CO2 Water-Alternating-Gas injection (CO2-WAG) is still a challenging task to simulate and predict accurately, due to the complex interaction of CO2/oil phase behaviour, 3-phase flow and the heterogeneity of the porous medium. In this paper, we focus specifically on the regime of viscous fingering flow in CO2-WAG in heterogeneous systems because of its importance in elucidating this complex interaction. This work presents a detailed simulation study of both immiscible and near-miscible CO2-WAG and continuous CO2 displacements with unfavourable mobility ratios for 1D and 2D systems. 2D heterogeneous permeability fields were generated as Correlated Random Fields (CRF) with specified degrees of heterogeneity (permeability range, described by the Dykstra-Parsons coefficients, VDP) and structures (defined through the dimensionless correlation range, RL = λ/L). Our central aim is to improve the modelling of CO2 displacement in the transition from immiscible to miscible flows in CO2-WAG processes. To do so, two key physical mechanisms that occur during near-Miscible WAG (nMWAG) processes have been studied in detail, namely compositional effects (denoted as Mechanism 1, MCE) and low-interfacial-tension (IFT) film flow effects (denoted as Mechanism 2, MIFT). The low IFT effects in MIFT manifest themselves in an increased mobility of oil phase due to enhanced film formation and flow processes. This latter mechanism (MIFT) is modelled as an increased oil relative permeability using different well-known models (Betté and Coats) parameterized by the gas/oil IFT (σgo), calculated in the simulation from the compositional PVT model via a built-in correlation (the McLeod-Sugden equation, in this case). A range of various combinations of oil-stripping effects (MCE) and IFT effects (MIFT) has been tested to evaluate the potential impact of each mechanism on the flow behaviour such as the local displacement efficiency and the ultimate oil recovery. Oil bypassed by viscous fingering/local heterogeneity, can be efficiently recovered by WAG in the cases where both MCE and MIFT are taken into account (as opposed to either mechanism being considered alone). We also show that the way these two distinct but related mechanisms (MCE and MIFT) operate in near miscible conditions cannot be observed in (i) a simple 1D system such as a slim tube experiment, or (ii) in a heterogeneous system under continuous CO2 injection. Using tracer analysis in our simulations, we demonstrate that a major recovery mechanism in near-miscible WAG displacement is viscous crossflow between non-preferential (bypassed) flow-paths and preferential flow-paths (i.e. between the viscous fingers). Due to the significance of IFT effects (the MIFT mechanism), we also present comparative results from two of the IFT-dependent relative permeability models (Betté and Coats) showing the impact of each model on the simulation of the near-miscible WAG flow behaviour.
- Fine-scale simulation
- Near miscible
- Viscous crossflow
ASJC Scopus subject areas
- Energy Engineering and Power Technology