Factors Controlling the Properties of Hydraulic Fracture Networks in Naturally Fractured Formations

Abstract

The interaction between natural and hydraulic fractures (HFs) results in the formation of complex fracture networks. A great deal of uncertainty exists around the geometry and connectivity of these fracture networks. The two primary objectives of this study are: (1) to show how the properties of the natural fracture network (orientation, density, length) control the created hydraulic fracture network, and (2) how microseismic (MS) and fiber optic data obtained during fracturing can be used to obtain better estimates of the fracture geometry in such complex systems. These effects are illustrated by using a new fracture propagation model that accounts for the interaction of the propagating hydraulic fracture with the natural fracture network and also allows us to compute and compare the fiber optic and microseismic data from a field site. A stochastic discrete fracture network (DFN) was constructed, incorporating the density, length, and orientation distribution of natural fractures at the FORGE site in Utah. Hydraulic fracture treatments within the DFN were modeled using the displacement discontinuity method (DDM) for stress. The coupling of strain with fluid flow in the created fracture network was achieved via the finite volume method. Fiber was installed in an observation well and Distributed Acoustic Sensing (DAS) measurements were obtained and analyzed. The influence of natural fractures on the DAS data is demonstrated by systematically varying the following: (1) fracture orientation angles ranging from 0° to 90° from the maximum stress direction, (2) fracture density ranging from 0.0002 to 0.005/m3, (3) lengths varying from 10 meters to 120 meters. The magnitude and moment of the microseismic events were computed to show the expected seismic clouds that would be generated. Fiber optic responses were also computed to show the expected results under different conditions. Finally, the geometry of the created fracture network was diagnosed and related to the microseismic and fiber responses. The intrinsic characteristics of the created fracture network can be identified in DAS waterfall plots. These fracture network characteristics change systematically based on the natural fracture orientation, density and length distribution. Most importantly, variations in these factors affect the number of isolated and branched fractures created in the fracture network. The effect of stress shadowing on the development of continuous fracture systems originating from different perforation clusters is clearly observed. Results are presented for the different sensitivity cases to illustrate the importance of the different properties of the natural fracture system on the final fracture network. This study, for the first time, incorporates and quantifies the impact of natural fractures on DAS and MS monitoring data. The findings of this study allow us to demonstrate how such data, together with geomechanical models, can be used to better characterize hydraulic fracturing networks in naturally fractured reservoirs, thereby improving hydraulic fracturing designs in such complex systems.