional geochemistry study, multivariate data analysis is worthy to be popularized

Key words:  mixed oils, multivariate data analysis, Alternating Least Squares algorithm, Tahe oilfield, Tarim basin