Pore structure analysis and permeability prediction of shale oil reservoirs with HPMI and NMR: A case study of the Permian Lucaogou Formation in the Jimsar Sag, Junggar Basin, NW China

Abstract

The pore structure characteristics and formation permeability are important reservoir information for the development of shale oil. The pore structures of the shale oil formation in the Lucaogou Formation in the Jimsar Sag are characterized with high-pressure mercury injection (HPMI), nuclear magnetic resonance (NMR), and fractal theory. The capillary bundle model, geometric model, and wetting phase model are used to obtain the fractal dimension of the capillary pressure curve. Among them, Df1 from the capillary bundle model is between 2.1937 and 3.4131, Df2 from the geometric model is within a range of 1.0414–3.575, and Df3 from the wetting phase model is from 0.5691 to 2.7047. For NMR spectra, the pore space is divided into micropores, mesopores, and macropores based on the peak value of the NMR T2 spectra, and the fractal dimension of different pore types is calculated, where Dfmi is from 1.564 to 2.301, Dfme is in a range from 2.082 to 2.941, and Dfma is between 2.831 and 2.991. The relationships between the fractal dimension and petrophysical parameters are analyzed, and the results indicate that among the three types of models, the fractal dimension of the capillary bundle model has the highest correlation with the core petrophysical parameters. Therefore, it is recommended to use the capillary bundle model to evaluate the fractal characteristics of pore structures. In particular, the fractal dimension is positively correlated with the throat coefficient and negatively correlated with the size of the pore throat. In contrast, the fractal dimension of NMR has a poor correlation with core petrophysical parameters. There is no obvious correlation between the fractal dimension of micropores and petrophysical parameters, while the fractal dimension of mesopores and macropores is only correlated with the throat-sorting coefficient, and the correlation coefficient is less than 0.15. Thus, the fractal dimension from NMR data is not suitable for studying the fractal characteristics of shale oil cores in the Lucaogou Formation. In addition, we predicted the permeability of shale cores using the Swanson model, Pittman model, and Winland model. The permeability prediction results show that the Swanson model has the greatest accuracy, followed by the Winland model. Among them, the accuracy index (ACI) of the Swanson model is 0.901, and the ACI of the Winland model is between 0.483 and 0.897. The permeability prediction by the Pittman model is poorest among these three models.