Modelling Multicomponent Flow in Tight and Shale Reservoirs

Abstract

Modeling of multicomponent flow in tight reservoirs has several challenges, including commingling of different transport mechanisms and lack of consensus on how to model such formations. This results in a variety of formulations being used to evaluate unconventional tight reservoirs, many of them often failing to address physics that become more relevant as permeability decreases, for example, diffusion. Since well-established methods used for conventional reservoirs cannot be directly applied to tight media, research on models that include relevant physics and improve predictability are extremely important.In this thesis, a fully implicit, fine-grid simulation tool was developed to evaluate production and enhanced recovery in tight unconventional formations. The tool is unique in the use of a rigorous mass transfer model including the real gas version of Maxwell-Stefan equations for diffusion, in association with physics of convection, phase equilibria and capillarity effects.Throughout the thesis, I demonstrate how the fine-grid simulation tool can be used to identify relevant transport mechanisms during production and injection of gases, to sensitize on how model simplifications impact hydrocarbon recovery predictions and to evaluate the impact of different physics on gas injection strategies.The model is used to investigate reservoirs in the dry gas, wet gas and gas condensate windows. In modeling dry and wet gas reservoirs, I demonstrate the effects of molecular partitioning on pressure decline and produced gas composition during primary production. The onset of liquid dropout is highly influenced by balancing forces of Knudsen and molecular diffusion, therefore, both physics need to be considered when simulating multicomponent flow in tight reservoirs.Subsequently, the model is used to evaluate injection of gases with different adsorption capacities during cyclic injection in unconventional tight reservoirs. It is observed that the mechanisms of hydrocarbon release are different depending on the type of gas injected, which affects the shape and velocity of displacement fronts. In particular, the model shows the advantages of injecting mixtures of carbon dioxide and nitrogen during enhanced recovery due to combined mechanisms of hydrocarbon release, i.e., competitive adsorption and partial pressure reduction.By applying the model to condensate reservoirs, I demonstrate how the interplay between diffusion and phase behavior will influence phase appearance and disappearance. It is then possible to track the location of condensate banks and composition profiles. A comparison with the widely used Klinkenberg formulation demonstrates the added value of using a more rigorous formulation in tight condensate reservoirs in terms of predictability of recovery factor per component and condensate yields.Finally, the model is used to investigate fluid partitioning between gas, liquid and adsorbed phase, by combining the real gas Maxwell-Stefan equation with a linear driving force model to evaluate the rate of adsorption. In this case, I demonstrate how the different adsorption capacities of molecules combined with physics of condensate vaporization during gas injection affects the displacement front in wet gas and gas condensate reservoirs. When adsorption and multiphase flow is present, re-adsorption of heavier hydrocarbons occur during condensate vaporization. Nevertheless, cyclic gas injection is still an efficient method of recovery of heavier hydrocarbons in condensate reservoirs.