Refracturing in Low-Permeability Reservoirs

A recent Applied Technology Workshop (ATW) held in San Antonio, Texas, “Refracturing in Low-Permeability Reservoirs,” explored refracturing techniques and overall reservoir effects. The technical agenda was divided into eight segments: keynote, candidate recognition, rock mechanics, refracturing design, economics, case studies, refracturing unconventional reservoirs, and poster sessions.

Kicking off the ATW was Chris Wright, President of Pinnacle Technologies, who brought participants up to date on where and how this domain surfaced, the various methods of candidate selection, and several brief case histories. While the technique of refracturing wells began in the 1940s, only a small number of wells have actually been refractured. The industry’s tendency to refracture poor-quality wells has led to a perception that the technique of refracturing may not be successful. There may be as much as 10 Tcf of proved developed nonproducing gas reserves available by means of refracturing. The candidate selection process has improved significantly over the past few years, leading to better understanding and improved results. Various formations have responded to larger fractures, water fractures, low-gel-load fractures, and other methods. The data must be gathered and analyzed and best practices employed to optimize the refracturing design.

Candidate Recognition

Three presentations addressed selection of candidate reservoirs for refracturing. In a session titled “GIS Technology for Refrac Candidate Selection: Process and Example,” BJ Services Applied Geoscience Manager Randy LaFollette described how a Geographical Information System (GIS) is used in the candidate selection process. The methodology is statistical, combining pattern recognition from aerial distribution of production performance and computational analyses/algorithms to provide short lists of refracture candidates. Individual well review occurs only on the “short list” wells as the last step before refracture treatments. A case history was used to explain workflow.

West Virginia U. Professor Shahab Mohaghegh explained candidate selection methodology using virtual intelligence in a session titled “Restimulation Candidate Selection in Tight Formations.” The process moves from field review to well-by-well review. This methodology attempts to merge a comprehensive production data set and analyze the data with reduced subjectivity while addressing the entire reservoir capability within reasonable geographic distance. The methodology is statistical and integrates several production-data-analysis techniques (e.g., type curve/dynamic contact angle). Case histories were discussed at the end of the presentation.

In “A New Refracture Candidate Diagnostic Test Determines Reser-voir Properties and Identifies Existing Conductive Damaged Fracture,” Halliburton Rocky Mountain Tech Team member David Craig discussed a diagnostic test to determine refracturing candidates on both a well-by-well and an interval-by-interval basis. The test requires a short injection above fracture-initiation pressure and an extended shut-in period with the pressure falloff recorded. The test has both qualitative (identify existing fracture damage) and quantitative (permeability, current reservoir pressure, current fracture properties) objectives.

Rock Mechanics

It was suggested by Pinnacle Technologies Chief Technology Officer Norman Warpinski that fracture reorientation could occur from stress-alteration effects, bridging within fractures, and pore-pressure effects from depletion. In his session titled “Refracturing: Concepts for Accessing New Reservoir Volume,” he presented an example of altered stress from the Piceance basin that showed several stress tests performed in an observation well 90 ft from a well being fracture stimulated. Each test indicated a higher rock stress as the nearby fracture treatment progressed.

ConocoPhillips Engineering Fellow Carl Montgomery, SPE Technical Director, Drilling and Completions, shared his experience with refracturing high-permeability chalk formations in the North Sea. The session, titled “High-Rate Acid Refracturing Mechanisms,” provided a different perspective on fracture mechanics, adding understanding of rock’s anisotropic, inelastic, and dilatant characteristics. He pointed out that reopening an original fracture will be difficult because it is plugged with the materials (desirable and undesirable) that have been pumped during original stimulation.

In his session titled “Refracturing—Considering First Principles,” ASRC Energy Services Technical Director John McLennan said he believes that if pore pressures are reduced there is a reduction in stresses that could lead to complex refracture geometry. He also stated that there is an optimum time for refracture stimulation—delaying refracturing allows a reduction in stresses, so as depletion develops far from the well, the reorientation possibility diminishes.

