A Statistical Factor Analysis Approach to Reservoir Quality Prediction in the Tertiary Fluviolacustrine Sandstones of the Yaojin II Field in the Qaidam Basin, NW China

Xu, Tianguang¹, Xiaoru Li², Jincheng Liu² (1) IHS Energy, Houston, TX (2) PetroChina-Qinghai, Dunhuang, China

An integrated geological and statistical Factor Analysis method is applied in this study to quantitatively evaluate and predict Tertiary fluviolacustrine sandstones of the Yaojin II field in western Qaidam Basin. Reservoir rocks in the Yajin II field are dominated by Tertiary lithic arkose and arkose, which were buried at depths ranging from 500 m to 2,000 m.

Core samples from six wells are analyzed, and resulted geological and petrophysical data are used for the study of sedimentary facies, diagenesis, and wireline logging correlation. Parameters, including porosity, permeability, and net thickness, are selected to conduct the Factor Analysis. A total of 15,350 row data are derived from wireline logging from more than 170 wells, and these row data are correlated with the core data. The following steps are performed to complete the Factor Analysis: 1) calculate descriptive statistics; 2) calculate a correlation matrix of all variables to be used in the analysis; 3) extract principal components; 4) rotate factors to create a more understandable factor structure; 5) assign factor scores; and 6) interpretation.

The results of Factor Analysis indicate that porosity and permeability are the principle components with a total variance contribution of 91.28%. Reservoirs are classified as four categories according to calculated factor scores. Good quality reservoirs are delta front sandstones with an average factor score of 36.66, followed by delta plain and marginal lacustrine sandstones with an average factor score of 29.97 and 23.56, respectively. Diagensis, especially compaction and cementation, also explains the distribution and trend of factor scores.