“A Simulation-Based Process to Predict the Impact of Hydraulic Fracture Parameters on EUR: A Tight Gas Example”, C. Yetkin, T. Firincioglu, NITEC LLC, A.M. Haney, Encana, SPE Eastern Regional Meeting, Lexington, KY, October 3-5, 2012. SPE 161350
A Simulation-Based Process to Predict the Impact of Hydraulic Fracture Parameters on EUR: A Tight Gas Field Example
This study develops a process that determines the critical hydraulic fracture parameters and quantifies their impact on the EUR by combining reservoir simulation with probabilistic analysis methods. The process is verified by a real field case example in a tight gas reservoir. The final product can be applied to other unconventional reservoirs to ultimately maximize revenues by planning superior fracturing operations and optimizing well spacing.
A detailed dual-porosity, 3 section reservoir model was created and history matched to model the flow mechanism. A fine layered (2-3ft) geostatistical model was utilized in simulation without upscaling. The dual porosity formulation enabled the simulation model to represent the hydraulic fracture – matrix interaction properly so that the flowback and formation water production could be matched also.
During the history matching phase, the parameters that control the impact of hydraulic fractures on the recovery were identified as follows:
Matrix-fracture exchange: this parameter represents the complexity of the fractures and is controlled by the surface area created during hydraulic fracturing.
Fracture conductivity: this is effectively the permeability of the hydraulic fracture
Half-length: this parameter impacts the extent of the hydraulic fracture, therefore the amount of matrix that has been accessed.
Job size: The size of the water volume injected during the hydraulic fracturing process
In this work, an internal proprietary technology that creates a response surface for the combination of the parameters defined above was utilized. The history matched simulation model was automatically modified to create the necessary cases to calculate a multi-dimensional response surface. The created response surface was then used to do Monte Carlo simulations to create P10 to P90 probabilities of the total gas production (EUR).
The results of the study allowed us to understand not only the mechanisms operating in the reservoir being studied but also the required hydraulic fracture parameters (ranges) to achieve a given EUR of a specific probability. The same algorithms were then be used to predict the future performance of other well spacing patterns and hydraulic fracture job sizes.