It is well-known that many unconventional reservoirs experience porosity and permeability changes with pressure change during production. In recent work, authors have incorporated geomechanical modeling into production analysis procedures to account for stress-sensitivity of permeability of unconventional gas reservoirs, such as shale gas. Such corrections are necessary for deriving both accurate estimates of reservoir and hydraulic fracture properties from rate-transient analysis and for developing accurate long-term forecasts.
Some shale gas reservoirs are unique in that dynamic changes may occur in both the induced hydraulic fracture AND matrix permeability, which could have a substantial impact on shale gas productivity. Stress-dependence of shale gas permeability has been quantified in the lab by several researchers, but measurements of this kind for propped or unpropped fractures under in-situ conditions are less routinely measured. For the latter, a variety of mechanisms, caused in part or wholly by stress changes in the induced hydraulic fracture, could lead to conductivity changes.
In the current work, we investigate the impact of both stress-dependent matrix permeability and fracture conductivity changes on 1) rate-transient signatures and 2) derived reservoir and hydraulic fracture properties. Stress-dependent matrix permeability is incorporated into rate-transient analysis using modified pseudopressure and pseudotime formulations, and fracture conductivity changes are approximated by applying a time-dependent (dynamic) skin effect.
We demonstrate that when rate-transient analysis incorporates both matrix permeability changes and dynamic skin, the resulting rate-transient signature looks very similar to other shale plays (long-term transient linear flow). Uncorrected data appear to have a very short transient linear flow period, followed by apparent boundary-dominated flow. The impact of the applied corrections on estimates of system permeability and fracture half-length is demonstrated as is the impact on production forecasts.