Abstract:

The E₂ s_4^{d^3} layer in the Leijia area of the western sag is characterized by lacustrine fine-grained sedimentary rocks, intermixed with clay, felsic, carbonate, and analcime minerals. Recent exploration efforts have been primarily directed towards tight oil and shale oil within this region. Previous research amalgamated drilling, logging, and analytical testing data to comprehend the geological characteristics of fine-grained rock reservoirs. Employing unsupervised learning techniques, particularly K-means cluster analysis, facilitated the classification of fine-grained sedimentary rocks into distinct types based on 303 X-ray diffraction whole-rock quantitative analysis data. The lithology of the E₂ s_4^{d^3} layer was categorized into carbonate rock type, felsic mixed fine-grained rock type, and analcitic mixed fine-grained rock type. Notably, carbonate rocks and analcite mixed fine-grained rocks exhibit superior brittleness characteristics, are prone to cracking, and have developed dissolved pores and caves along fractures, thereby enhancing storage and seepage properties. Evaluation of the E₂ s_4^{d^3} reservoir revealed a small pore throat radius, poor sorting, small homogeneity coefficient, and inadequate pore structure. Effective reservoir formation predominantly relies on micropores and microfractures. Through the overlapping favorability method, six factors influencing reservoir development were integrated and assessed, including brittleness index, thickness of carbonate rock and analcitic mixed fine-grained rock, fracture density, average porosity, average permeability, and organic carbon content. Favorable zones for reservoir development are primarily distributed along the long axis of the lake basin, encompassing specific wells such as Lei 15, Lei 84, Lei 59, Shu 90, and Lei 93. This summary provides a comprehensive overview of the geological study, encompassing methodologies, findings, and implications for oil exploration in the Leijia area.

Key words: Liaohe Depression, western sag, fine-grained sedimentary rocks, unsupervised learning, overlapping favourability method, reservoir evaluation