Identification of distinctions of immiscible CO2 huff and puff performance in Chang-7 tight sandstone oil reservoir by applying NMR, microscope and reservoir simulation

CO2 huff and puff (HNP) is one of the most effective methods to improve tight oil recovery after the primary depletion process. The seepage mechanisms between CO2 and crude oil are complicated in porous media during CO2 HNP process. Therefore, in this paper, the CO2 HNP process of Chang-7 tight oil reservoir, Ordos Basin, China, was studied by nuclear magnetic resonance (NMR) technology, microscopic observation and numerical simulation. Experimentally, using NMR technology and microscopy methods, the distribution characteristics of residual oil during CO2 HNP process were measured intuitively. Numerically, a group of core-scale and field-scale simulations considering molecular diffusion and asphaltene precipitation were established to further verify and elongate the experimental results. The results show that at the initial state, the crude oil in the tight core was mainly distributed in nanopores, sub-micro-nanopores and sub-micropores, where the oil content exceeded at least 73% in these pores. During CO2 HNP process, the oil recovery was more pronounced for the 1st and 2nd rounds than for 3rd to 5th rounds. Notably, even if the cores with more nano-pores were more favorable for the 4–5th CO2 HNP rounds, the oil molecules in nanopores were still difficult to be available. Moreover, the CO2 sweep scope could be divided into displacement affected region and diffusion affected region. CO2 could effectively drive the crude oil in the displacement affected region. While the oil could be successfully displaced by dissolved gas flooding in the diffusion affected region only under the appropriate conditions. Meanwhile, the core-scale numerical models confirmed that it was necessary to consider molecular diffusion and asphaltene precipitation factors, which would make the simulation results in line with the experiment. In terms of the ultimate oil recovery, the field-scale model only considering the diffusion (2.456%) > the model both considering the diffusion and asphaltene (2.436%) > the model without considering the diffusion and asphaltene deposition (2.412%) > the model only considering the asphaltene deposition (2.388%).