Source rocks, such as organic-rich shale, consist of multi-scale pore structures which include pores with sizes down to nano-scale contributing to the storage of hydrocarbons. During my Ph.D. studies, using molecular simulations, I showed the applications of the concept of composition redistribution of the produced fluids in source rocks. The hydrocarbons in the source rock partition into nano-confined fluids with significantly varying physical properties across the nanopore size distribution of the organic matter. This partitioning is a consequence of multi-component hydrocarbon mixtures stored in nanopores showing a significant compositional variation with the changing pore size and pressure. It is firstly observed that this variance leads to capillary-condensation of fluids in nanopores at the lower end of the organic pore size distribution. Condensation impairs the transport ability of the fluids left behind in nanopores and consequently, their recoveries are reduced significantly. In the light of these microscopic scale observations, I developed a new volumetric method for predicting hydrocarbons in-place honoring the compositional variability across the measured pore size distribution in the presence of nano-confinement effects. My approach allows the reservoir engineer to differentiate mobile bulk hydrocarbon fluids from the fluids under confinement and from the capillary-condensed trapped fluids. The low recoveries from the organic nanopores makes the source rocks potential resources for enhanced oil recovery. In addition, as part of my thesis work, I considered lean gases (such as COv2 and Nv2) injection for enhanced nano-confined oil recovery. The concept of gas injection is firstly developed in equilibrium molecular simulation research. I showed that lean gas injection could influence the vaporization pressure of the confined hydrocarbon fluid mixture and strip additional hydrocarbon molecules from the organic pores. This mechanism is known as the vaporizing gas drive in the literature. On the other hand, the several limitations of COv2 injection are found during my study, and ethane injection is alternatively considered for enhanced nano-confined oil recovery. I studied the ethane injection extensively and compared with COv2 injection. Ethane has a better stripping ability against heavy hydrocarbons, and also enhances the mobility. At reservoir scale, I propose a new robust method of simulation-based history-matching and optimization for future reserve prediction. This approach considers the total fracture surface area for the drainage of hydrocarbons as a key quantity in production from horizontal shale gas wells with multiple-hydraulic fractures. The effective fracture surface area is estimated by incorporating an analytical solution of production rate transient data associated with the formation linear flow and introduced as an additional constraint to the optimization. Stress-dependent models are employed for the fracture width and the matrix permeability change during the production. The new approach not only predicts the reserve but also the time for the fractures to close significantly when the fractures no longer produces economically. This time indicates the life of the well.
