Project Description

SPE 155499 PVT in Liquid-Rich Shale Reservoirs

Curtis Hays Whitson (Norwegian University of Science & Tech) | Snjezana Sunjerga (INA Naftaplin)

SPE Annual Technical Conference and Exhibition, San Antonio, Texas, USA, 8-10 October 2012

Download PDF here


This paper addresses sampling and PVT modeling of liquid-rich fluids produced from ultra-tight formations – “liquid-rich shale?? (LRS) reservoirs.1 Proper PVT treatment in these unconventional reservoirs is important to provide improved short- and long-term oil and gas production forecasts, and define the initial oil and gas in place.

We give recommended practices for sampling, laboratory PVT tests, developing PVT (EOS and black-oil) models, and estimating the in-situ reservoir fluid system (composition, saturation, and initial gas-oil ratio). Fluid systems studied include a wide range from lean gas condensates to volatile oils, typical of what is found in the Eagle Ford, Avalon, and other liquid-rich shale plays in North America – with producing oil-gas ratios ranging from 10-1,000 STB/MMscf.

LRS producing wellstreams, usually expressed in this paper as a producing oil-gas ratio (OGR) or “liquid yield?? 2, are always much leaner than what would be produced from a conventional, higher-permeability reservoir containing the same initial reservoir fluid system. Conventional reservoirs typically produce an initial mixture (for months or years) that is quite similar to the in-situ initial reservoir fluid. The anomalously-low producing OGR of LRS wells is due to very low permeabilities that lead to large drawdowns and fluid flow with localized and large gas-to-oil mobility ratio gradients near the fractures. We show that the loss in oil is a staggering factor of 2 to 50! The liquid yield will be approximately constant from the early days of initial testing throughout the well’s entire life.

The degree of oil recovery in LRS wells is associated mainly with two issues. First and foremost, whether the reservoir is initially saturated with oil (Soi=1-Swc) or gas (Soi=0). For example, with the in-situ solution OGR of ~350 STB/MMscf (initial GOR of ~3,000 scf/STB), the producing OGR might be 100 STB/MMscf for an oil reservoir (Soi=1-Swc), while it might be less than 10 STB/MMscf for a gas reservoir (Soi=0). Second, for oil LRS reservoirs, oil recovery loss is greatest for near-saturated initial conditions, with oil recoveries increasing as the oil reservoir becomes more undersaturated; degree of undersaturation does not have an impact on the large oil recovery losses seen in all LRS gas reservoirs.

Another important result from our study is showing how liquid yield (OGR) evolves with time for LRS wells. It is shown for planar “slab?? fracture geometries that the expected infinite-acting behavior is a constant OGR that may last many years or decades. A less-constant intermediate-to-long-term OGR development is found in naturally- or induced-fracture “networks?? consisting of a collection of matrix blocks surrounded by fractures. OGR variation depends on network fracture “block?? size.

The paper shows that it is necessary to combine single-well, finely-gridded numerical modeling of LRS wells to properly develop valid PVT models and in-situ fluid description. Conventional “PVT?? sampling and initialization procedures are alone inadequate for liquid-rich shale systems, but additionally require proper treatment of near-well reservoir flow and phase behavior to properly link the significant contrast in producing wellstreams and in-situ fluids.

Finally, we propose a special PVT laboratory test for LRS systems.