Compositional Modeling of Tight Oil Using Dynamic Nanopore Properties

A typical tight oil reservoir such as the Bakken has matrix pore sizes ranging from 10 nm to 50 nm. At such small scales the confined hydrocarbon phase behavior deviates from bulk measurements due to the effect of capillary pressure. In addition, compaction of pore space can bring about order of magnitude changes for tight oil formation properties during pressure depletion further exacerbating these deviations. Without considering these facts a conventional reservoir simulator will likely not be able to explain the inconsistent produced GOR observed in the field compared to simulated results. The effect of these inaccuracies on ultimate recovery estimation can be devastating to the underlying economics. This paper presents a compositional tight oil simulator that rigorously models pressure dependent nanopore-impacted rock and fluid properties, such as suppression of bubble point pressure, decrease of liquid density, and reduction of oil viscosity as well as their interactions with pore space compaction. The cubic Peng-Robinson equation of state is used for phase behavior calculations. Capillary pressure is evaluated by standard Leverett J-function for porous media. Modifications to the stability test and two-phase split flash calculation algorithms are provided to consider the capillarity effect on vaporliquid equilibrium. The simulator can capture the pressure-dependent impact of the nanopore structure on rock and fluid properties. As a result, the problem of inconsistent GOR is resolved and the history matching process is greatly facilitated. It is shown that inclusion of these enhanced physics in the simulation will lead to significant improvements in field operation decisionmaking and greatly enhance the reliability of recovery predictions. Introduction The recent advances in massive hydraulic fracturing techniques have enabled the oil industry to economically extract hydrocarbon from ultra-tight, unconventional resources, such as shale gas, liquid rich shale and tight oil. The success in North America has stimulated the development of unconventional plays worldwide. For example, a marine shale play in southern China has showed large potential and attracted great attention (Wei et al. 2012; 2013a, b). However, despite the great success and potential, the understanding of fluid flow mechanism in shale and properties in confined pore space is still poor. The flow mechanism in the shale matrix is complicated by organic and inorganic portions of the matrix with distinct wettabilities. Yan et al. (2013 a, b, c, d) proposed a micro-model to model single-phase gas and two-phase gas-water flow in shale matrix blocks by considering different flow mechanisms in organic and inorganic nanopores. Yan’s work also upscaled the single-phase gas flow to well-scale modeling via the apparent permeability approach. On the other hand, the fluid properties in the confined nanopore space deviate from the corresponding bulk measurements in which zero vapor-liquid interface curvature is assumed. This assumption is generally held when the vapor-liquid equilibrium takes place in PVT cells. But, when the fluid is confined in pore spaces of nano-size, the significant interfacial curvature may cause a large capillary pressure difference between liquid and vapor phases. The effect of capillary pressure on vapor-liquid equilibrium is not new to the oil industry. A number of researchers have conducted both experimental and theoretical investigations with general conclusions that capillarity effect on vapor-liquid equilibrium is negligible for conventional reservoirs (Leverett 1941; Sigmund et al. 1973, 1982; Shapiro and Stenby 1997; Shaprio et al. 2000). Perhaps due to this reason, essentially all the current commercial simulators assume no pressure difference between vapor and liquid phases during flash calculations. However, ignoring capillarity in vapor-liquid equilibrium might not be a valid assumption for unconventional reservoirs. A typical tight oil reservoir such as the Bakken has matrix pore size ranging from 10 nm to 50 nm. At such small scales, the confined hydrocarbon phase behavior is believed to deviate from bulk measurements due to the extra capillarity effect. Rock wettability is another factor to consider when dealing with capillary pressures. Wang et al. (2012) performed a wettability