Most completion engineers use two-dimensional fracture models to design hydraulic fracture treatments. These engineers Also use two-dimensional reservoir models to analyze post-fracture production and pressure buildup data. Normally, a completion production and pressure buildup data. Normally, a completion engineer will concentrate only upon a single zone and will often design a well completion based upon the properties of that zone as computed from log data, core data, drilling data, or other information about the interval.
Once the interval has been evaluated, the completion engineer will pick the location of the perforations based upon the location of the net pay zone. Estimates of porosity, water saturation, and net pay thickness are used to determine where to shoot the perforations pay thickness are used to determine where to shoot the perforations and to estimate the value of fracture height that will be input into the two-dimensional fracture design model.
By using a two-dimensional fracture design model, the completion engineer is dictating the fracture height and the fracture shape. If the engineer uses the wrong fracture height, then the calculated values of fracture length and width will also be wrong. Most of the time, an engineer tends to underestimate fracture height and, thus, overestimate fracture length.
Since 1983, the Gas Research Institute (GRI) has sponsored research concerning how a completion engineer can use three-dimensional fracture propagation and reservoir simulation models to design hydraulic fracture treatments. The GRI research program has focused on how one can develop permeability-thickness program has focused on how one can develop permeability-thickness (kh) data, porosity-thickness (oh) data, and in-situ stress (o) data to describe the three-dimensional properties of a complex gas reservoir. By determining the values of these parameters in various layers of the formation, one can develop a kh profile, a oh profile, and a (3, profile; these profiles are needed to run three-dimensional models. The kh and oh data can be used to describe (1) the location of hydrocarbons in a complex system and (2) where maximum flow rates will be achieved from these layered gas reservoirs. The o, data can be used to determine fracture growth patterns within the complex, layered reservoir.
Using all three profiles in a three-dimensional fracture propagation model, one can determine the shape of a hydraulic propagation model, one can determine the shape of a hydraulic fracture versus the location of the perforations at the wellbore. If the completion engineer makes multiple computer runs with the fracture design model using different locations for the perforations, one can determine the effect of perforation location on created and propped fracture dimensions. propped fracture dimensions. One can then input the propped fracture dimensions into a three-dimensional reservoir simulator, and compute values of flow rate versus time for different perforation schemes and fracture treatment volumes. The 3-D fracture and 3-D reservoir models can be used to optimize the completion of a given well. The location of the perforations and the specific details concerning the hydraulic fracture treatment can be based upon the kh profile, the oh profile, and the o profile.
This paper presents the methodology required to develop the input data one needs to use three-dimensional models in designing the well completion. After explaining how the various profiles can be determined, the paper presents example profiles that have been obtained during the GRI Tight Gas Sands research project. Two hypothetical examples are also presented to illustrate (1) the importance of locating perforations at the optimum position in the wellbore and (2) knowing the stress profile with precision.
When two-dimensional fracture design models are used, computed values of propped fracture length will normally exceed the apparent values achieved in the field. Estimates of propped fracture length determined by analyzing production and pressure buildup data will often be much shorter than the designed value of fracture length.
There are three possible reasons why actual fractures turn out to be shorter than calculated fractures when two -dimensional fracture models are used to estimate the computed values. These three possible causes are (1) excessive height growth, (2) Poor proppant transport, and (3) insufficient proppant concentration. The proppant transport, and (3) insufficient proppant concentration. The latter two reasons can be easily corrected in the fracture design process. process. P. 119