Refracturing Design

Schlumberger Senior Development Engineer Xiaowei Weng presented and explained a modeling concept for fracture reorientation based on the work of Mack and Elbel (1993).1 The production from a fractured well could cause stress reorientation near the initial fracture plane, inducing a new fracture orthogonal to the initial fracture. Three distinct periods of fracture propagation during refracturing treatments were described. Also, part of Weng’s presentation, titled “Effect of Production Induces Stress Field on Refracture Propagation and Pressure Response,” comprised evaluations of two field cases in the Barnett shale comparing the model prediction to the actual-treatment pressure response. Comparison also was made with tiltmeter observations. The implications for refracturing design and candidate selection were discussed.

BJ Services Region Engineer Jason Miller, in “Fracturing Fluid Viscosity Impact on Codell Refracs,” explained the fracturing-fluid evolution and the methodology to measure well performance in the Wattenberg field. Hundreds of wells were studied, and it was found that the fracturing-fluid viscosity profile (defined by the rate of viscosity buildup, peak viscosity, and rate of viscosity degradation) is the main parameter for the success of the refracturing treatment. A correlation algorithm that was based on several key reservoir and completion parameters was developed to successfully predict the expected performance for each well after the refracturing treatments.

In “The Evolution of Producer Refrac and Injector Refrac Designs—Kuparuk River Field, Alaska,” Carbo Ceramics Senior Staff Engineer Terry Palisch explained the candidate selection and design methodology for producer and injector wells in the Kuparuk River field, Alaska. Palisch connected the design methodology for these moderate-permeability oil wells to the design methodology for gas wells with low permeability. The presentation emphasized that the success of a refracturing program first lies in the candidate selection methodology. In refracturing treatments, good wells make good results, and poor wells make poor results.

Tiltmeter Mapping

Pinnacle Technologies Senior Staff Engineer Steve Wolhart’s presentation “Mapping of Refracture Treatments” (addressing vertical- rather than horizontal-well fracture mapping) described current methods used to map fracture/refracture treatments, citing examples from the Lost Hills, Vann, and Newark East fields to illustrate the phenomena. Surface tiltmeter mapping is used in various field areas to identify fracture orientation, length, and drainage area to allow for better fracture-treatment design. According to Wolhart, tiltmeters help to show that refracture treatments grow in a different azimuth from that of the original fracture treatment. Microseismic mapping is perhaps more accurate than surface tiltmeter mapping, allowing for determination of fracture height and direction.

“Production Allocation for Commingled Wells” by Schlumberger Engineering Adviser Bob Poe described a method of using a pulsed-neutron-logging tool to help identify the -contribution to flow from various sands in a commingled completion. It is possible to quantify the contribution from zones producing “behind pipe” when the tubing is run below the top of the perforated interval. By running multiple pulse neutron logs over the life of the well, individual sands that were not con-tributing to the well’s production could be identified and targeted as refracturing candidates. If these sands were found to be nonproductive, they could be avoided altogether in future treatments.

Modeling that refutes the probability of fractures being reoriented in formations because of pressure depletion created by the initial fracture was also described. Poe’s work indicates that it would take a long time to see the sort of stress reversal required to create favorable conditions for reorientation.

In “Novel Technology To Diminish Risk Associated With Refracturing,” Schlumberger Production Enhancement Group Manager Wayne Rowe described a method that uses a borehole sonic tool to identify the stress regions within the borehole. This information, along with the economic criteria, can determine the risk/rewards of refracturing sands that may have some depletion from offset producers. This technique allows the quantification of the maximum horizontal stress, which previously was not determined without costly formation-sampler testing. The benefit of this approach is identification of those sands that have been significantly depleted in offset wells, giving the option of not perforating these sand sections and removing thief zones from initial fracture treatments.

Case Studies

The ATW offered four low-permeability-reservoir case studies. BJ Services Region Engineer Mark Shaefer, in a presentation titled “Green River Basin Refracs,” described how the brief G-function test could help identify reservoir properties and how this helped design fracture treatments for the Frontier formation. This analysis helped change the design from large sand fractures to somewhat smaller treatments with manmade proppants, giving better results. The reciprocal productivity-index analysis technique was used to progressively improve treatment design and well performance.

