Supervised Reservoir Facies Classification for Pre-Paleogene Buried Hill Formation and Hydrocarbon Geological Significance, Offshore South China

Abstract The pre-Paleogene buried hills in the South China Sea are crucial succession of deep oil and gas reserves. Buried hill X, formed from a composite metamorphic sandstone and granite, is controlled by multiple phases of tectonic activity, exhibiting a well-developed fracture network but poor matrix properties. Effective classification of buried hill reservoir facies is crucial for subsequent exploration and evaluation, as well as optimization of geological model for further reservoir prediction. To effectively classify reservoir facies, a comprehensive analysis was conducted using data from conventional logging, gamma ray spectroscopy, dual borehole imaging logging, sidewall core, and mud gas. First, lithology was identified using the thorium to uranium ratios and conventional logs, resulting in two main lithologies: metamorphic sandstone and granite. The granites were further subdivided into Type A, Type I, and Type S. Type A is of alkaline or non-orogenic origin, Type I is of igneous source region origin, and Type S is of sedimentary source region origin. Second, dual borehole imaging logging were used to identify natural fractures, classifying three types of effective fractures: vuggy open fractures, fully open fractures, and partially open fractures, and two types of closed fractures: clay-filled fractures and tight cemented fractures. Analysis of fracture occurrence and parameters revealed significant differences in fracture development across different lithologies. A negative correlation was found between matrix properties and fracture development; that is, poorer matrix properties correlated with more developed fractures. Based on the fracture calculation parameters from imaging logging, and petrophysical interpreted porosity and permeability, a supervised reservoir facies clustering analysis involving machine learning and manual intervention was conducted. According to regional experience, reservoir permeability and fracture aperture have greater weights than other parameters. The results yielded six reservoir facies categories, IP1 to IP6, with IP1, IP2 and IP3 exhibiting the best reservoir quality, IP4 and IP5 representing medium quality, and IP6 representing the worst quality. Finally, the relationship between reservoir facies and lithology was established, revealing that IP1 and IP2 reservoirs were the most prevalent and of the best quality in metamorphic sandstone, while IP6 reservoirs were the most prevalent and of the worst quality in A-type granites, and S-type and I-type granites exhibited medium-quality reservoirs. This supervised reservoir facies classification establishes a connection between reservoir quality and different lithologies, providing a more systematic understanding of the vertical variations in buried hill reservoirs. Multi-well reservoir facies classification results can also be incorporated into geological reservoir modeling to predict the 3D distribution of favorable reservoir facies. Simultaneously, the reservoir facies result also facilitate drill stem testing zone selection and reduce the exploration cost of offshore buried hill oilfields.