Geo-engineered completions improve recovery in unconventionals

Applying engineered workflows to unconventional reservoirs has improved production by applying stress variability/contrast or MSE to generate fracability indices to group stages with similar rock characteristics. Multi-stage completion designs have quickly evolved along a similar path. Horizontal well IP rates and EURs have improved, using these integrated completion designs.  

To improve efficiencies, a service provider introduced geo-engineered completions, based on cross-functional expertise and software to integrate petrophysical, geomechanical, drilling and production data into a completion design. Although, these geo-engineered completion designs evolved from engineered workflows, they combine multiple inputs to optimize stage length and perforation cluster positioning. 

In cases where geo-engineered designs were used, wells achieved superior production results, compared to geometric, high-intensity plug-and-perf (PNP) designs. They also displayed improved EURs over other wells with increased lateral lengths, proppant loading and stage counts. In one case, a geo-engineered design, with fewer stages and clusters, achieved higher production than offsets while injecting less proppant and fluid, lowering completion costs.

INTRODUCTION  

Operators working North America’s unconventional plays have implemented a combination of new techniques and completion technologies to increase yields. These include longer laterals used in conjunction with multi-well pad locations. Completions have evolved into “mega-fracs,” that utilize high volumes of lightweight sand/proppant along with slickwater and hybrid fluid systems, deployed in multiple stages and perforation clusters. These trends have improved EURs and net present values in all basins, but most noticeably in the Permian, Fig. 1. 

Fig. 1. Evolution of typical production curves, Midland basin. Source: Pioneer Natural Resources.

Yet, against a backdrop of higher fracture intensity, lower completions costs, and improved average well EUR, recovery factors for unconventional horizontal wells remain in the 3%-to-10% range. After more than a decade of unconventional development, operators have determined that each shale formation has a unique set of attributes. The characteristics that drive production in Bakken wells are different from those driving output in Wolfcamp wells. As such, Bakken completion practices should be different from those used in Wolfcampian formations. Going forward, operators must capture and utilize all available formation attribute data, to improve IPs and recovery factors. 

GEO-ENGINEERED COMPLETIONS

An early success using a geo-engineered completion was documented in a Woodford shale well in Logan County, Okla. The automated workflow resulted in 16,700 additional boe being produced during the first 60 days, compared to an offset that was completed using a purely geometric design, Fig. 2.

Fig. 2. An early geo-engineered completion significantly increased IP in a Woodford shale well.

Based on the original work, an updated geo-engineered design was conceptualized, which integrated perforation cluster/sliding sleeve and packer placement, frac simulation modeling and the use of public domain data to determine the best existing practices for fluid and proppant placement. The work enabled the method to gain additional technical credibility within the industry.

Incorporating new elements. The validity of any completion design methodology centers on the accuracy, efficacy and repeatability of the workflow. Although comprehensive in the number of inputs capable of being integrated into the model, the original methodology lacked several elements desired by operators’ asset teams. The recommendations for improvement included optimized perforation cluster placement using the same data inputs, and a side-by-side comparison of stage designs (e.g. geo-engineered vs geometric). Additionally, a hydraulic fracture simulation using the recommended positioning of stages and clusters, and an associated stage-by-stage treatment schedule, also were proposed as new design elements. 

Fig. 3. A log shows reservoir, geo-mechanical and composite indices, completion tools, stage comparisons, stimulation modeling and treatment schedule.

A single integrated planning tool was developed that contains a large and diverse data set, Fig. 3. The log is intended for horizontal multi-stage completion design, and was created in an attempt to make stage and perforation cluster decision-making easier for cross-functional asset teams. If any particular data set is not available (e.g., cross-dipole sonic or micro-image log) a synthetic proxy can be inserted or a substitute track, such as hi-resolution gas ratio chromatographic analysis, can be used.

PERFORATION PLACEMENT 

Work by Wutherrich (SPE 155485) outlined three considerations for optimizing perforating strategy when using a limited entry approach:

• Perforations should not cause excessive flow restriction. 
• Perforations are optimally designed when every set of openings receives an equal volume of fluid, to ensure connection with the reservoir.
• Placement strategy should provide sufficient reservoir drainage.

