The development of theoretical hydraulic-fracturing models started several decades ago, setting a basis for hydraulic-fracture design, optimization, and diagnostics. As a reservoir stimulation technique, the problem of hydraulic fracturing in essence is the prediction of the geometric shape and dimensions (length, width, and height) of the growing fracture as a function of time. This mathematically translates to a set of nonlinear, integro-differential equations with moving boundaries, solved numerically over a number of time-steps (SPE 168600).
When it comes to multistage fracture treatments (Fig. 1), where many fractures are being “pumped” simultaneously within each stage, the problem’s complexity increases exponentially. The old adage of four inputs being necessary for optimal fracture design (reservoir permeabilities, in-situ stress distributions, a sound geological model, and a fluid-loss model in naturally fractured rocks) is no longer the case as multiple-fracture (multi-frac) interactions and interference (intra-stage, intra-well, and inter-well) should also be incorporated (SPE 1210-0034-JPT). Classic treatment-design software, although convenient to use, struggle to predict multi-frac growth geometries with sufficient certainty, leading to significant disparities between simulations and field observations.
Four real-world challenges associated with multi-frac modeling from horizontal well operations are described below. Although each formation presents its own set of challenges rendering it necessary for operators to learn from experience, the development of computational techniques to facilitate the simulation of these common occurrences will yield more-realistic models.
1. Nonsimultaneous fracture initiation within a stage
Fracture-propagation pressures are significantly smaller than fracture-initiation pressures (approximated by the recorded “formation-breakdown” pressures) in the same rock. Hence, earlier fracture initiation from one perforation provides a low-energy pathway for the fracturing fluid, which would “prefer” to enter an existing fracture and propagate it rather than initiating new fractures from nearby perforations.
2. Dominant (or “runaway”) fracture creation
Dominant (or runaway) fractures are those whose growth is significantly larger compared to that of the other fractures. Laboratory-scale experiments where more than one fracture was pumped from a single fluid source displayed dominant fracture creation as a consistent observation (Michael). This observation qualitatively agrees with acoustic sensing data from the field-scale operations (Fig. 2). Dominant fractures end up receiving the entire fracturing fluid by the end of a stage (SPE 194334), while the remaining fractures cease propagating. The contribution of these few dominant fractures in the post-stimulation well productivity could be major.
3. Inactive perforation clusters
Fracturing fluid can bypass entire perforation clusters, initiating fractures from perforation clusters further downstream along the horizontal lateral (SPE 194334). This occurs both by inactive perforation clusters located between active clusters and by inactive perforation clusters located adjacent (upstream or downstream) of active clusters (Michael).
There are many possible causes of a perforation cluster failing to generate fractures, including nonsimultaneous fracture initiation and dominant fracture creation. Limited-entry techniques accomplished by varying perforation numbers and diameters (SPE-530-PA), later proposed for mitigating interference between simultaneously growing fractures from horizontal wells (“stress-shadowing”), can potentially be effective in eliminating inactive perforation clusters if specific patterns are detected.