The session “Refracturing With Lightweight Proppants” by BJ Services Regional Tech Manager Mark Malone described the attributes of lightweight proppants. The main focus was that you could get better transport in the fracture with a lower-gel-loading fluid without significant settling. This results in a cost reduction in fluids. Malone described a sand-banking model run at the U. of Oklahoma and how less banking was observed with the lightweight proppant in slickwater than with sand. He also described the phenomenon of partial monolayer formation and pointed out the potential for improved flow with the layering. He then discussed several case histories in which significant improvements in fluid rates were achieved with the lightweight proppants.

The session by Kerr-McGee Rocky Mountain Corp. Asset Manager David Howell, “Refracturing Codell Formation,” described the success that Kerr-McGee (KMG) has had in its refracturing efforts in the basin-centered resource play of the Wattenberg field in Colorado. Working with several companies, KMG has identified several factors that contribute to the success of the refractures (e.g., water-block removal, degradation of initial fracture, and weak initial proppant). Current estimates are that reserves are being added with the refractures (125 MMcf/well/refrac) and that the industry could add as much as 185 Bcfe through the refrac process.

In his talk “Codell Tri-Fracs in Wattenberg Field,” Noble Energy Operations Engineer Michael Zoll outlined a process in which wells that would benefit from the “tri-frac” technique (third frac on the same sand) were identified. Singling out innumerable factors that could affect the outcome of the “tri-frac” analysis, Noble settled on a complicated algorithm that required constant adjustment to model actual results.

Refracturing Unconventional Reservoirs

The Nick Steinsburger talk, “Refracturing the Barnett Shale,” covered a circumstance in which hydraulically fractured vertical wells on 80-acre spacing were draining only 4% of the gas in place. Crosslinked gelled water fracturing fluids promoted simple fracture growth and inhibited connection to the anisotropic natural fracture system characteristic of the Barnett shale. Fracture mapping from microseismic measurements in observation wells confirmed this concept.

BJ Services Region Engineer Devin Brown presented “Refracturing the Fruitland Coal.” Fruitland coal formation refracture treatments are typified by high fracture gradients (i.e., 1 psi/ft or greater). There were 138 refracture treatments performed for BP in a program begun in 1998. In 2001, the per-well improvement from the refracture treatment was 130 Mscf/D; in 2004, because of improved candidate selection and treatment design, per-well improvement from the refracturing treatment increased to 350 Mscf/D.

A talk titled “Evolution of a Refracturing Program in the Niobrara Formation, DJ Basin,” by Halliburton Energy Services Technical Manager Mike Eberhard, summarized a treatment program in the Wattenberg field, where there are more than 14,000 wells in the DJ basin, of which many are good prospects for recompletions and refracture treatments in the Niobrara formation. A typical refracturing procedure is to set a sand plug and clean out the well to the bottom of the Niobrara. The Niobrara intervals are reperforated with 6 spf, the well is fracture treated, and the Niobrara is produced by itself for at least 2 months. The well is then cleaned out with a foam unit to restore production from the lower intervals.

The talk, “Dendritic Refracturing of the Austin Chalk,” by BJ Services Region Tech Manager Gary Schein highlighted a program in the naturally fractured Austin chalk reservoir. Originally, most wells were treated with crosslinked gelled water and large volumes of sand. Later, debris recovered from wellbores was found to be calcite-cemented fracturing sand, suggesting that the proppant pack degraded with time. Response to retreatment with large-volume slickwater fractures was favorable, displacing oil from the small pore system characteristic of the matrix of the limestone into the natural-fracture network. To enhance the latter effect, wells were left shut in for several weeks following the treatment to enhance the imbibition process. Another benefit of refracture treatments was the rubblization of the salt-/scale-damaged proppant pack. Popular use of benzoic acid flakes proved most effective as a diverting medium. Injection rates have greatly increased over time to more than 100 bbl/min.

Reference

Elbel, J.L. and Mack, M.G.: “Refracturing: Observations and Theories,” paper SPE 25464 presented at the 1993 SPE Production Operations Symposium, Oklahoma City, Oklahoma, 21–23 March.