Over the last five years a number of publications and technical papers have addressed the impact engineered completions can have on the success of shale reservoirs. Early efforts to determine this impact focused on whether the degree of lateral variability warranted the effort and costs associated with an engineered approach. These studies pointed to the mismatch between the wellbore trajectory and geological trajectory, which is inevitable when drilling thousands of feet of lateral section each day.

Lateral variability manifests itself as variability in the stress profile along the lateral. The variations in stress have a direct impact on the completion efficiency within each stage because the frack treatment tends to follow the path of least resistance. If perforation clusters occur within a stage at relatively lower stress, then they will break down early and be overtreated, while the higher stress clusters will be undertreated. This occurrence leads to inconsistent treatment results and a corresponding inconsistency in production from stage to stage and from well to well. This result is referred to as the “statistical nature” of shale reservoirs.

Despite this, the engineered completion continues to be a workflow that has failed to gain universal acceptance. The engineered completion is still viewed as a “science project,” not a mainstream workflow. Some of the reasons for this include:
• The high cost of data acquisition;
• The negative impact on operational efficiency;
• Challenging logistics in a high-volume arena; and
• The difficulty in quantifying the impact of the optimized completions on individual well production.

LateralScience has been designed to address these factors and pave the way for the deployment of engineered completions on every well. It is an innovative, out-of-the-box approach that focuses on a different source of reservoir information—drilling data.

The data being leveraged are common drilling parameters including weight on bit, RPM, torque (TOR) and differential pressure (ΔP). A key advantage to this method of reservoir evaluation is that these data exist on every lateral ever drilled, which directly addresses the first two objections to universal acceptance previously listed.

In 1964 Robert Teale published work that proposed the use of mechanical specific energy (MSE) as an effective way to optimize drilling practices using the drilling data listed above. The MSE parameter defines the amount of work required to drill a unit volume of rock, which will vary according to the geomechanical properties of the reservoir. Generally, completion engineers are not familiar with MSE, but they do recognize the parameter known as unconfined compressive strength (UCS). This parameter is typically measured in a laboratory on core samples and is then used to understand how the reservoir will break down under pressure during a hydraulic fracturing operation. The relationship between MSE and UCS is important; they are related by the drilling efficiency (Deff) of the rig (UCS = MSE*Deff). Assuming that Deff is reasonably constant, MSE can be used as a good qualitative proxy to UCS. This assumption will be reasonable over short intervals of the lateral (i.e., the length
of a frack stage, ~ 76 m [~ 250 ft]).

The work done by Teale successfully addressed vertical wells, but it was insufficient for the horizontal wells, which became popular some 30 years later. The actual equation currently deployed is still the same science, but it had to be adapted to account for the mud motors in the tool string. Teale’s equation relies on the TOR measurement as an important indicator of the reservoir strength, which is typically adversely effected in most horizontal holes. The LateralScience approach leverages ΔP rather than TOR to better define rock strength in the lateral section of a shale well.

Workflow
For ease of use, the LateralScience workflow converts MSE into a facies log, with each facies representing a range of MSE values. The yellow facies is easier to drill than the orange facies which, in turn, is easier to drill than the red facies. While examination of the entire well is helpful, the real work is done one stage at a time. The workflow starts with the original geometric design, shown in Figure 1A. In this particular stage, 10 perf clusters are spaced 7.3 m (24 ft) apart. With the original design, perf cluster two appears to be an issue since it will break down early and ultimately receive a disproportionately high percentage of the stimulation. Figure 1B demonstrates how the perf clusters can be repositioned to alleviate this issue while maintaining a reasonable spacing between clusters. This optimized perf design increases the odds of getting an evenly spaced treatment across at least eight of the 10 perf clusters, which improves the productivity of this particular stage.

This facies-based approach is simpler to deploy than other engineered completion workflows, addressing the concern of being able to deploy this technique in a high-volume, high-efficiency manner.

FIGURE 1. (A) The original geometric completion design with 10 perf clusters is shown. (B) The optimized completion design minimizes the effect of lateral heterogeneity. (Source: C&J Energy Services)

Case study
An important step in validating this approach is to demonstrate that LateralScience can differentiate between a well whose production is adversely affected by lateral heterogeneity and a well that is producing optimally. This case study compares two wells drilled in Ellis County, Okla., in the Lower Cleveland sand, a silty sandstone. These two wells were selected by the operator for evaluation because they were in close proximity (<1.6 km [1 mile] apart) and were drilled parallel to each other (due north). The two wells were completed in a similar fashion and yet had significantly different production results. Both wells were completed geometrically with identical schemes (20 stages, four clusters per stage, stage length of 72 m [235 ft]) and were treated with identical frack programs.

A 1,006-m (3,300-ft) section of the LateralScience facies log from each of these two wells is shown in Figure 2, showing that the LC-1H well has significantly less lateral heterogeneity than the offsetting LC-4H well. The variability in rock strength at each perforation cluster is shown in Figure 3, giving valuable insight into variability in production between the two wells.

FIGURE 2. Two 1,006-m LateralScience facies plots of the LC-1H (top) and LC-4H (bottom) wells show that one well has significantly less lateral heterogeneity than the other. (Source: C&J Energy Services)

Assuming the weakest facies in each stage will be effectively treated, LateralScience predicts that 63 of 80 perf clusters in the LC-1H well were effectively stimulated. The same analysis for the 4H well yields 42 effectively stimulated perf clusters. The perf efficiency is 50% better (63/42) in the LC-1H well, which is in excellent agreement with the actual production data.

In the first year, the LC-1H well averaged 214 bbl/d of oil and 16,169 cu. m/d (571 Mcf/d), while the LC-4H well averaged 135 bbl/d of oil and 9,288 cu. m/d (328 Mcf/d) (58% more oil and 74% more gas). The excellent agreement suggests that the difference in production is primarily due to the effects of lateral heterogeneity and that LateralScience can detect this difference on future wells and enable the engineered completion design.

FIGURE 3. The perf clusters in the LC-1H (left) well show far less variability within each stage. (Source: C&J Energy Services)