Evaluating Gas-Injection Performance in Fractured Reservoirs Using Discrete Fracture Network Models

Abstract

This reference is for an abstract only. A full paper was not submitted for this conference. Abstract This paper presents an alternative approach to modeling gas injection performance in naturally fractured reservoirs. The two conventional methods of modeling fractures have been to either use a single porosity medium with enhanced tensor permeability or a dual porosity system, where the matrix feeds the higher permeability fracture network. The former approach captures variability in well productivity, which is important during primary production but may not be suitable for modeling displacement mechanisms at the large simulation grid block scale. The dual porosity approach may provide some more flexibility in accounting for displacement mechanisms (imbibition, drainage, gravity effects etc.), however, calibrating dual porosity models (using an appropriate matrix to fracture transfer coefficient) may again be difficult. We discuss an alternative methodology for generating, simulating and evaluating displacement performance using Discrete Fracture Network (DFN)models. While computationally expensive, the DFN models for smaller sectors could be used to calibrate displacement performance of full field dual or single porosity models. We generated stochastic realizations of fracture patterns conditioned to available data on fracture spacing and orientation using a multi-point statistics approach. These stochastic realizations were used to investigate the impact of fracture characteristics on displacement behavior. Fractures were explicitly treated as discrete entities having extremely low pore volume and high transmissibility. The rock-fabric permeability realizations were generated using sequential gaussian simulation, honoring well log data. The models used to demonstrate the DFN approach were 1500m X 1500mspatially, with thickness varying from 200– 400m. They comprised of one up-dip gas injector and two down-dip producers. Simulations with the DFN approach show the injected gas moving rapidly through the fracture network, with significant bypassing of oil. The effective permeability and dual porosity models on the other hand, displayed more stable gravity drainage behavior. Displacement characteristics observed from the DFN sector models were then used to calibratefull field dual porosity models. Simulations showed that with higher rockfabric permeability, the displacement efficiency was significantly improved. Sensitivity studies were conducted to investigate the impact of fracture spacing, rock fabric permeability and well intervention techniques on recovery using the DFN approach.