Objective The Ordos Basin is a pivotal region for unconventional oil and gas production in China. The Yanchang Formation Chang 7 reservoir possesses abundant shale oil resources and significant exploration potential. Reservoir geomechanics evaluation is critical for guiding the efficient development of these resources. Characterizing the geomechanical properties of the Chang 7 shale oil reservoir and establishing a quantitative prediction model are essential for evaluating reservoir fracability and optimizing sweet spot intervals.

Methods In this study, representative core samples were collected from the Chang 7 reservoir in the southern Ordos Basin. Geomechanical parameters, including elastic modulus, Poisson's ratio, and present−day in−situ stresses, were determined experimentally. Subsequently, a geomechanical parameter prediction model was constructed using the BP neural network based on logging data to achieve quantitative evaluation of the reservoir.

Results The results indicate the following: The BP neural network−based model demonstrates high accuracy, with minimal error between predicted results and measured values; The key geomechanical parameters of the Chang 7 shale oil reservoir exhibit significant heterogeneity. Specifically, the elastic modulus is between 16.26 GPa and 59.12 GPa, the fracture toughness is between 0.2~1.2 MPa·m^0.5, the horizontal maximum and minimum principal stress range from 20 MPa to 43 MPa, 12 MPa to 38 MPa, respectively; A fracability evaluation index F was established based on these reservoir geomechanical parameters to classify the reservoir quality. The reservoirs are categorized into four levels: Class Ⅰ reservoirs F>2.00, Class Ⅱ reservoirs 2.00>F>1.50, Class Ⅲ reservoirs 1.50>F>0.10, and Class Ⅳ reservoirs F<1.00. The BP neural network is an effective method for the precise prediction of reservoir geomechanical parameters.

Conclusions These findings provide scientific guidance for the optimization of hydraulic fracturing designs in the region.