Rapid identification of an effective bauxite gas reservoir by principal component analysis

Abstract

Abstract In recent years, industrial gas flow has been obtained from the bauxite gas reservoir in the southwestern Ordos Basin, which has made the identification of aluminium-bearing rock reservoirs a popular topic. To accelerate the exploration and development of this type of gas reservoir, major element testing, rock thin section identification and principal component analysis (PCA) were conducted, and a method for rapid and accurate identification of bauxite reservoirs via conventional logging was established. The test results clearly revealed the vertical stratification of major elements and three lithologies in the aluminium (Al)-bearing rock series in the study area. The log response characteristics of effective gas reservoirs were summarized, providing a basis for subsequent research on identifying effective bauxite reservoirs via mathematical dimensionality reduction of logging curves. The porosity comparison of strata with different lithologies suggests that dissolution pores are more developed in Al-rich layers, providing insight for identifying effective reservoirs by Al2O3 content. On the basis of the above findings, a lithological identification chart of Al-bearing rock series was established via principal component analysis (PCA), and an effective bauxite reservoir logging identification model based on Al2O3 content prediction was developed. The results show that using the dimensionality reduction method for principal component analysis of logging curves with overlapping information can avoid model distortion caused by multicollinearity. The research results can be used to identify bauxite reservoirs quickly and accurately without other test data.

Keywords

• Bauxite
• Aluminium-bearing rock
• Principal component analysis
• Linear dimension reduction
• Effective reservoir
• Logging identification