Detailed experiments were performed in four instrumented boreholes to determine the growth pattern of an induced hydraulic fracture and its impact on flow through an ultra low permeability formation. The instrumentation allowed measurement of fracture pressure and width in three boreholes which were intersected by a fracture induced from a central well. The fracturing fluid was synthetic pore water to prevent formation damage.

The experiments showed that because of the difference between the instantaneous fracture velocity and fluid velocity inside the fracture the fluid front lags behind fracture front. There was a large pressure drop along the length of the fracture which caused rapid narrowing of the fracture width, poor hydraulic communication along the fracture length, and localized growths and fluid movements.

The data confirmed an off-balance fracture growth pattern, consisting of tensile and shear fractures and many branches.

At the end of the treatment, even though the fracture contained some residual opening, it had very little conductivity along its length. At the same time, because of its off-balance growth the fracture had the capacity to transmit fluid locally and trap and hold part of the treatment fluid inside the ultra low permeability formation.

These experiments have raised new questions about some of our common assumptions in hydraulic fracturing. These include the pattern of fracture propagation, pressure distribution inside the fracture, fracture closure and re-opening, in-situ stress measurement, etc.


Over the years, many questions have been raised about the accuracy of theoretical models used by the oil and gas industry for designing hydraulic fracturing treatments, as well as for post-treatment diagnostic purposes. These models are based on the assumption of a single tensile fracture propagating progressively away from the wellbore in a piston-like manner. The concern about accuracy of these models arises from the fact that the observable behavior of actual fracturing treatments (pressure, production increase, fluid and proppant flowback pattern, etc) differs significantly from prediction of these models.

Sandia National Laboratories performed the first investigation and instrumentation of actual treatments. In a series of comprehensive reports Warpinski et al 1,2,3,4, Smith5, and Cuderman6 reported fracture behaviors substantially different than predicted by theory. Later, Warpinski et al7 verified the same behavior in fractures created at greater depths. By coring through the fractured area they recovered numerous fracture branches spread along the fracture direction. They also reported substantially larger pressures at the wellbore than locations away from it and intercepted by observation holes.

In a recent article, Daneshy8 articulated the concept of off-balance fracture growth. In this mode, the propagation of the main tensile fracture is associated with numerous shear fractures and branches, such that the created network consists of a large number of misaligned fracture segments propagating randomly inside and outside of the existing fracture domain, Fig. 1. Shear fractures are characterized by the misalignment between the two fracture faces and much narrower width which consequently hinders fluid flow through it. Branches are characterized by their small dimensions and limited extent. Together, these features result in a fracture network that is narrower and shorter and shows much more resistance to fluid flow than predicted by theory. Although such fractures have a general large-scale orientation, on a smaller scale they may follow any horizontal or vertical plane.

The experiments reported here (Enachescu et al9) were conducted for two purposes; to observe the pattern of an off-balance fracture opening and closure, and, to determine if and how fluid dissipates in a hydraulically fractured ultra low permeability formation (Opalinus Clay). Furthermore, by leaving the fracture inactive for a period of time, insight was sought as to whether the fracture would heal or not, and if yes, how?

This content is only available via PDF.
You can access this article if you purchase or spend a download.