Dynamic saturation distribution of fluids in oil reservoirs is arguably the most important piece of information petroleum engineers and geophysicists need. It allows them to better characterize oil fields, choose right enhanced oil recovery techniques and make accurate recovery forecasts. The main objective of the present paper is to develop and test a novel non-intrusive technology that enables real time in-situ monitoring of fluid saturations in porous media during immiscible floods.
To achieve this goal we designed, constructed and commissioned a high pressure (up to 25MPa) and temperature (up to 200°C) core-holder with an antenna that allows for low power electromagnetic sweeps in the radio frequency range. A novel inversion algorithm was created and incorporated with our core-holder-pump system that enables dynamic acquisition of the oil and water saturations. Experiments were conducted through sand packs saturated with water and then displaced with oils of varying viscosity (drainage) followed by water flooding (imbibition). Reflection and transmission coefficients in the frequency domain were measured while flooding using a commercial vector network analyzer connected to the core holder. Recorded frequency data was processed with the novel high-resolution inversion technique to obtain impulse reflection and transmission responses in the time domain. These responses were further normalized to dynamically track the water-oil volumetric saturations within the porous media. Every displacement experiment has been performed in both vertical and horizontal positions to emulate water-oil override and underride scenarios. Material balance calculations were performed to validate fluid saturation profiles for all imbibition and drainage experiments every 5 % of the pore volume of the fluids injected.
Frequency domain reflection and transmission electromagnetic responses were measured every 1 % of the fluid injected. Material balance was used to validate the measured saturation profiles. The goodness of fit was calculated between these two independent measurements for every flood experiment performed. The mean-square error was calculated to be around 1.26% of the total pore volume on average. Late breakthroughs and piston-like displacement were observed in the floods with favorable mobility ratios. In contrast, much earlier breakthroughs have been registered in all the experiments with unfavorable mobility ratios due to fingering. All our observations are in agreement with the current theory of immiscible displacement in porous media.
The novel automated system paired with the inversion algorithm were developed to allow for virtually real time monitoring of the fluid saturations during imbibition and drainage displacement cycles. Our technology is shown to be a promising candidate to compliment resistive logging measurements in the field.