Hydraulic fractures do not completely close during the first twenty-four hours after the stimulation procedure is completed. Typical reservoir rocks may require as many as two years to close sufficiently that production is not affected by portions of the fracture which do not contain proppant. Further, many reservoir rocks do not close sufficiently during the first ninety days after a stimulation to prevent proppant movement and settling. Thus, the slowly closing hydraulic fracture presents problems in both fracture design and production performance evaluation in the early life of a fractured well. The effects of slowly closing hydraulic fractures are directly observed and measured using traditional pressure and production testing procedures. Radioactive-tracer logging can be used to observe the settling of proppant during the first ninety days after fracturing. Predictions of the closure rate and final crack width can be simulated using finite-element analysis which includes consideration for plastic deformation and creep.
Production performance predictions made in the early life of low permeability and low production rate reservoirs have been historically plagued by initial reservoir performance which indicates a higher reservoir quality than ultimately is observed in cumulative production. The typically observed manifestation of this situation is a high initial production rate followed by a very rapid decline to a "stabilized" production rate and decline. Within fields, the initial production rate and decline are often not consistent between wells and are often not a reliable indicator of ultimate well performance.
Explanations for this behavior often involves consideration of reservoir non-homogeneity, presence of boundaries and permeability changes, connection with in-situ fracturing, and possible long-term damage caused by completion or reservoir fluids. Each of these conditions are possible contributors to unusual well behavior. However, a not-propped, initially high-conductivity fracture which connects productive reservoir with the well bore and which slowly closes completely during a period of several months to years could also explain similar behavior. The slowly-closing nature of the fracture is easily explained by rock creep within the rock after the fracturing stimulation.
Laboratory experiments performed on sandstone and cement paste to measure the pressure response during both propagation and closure of hydraulic fractures created fractures which did not completely close at the end of the tests. This fact was noted and dismissed as not being critical to the study being conducted. The authors did, however, state the same phenomenon of a not-closed fracture may also occur in the field.
The concept of a slowly closing fracture was tested in several reservoirs using production testing methods which permit determination of producing reservoir transmissibility, effective fracture length, and skin damage. This study demonstrated the effective reservoir transmissivity and fracture lengths often decrease during the first one-to-two years, especially in low permeability reservoirs. A slowly-closing not-propped fracture which gradually reduces the reservoir interval and infinite conductivity fracture length connected to the wellbore provides a reasonable explanation for this behavior.
The concept was further tested by measuring movement of radioactively traced proppants during the first ninety days after hydraulic fracturing. Wells were logged once during the first week after the fracture stimulation. The traditional belief is the fracture would have sufficiently closed in this time to fully entrap the proppant.
P. 151^
