Table 1

Microseismic Hydraulic Fracture Mapping

Downhole Tiltmeter Hydraulic Fracture Mapping

Mixing Proppant and Fluids Downhole Reduces Risk and Lowers Cost

Waterfrac Success Depends on Reservoir Selection

Table 2 

Damage Removal Using Microbiological Tools

Restimulation Candidate Selection Methodology to be Tested

New Technologies Focused on Squeezing More from Existing Wells

References

Improvements in Fracture Stimulation Technology
by Karl Lang, Hart/IRI Fuels Information Services
Excerpts in PTTC Network News, 1st Quarter 2001

During 1999 the oil and gas industry spent more than $1.5 billion fracture stimulating gas and oil wells worldwide. Approximately $850 million of that money was spent on about 20,000 North American fracs. While hydraulic fracturing is an economic necessity for some wells (e.g., low permeability sandstone natural gas completions), the technique is finding increased application in high permeability sandstone completions offshore (fracpacks), as well as in the completion of disposal wells for oilfield wastes. It is clear, however, that the most important contribution of improved fracturing technology will be in helping to meet the growing demand for domestically produced natural gas.

Improvements over the past two decades have been based on advances made in hydraulic fracture theory and modeling, equipment, fluid systems and proppants. In spite of these advancements, however, the industry’s ability to determine the geometry (length, height and direction) of the hydraulic fractures it creates is still limited. The most important questions asked by any engineer responsible for fracturing are:  Can I be certain the dimensions and direction of the created fracture are optimal? Can I achieve the same (or more) productivity improvement at a lower cost? This article looks briefly at how engineers are looking to new technologies to answer these fundamental questions.

Hydraulic Fracturing Diagnostics

Optimizing the hydraulic fracturing process is difficult because engineers must rely on indirect measurements such as treating pressure or post-fracture production data to infer the results of the treatment. The physical mechanisms that underlie fracturing are not universally agreed on and the various commercial models based on interpretations of these mechanisms can give widely varying estimates of fracture geometry for given treatment parameters (Warpinski, et al. 1994). Gas Technology Institute (GTI – formerly GRI), together with others, has developed commercial fracture mapping technologies that attempt to overcome these limitations by directly imaging the dimensions and direction of a fracture as it is created.

The most common fracture diagnostic techniques (Table 1) differ in the fracture attributes they measure, the certainty of the measurement and whether they provide a far-field or near-wellbore look at the fracture (Wolhart, 2000). Mainstream techniques, either direct or indirect, typically provide only a limited amount of information on far-field fracture geometry. For example, treating pressure analysis and production data analysis can result in non-unique solutions and radioactive tracers, useful in determining whether a zone accepted stimulation fluid, provide only a near-wellbore look at minimum fracture height. Two new technologies, microseismic fracture mapping and downhole tiltmeter fracture mapping can help engineers answer the questions: Does the fracture cover the pay zone and is it confined to the pay? What are the optimum number of stages, treatment size and proppant specifications? What is the fracture azimuth, dip and length and how do they affect choice of well direction and spacing?

Table 1: Hydraulic Fracture Diagnostic
Techniques (modified from Wolhart, 2000)  

 " = High Certainty          ! = Low Certainty

Microseismic Hydraulic Fracture Mapping

GTI’s fracture diagnostics program, carried out in conjunction with The Department of Energy (DOE), was started in 1993 using the GTI/DOE M-Site (located in the Piceance Basin) as a field laboratory where GTI could develop, test and verify hydraulic fracture diagnostics, mechanics and theory on a field scale (Peterson, et at., 1996). What emerged from the work done at M-Site were two technologies for measuring far-field fracture geometry, each of which is described in more detail below.

GTI’s FRACSEIS^SM microseismic fracture mapping service (now offered by Pinnacle Technologies Inc.) uses seismic sensors placed in an offset well to detect microseisms (micro earthquakes) generated during treatment. The location of these microseisms is recorded and used to create a real time image of the hydraulic fracture (Warpinski, et al. 1998). A formation is stressed during a hydraulic fracture treatment because of leakoff-induced pore pressure increases and net treating pressures. This leads to small shear slippage’s similar to earthquakes along faults, although with much lower amplitude. These micro-slippages (microseisms) emit elastic waves that can be detected by geophones. The microseisms are located and the data is used to create maps of the hydraulic fracture geometry. There are two ways to gather the microseismic data. The first uses seismic receivers located in two or more offset wells. The second approach employs a vertical multiple-array of receivers in a single offset well. To date, about 60 fracture treatments have been mapped in the United States, Canada and Mexico.

