A laboratory study on the dielectric spectroscopy of sandstone and the improvement of dispersion model

• Published in: Journal of petroleum science & engineering 2022-09, Vol.216, p.110655, Article 110655
• Main Authors: Chen, Shichang, Ke, Shizhen, Jia, Jiang, Cheng, Linfeng, Shi, Hongwei, Zhang, Yumeng
• Format: Article
• Language: English
• Subjects: CRIM model; Dielectric spectroscopy; Interface polarization; Maxwell-Garnett; Textural model
• ISSN: 0920-4105; 1873-4715
• DOI: 10.1016/j.petrol.2022.110655

Dielectric spectroscopy can be used to identify the fluid properties of complex reservoirs, but no existing model can accurately describe the dielectric dispersion behavior of rocks in a wide frequency band. In this study, dielectric spectroscopy measurements were performed on 20 natural sandstone samples with different water saturation levels and salinity in the frequency band of 20–1000 MHz. A model named textural-EP was established based on the textural model by changing the dielectric constant of water in the pores, which further improved the accuracy of the rock's dielectric spectroscopy. Experimental results show that the effect of interfacial polarization cannot be ignored in rock dielectric spectroscopy at low frequencies when the salinity of the pore water is high, resulting in poor correlation at low frequencies when using the textural model to fit the rock dielectric spectroscopy. Another model named textural-EPMW with correction terms was obtained by adding the Maxwell–Wagner formula to the textural model, which improved the shortcoming of the textural model's poor correlation with the rock dielectric spectroscopy in low-frequency bands. The results obtained using the textural-EPMW model have a high consistency with rock's dielectric spectroscopy in a wide frequency band; thus, the textural-EPMW model is more suitable for describing the dielectric dispersion behavior of rocks. •The dielectric properties of water in the pores are different from those of pure water.•The dielectric constant of water in the pores can be expressed using exponential functions and power functions.•The improved dielectric dispersion model can correspond well with the experimentally measured data.