This paper presents a new hydraulic fracture (HF) tip mechanism observed in a statistically homogeneous porous rock, HF tip mechanisms are widely recognized to be responsible for discrepancies between field observations and HF model predictions. Current proposed tip mechanisms used in HF geometry simulations require field calibration and are not truly predictive based on lab- and/or field-scale material property data. Matching field pressure data often gives significantly different material property values from those measured in the lab. This paper reports studies of a hydraulic fracture created in an 18-inch cubic block of Lueders limestone that reveals random "bridges" connecting the opposing fracture faces. Observation of the bridges suggests the existence of a crack tip zone that resembles the process zone widely reported for various other materials. The existence of a process zone significantly changes many fracture parameters in comparison with linear fracture mechanics predictions (e.g., fracture length and width, net pressure, etc.). A new model is proposed to account for the observed crack tip process zone (PZ). The model is comprised of two complementary parts:
a statistical characterization and simulation of the bridging phenomenon, and
a thermodynamics-based model of the growth of the crack and PZ.
The main challenges and future research topics in Hydraulic Fracturing are also discussed.
Critical evaluation of hydraulic fracture (HF) treatment data in the 1980s has exposed frequent discrepancies between field results and theoretical predictions. The most important for the following discussion are the observations that (1) the measured field net fracture pressure is larger than conventional HF simulators predict, and (2) the net pressure is rather insensitive to rate variation and fluid viscosity. In addition, available direct observations of hydraulic fractures show they may be not a single crack, but a zone of multiple cracks somehow linked together in a large hydraulic fracture zone on a scale much greater than the scale of rock heterogeneity. Although the actual mechanisms of this phenomenon are not yet well understood, these observations suggest that the presence of this tip zone may be responsible for the observed discrepancies between HF simulators and field data.
Fracture treatment designs based on the conventional fracture models required heavy viscous gels and large pads to create a sufficient fracture opening for proppant transport. The models also predicted long propped fractures. The practical implications of using such models were to pump very expensive and conservative treatments with excessive fluid additives and gel volumes. These erroneously optimized treatment designs often resulted in creating fractures shorter than predicted. In addition, the conservative designs with more-than-required clean fluid volumes might cause proppant settling and convection that would leave some productive intervals without proppant and undrained.
Recognizing the limitations of conventional hydraulic fracture models, various net pressure calibrated models and procedures were introduced into field practice. New frac job designs based on such models do not require high-viscosity gels, suggest much smaller fluid volumes can be used to place large quantities of proppant, and predict shorter and wider fractures. Therefore, the tip-calibrated models result in a radically different economic optimization of HF treatments and offer new engineering ideas for technical job improvements.
P. 275