Later work by Baihly definitively made the case for understanding lateral heterogeneity prior to adopting a horizontal completions design. Additionally, industry experts discussed the importance of optimal positioning of perforations within stages in areas of similar stress, so that induced fractures can be distributed effectively. Geo-engineered completion designs account for similar stress considerations, as well as possible interaction with natural fracture networks. 

Fig. 4. By minimizing intra-stage variability using the reservoir (track 2) and geomechanical (track 3) information enables engineers to optimize perforation placement. Track 4 shows the presence of natural fractures, with red circles, indicating fractures that might be reactivated.

Automated stage algorithms and workflows result in a geo-engineered stage design, where the length and placement of the stages is based on minimizing the intra-stage variability of fracture potential, based on a combination of reservoir and geo-mechanical attributes. However, given the heterogeneity typically observed in these attributes along the length of a deviated or horizontal wellbore, a degree of variation is still observed within each stage. This remnant intra-stage variability can be used as the basis for the placement of the perforation clusters within each geo-engineered stage, Fig. 4.

The presence of natural fracture networks, and their current stress state, can be important attributes in choosing the location of perforation clusters. The question, “are natural fractures good or bad?”, arises frequently in discussions regarding completion design. The answer is that they can be both. Natural fractures can provide points of “weakness” that promote the initiation of hydraulic fractures, and can be targeted to promote fracture height and half-length. However, if these fractures are open, or if their stress state is such that they will be reactivated when the pore pressure is increased during hydraulic stimulation, these points of relative “weakness” can cause the hydraulic fractures to propagate out-of-zone and potentially connect with adjacent aquifers above, and more deep-seated gas rich formations, resulting in high water and gas cuts during production.

It also may be desirable to target intervals without natural fractures, and fractures with a low or no possibility to be reactivated within a stage to promote complexity in the hydraulic fracture network in “fresh” rock. The presence and location of natural fractures and faults can be interpreted from seismic, and confirmed using a variety of formation evaluation techniques. Once the presence or absence of natural fractures has been confirmed, their current stress needs to be evaluated, together with an assessment of the likelihood that they will be reactivated. 

Fig. 5. Sonic shear data is used to understand the stress network and determine which natural fractures will contribute to production and fluid movement.

An accurate estimation of in-situ stress magnitude and direction is a prerequisite for robust and reliable geomechanical analysis. This method combines fracturing and image log data for determining the complete stress field, using data from a single wellbore. Fracturing data (leak-off tests) and image log information (tensile fracture characteristics) are utilized to determine the in-situ stress field, using a novel inversion algorithm capable of efficiently sampling from the complete in-situ stress solution set conforming to the input data and satisfying a priori domain constraints. In-situ stress magnitude and direction can then be used as inputs to critically stressed fracture analysis, to evaluate the probability that natural fractures will be reactivated during hydraulic stimulation, Fig. 5.

In the absence of a natural fracture network, the placement of perforation clusters can be based on minimizing the intra-stage contrast in other reservoir and geomechanical attributes, such as closure stress gradient. Placement of the perforation clusters in Stage 4 (Fig. 4) was based on the absence of natural fractures (track 4), and then minimization of the contrast in closure stress gradient across the three perforation clusters.

While a degree of automation can assist in selecting the location of individual perforation clusters, a detailed workflow, in combination with human intelligence and experience, can result in repeatable and verifiable perforation cluster design.

IMPACT OF PUBLIC DATA

The public data obtained from outside an operator’s normal channels can be used to determine which completion practices are the most successful in a specific area. Public databases in nearly every state with drilling activity allow free access to view permits, completions, and production records for all wells. In addition, access to the chemical disclosure registry database, provides details on proppant, fluids and chemicals used in completing each well. Since 2011, information on more than 112,500 horizontal wells has been made available to the public on FracFocus.

The information supplied by business intelligence firms comes, in many cases, with analytical tools, such as maps, area search functions, EUR estimations, and decline calculations. This method of preparing for a completion design takes much of the guesswork out of deciding which proppant loading to consider or fluid type to use. More importantly, because the data are tied to production records, it is also possible to know which combination of proppants, fluids and chemicals yields “best-in-class” performance within a particular area. Production and proppant loading are also provided in formats normalized for comparison of wells with varying lateral lengths. These databases are available in the U.S. and Canada, and there are plans to provide similar business intelligence in Argentina.

