Abstract
AbstractIn conventional reservoirs, 2D-NMR fluids evaluation targets the free fluid part of the total porosity with the assumption that the bound fluid is irreducible water. As such, pulse sequences are designed for long relaxing fluids, and the interpretation commonly assumes free diffusion of hydrocarbon molecules in water-wet pores. This is clearly not appropriate for unconventional reservoirs such as shale gas and shale oil where the fast-relaxing fluids of interest reside in the bound fluid region.We show a revised 2D-NMR model that focus on fast relaxing fluids. In unconventionals, the three causes of fast relaxation are: small pore size, heavy oils and wettability alteration. The fast relaxation has the following consequences with respect to diffusion. In small pores, fluids cannot diffuse freely, and hence, the free diffusion lines of water, gas and oil must be corrected accordingly. In heavy oils, the oil relaxation can be enhanced by a wettability change to an oil/mixed-wet system. Another case of hydrocarbon-wet systems is hydrophobic kerogen. Consequently, the oil diffusion line as a function of viscosity (T2) must also be modified before the 2D map interpretation. This can be accomplished within the framework of the restricted diffusion model previously applied to water and gas that captures both the effects of surface relaxation and geometric restriction to molecular motion.The results of the revised 2D-NMR model are shown through modeling and log examples in a shale gas reservoir and a shale oil reservoir. In real rocks, there is also a need to take into account simultaneously two models: an unconventional model as described above and a conventional model for long relaxing fluids associated with the matrix of the rock. The 2D-NMR log results are compared with lab results. Essentially, we show that adding diffusion and T1 information to standard T2 relaxation logs improves both the understanding and evaluation of unconventional reservoirs.
Cited by
12 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献