INTEGRATING SIMULATION MODELS 

Most operators have preferred stimulation modeling software. Providers offer a choice of software to enable the exchange of compatible data with clients. However, there are no perfect hydraulic fracture simulators, and all modeling software has some negative features. Examples include mandatory inputs that are difficult to obtain, the inability to make an adjustment to a single-stage element without re-running the entire well model, and inflexibility regarding cluster length/stage length variability or positioning.

Conversely, geo-engineered workflows are designed to readily accommodate changes. They also have rapid processing times and the flexibility to interface with multiple software models. Multiple fracture simulations are run to compare geometric and/or engineered designs with geo-engineered stage plans, or to determine the sensitivity to various cluster positioning concepts. Simulations are often needed to compare potential well performance from mega geometric frac and geo-engineered designs. With many comparisons to model, it follows that run-time speed is a critical factor in simulator selection.

Fig. 6. Different geo-engineered completion designs, integrated with 2D and 3D frac software output, for single and multi-stage symmetric and asymmetric simulations

Simulations are typically included as a part of the frac treatment schedule report. By incorporating the multiple simulation images as part of a horizontal geo-engineered log format, it is easier for teams designing completions to see the impact of stage and perf cluster decisions visually, Fig. 6.

Three tracks in the geo-engineered integrated log format (Fig. 3) depict the hydraulic fracture design after simulations are done. The track immediately adjacent to the stage model image track is an
easy-to-view representation of fluid and proppant treatment. These selections are displayed in the steps depicted, in a color bar chart image in the geo-engineered log. The track adjacent to the proppant and fluid column displays pressure and pumping rate data as part of full treatment schedule.

CASE STUDY

A Delaware basin well was selected for completion in the Wolfcamp formation. Data from four nearby wells were used as input for a trial of a new geo-engineered design. The offset data included a variety of logs, surveys, tracer and production measurements. The well was down-dip to nearby producing wells and was not considered a high-potential candidate.

Fig. 7. A new-generation geo-engineered completion improves recovery compared to similar wells in the same area.

The use of a geo-engineered completion method to select stage lengths, and to position perforation clusters, resulted in the well’s performance significantly exceeding the operator’s expectations and delivering approximately the same hydrocarbon production as another operator’s nearby well. The performance improvement occurred despite stimulating the subject well with approximately 50% less fluid and proppant compared to the offset, Fig. 7.

PATH FORWARD 

Geo-engineered design algorithms and the weighted attribute methodology have contributed to advancing the science and engineering behind horizontal completion design. But the quest to improve recovery factors, EURs and associated net present values in unconventional reservoirs involves at least two recognized challenges: 1) geo-engineering designs and degradable diverters to address perforation cluster and/or sliding sleeve placement; and 2) improving lateral placement, which is being addressed by others within the industry.

Two-pronged strategy. The industry needs to develop an integrated, repeatable workflow that couples lateral placement and completion design optimization. Work done by Laredo Petroleum in the Midland basin outlined the need for a comprehensive earth model. In trying to determine the attributes driving production in the first six months of a well’s life, they found a low correlation with bivariate statistical analysis. It was assumed that no single variable was capable of predicting production. Laredo’s adoption of an earth model coincided with the collection of data regarding 180 geoscience and engineering attributes. That yielded the opportunity to perform multivariate analysis to predict well performance. 

Among the findings, it was revealed that 11 of the 180 attributes identified as part of the multivariate analysis had 97% correlation coefficient with actual production.

CONCLUSION 

The implementation of geo-engineered designs has been successful and represents another achievement in completion design workflows for single-well applications. Advances in horizontal completion design are evolving rapidly and have demonstrated the following:

• Mega-frac jobs with high proppant loading and large fluid volumes work in nearly every basin.
• Geo-engineered completions and degradable diversion improve cluster efficiency.
• A proven 3D earth model is essential in determining accurate lateral landing in sweet spots. 
• Public data have value. The trend toward statistical modeling is helping to determine a reservoir’s unique DNA attributes that lead to optimizing production.
• Determining the factors/attributes driving production requires large amounts of data from multiple wells as input, to perform multivariate analysis for each potential interval.
• In the absence of large amounts of data (exploration settings), the combination of high proppant loading, geo-engineered design and degradable diversion provides logical steps in establishing baseline performance to approximate best-in-class production within a given area.