Downhole Tiltmeter Hydraulic Fracture Mapping

Another new technology for hydraulic fracture mapping employs downhole tiltmeters (Warpinski, et al., 1999). As with microseismic monitoring, multiple instruments (tiltmeters in this case) are run on wireline in an offset well to measure the earth’s tilt due to the hydraulic fracture. Creating a hydraulic fracture involves parting the rock and deforming the reservoir. Downhole tiltmeter mapping simply measures the fracture-induced deformation in an offset well (or wells) versus time and depth (Wright, et al., 1998). A tiltmeter is in effect a very sensitive “carpenter’s level” that can detect tilts down to one nanoradian. The data is inverted to obtain created fracture height, length and width. To date, downhole tiltmeters have been used to map over 400 fracture treatments in multiple basins in the United States, Mexico and Canada. This technology is now being used by Pinnacle Technologies, Inc. to field a new variation of the downhole tiltmeter system that can be deployed on wireline within a treatment well itself.

Treatment well tiltmeter mapping involves running an array of downhole tiltmeter instruments (typically 3 to 10 tools) in the treatment wellbore prior to the fracture treatment. The tools are coupled to the wellbore with magnetic or bowspring centralizers. A fluid-only treatment (no proppant) is pumped (waterfrac, acid frac or minifrac) and the induced tilt at each location is measured in real time and telemetered to the surface.

“Deploying a tiltmeter array in the well being treated, rather than in an offset well, makes this technology much more accessible to the smaller operator trying to do a better job of fracturing,” says Kevin Fisher, Business Development Manager with Pinnacle. “Our tool can be used during a pre-treatment minifrac to obtain an actual fracture height and width measurement that can then be used to fine-tune the fracture model for the main fracturing treatment. The tool can also be run during an actual treatment, so long as proppant is not used.” Measuring fracture height directly in the treatment well eliminates the need to shut in production on monitor wells and allows the tiltmeter technology to be used in areas where well spacing is not close enough for offset well mapping. Through January 2001, Pinnacle has run the new array during five minifracs and fracs in California diatomite wells, and also during an acid frac in Oklahoma’s Hunton, with good results. The tiltmeter signals obtained from the treatment well array are up to 1,000 times greater than those obtained in offset wells.

Both microseismic and downhole tiltmeter fracture mapping techniques are capable of providing real-time measurements of fracture geometry. Using these techniques to fine-tune hydraulic fracture models in real-time will allow engineers to make decisions in the field on how to change rates, adjust fluid properties, change proppant schedules or adjust pad size in order to optimize the frac job while it is being pumped. “Our goal is to use the fracture mapping data to calibrate the model, so the model can be used to accurately predict fracture behavior, rather than to simply explain treatment results,” says Fisher.

Mixing Proppant and Fluids Downhole Reduces Risk and Lowers Cost

Pumping fracture treatments at high pressures is expensive and increases the likelihood that casing, tubulars or equipment might fail, creating a safety hazard. For these reasons, older wells are often rejected as candidates for high pressure treatments; the payout may be good but the risk of rupturing the casing or tubing is significant. Any fracturing procedure that would allow an operator to perform a treatment at lower pressures could save money and increase reserves. A DOE-sponsored project in New Mexico has shown that mixing fracture fluids at the bottom of the well, rather than on the surface, may lead to a better, safer, lower cost technique for fracturing wells. Another benefit of this approach is an increased ability to alter the treatment mixture at the perfs, during the treatment.

RealTimeZone Inc. (RTZ), of Roswell, NM, used their downhole mixing technique for the first time last November in a 12,300-foot Morrow gas well in the Sand Point field of Eddy County, NM. According to George Scott, a manager at RTZ, surface treating pressures were even lower than expected. “The treatment consisted of a methanol gel with 7,000 pounds of bauxite proppant pumped down the annulus and 40 tons of liquid CO₂ pumped down the tubing,” said Scott. “Tubing pressure never got above 6000 psi, and the casing side was never above 5000 psi. Pressures averaged around 5000 psi, but if we had pumped the job in the conventional manner, the pressures would have averaged closer to 10,000 psi.” Liquid CO₂ was used because after the proppant has been placed, the drop in treating pressure turns the CO₂ from liquid to gas, allowing the fracturing fluid to be produced back from the formation at a faster rate. The Sand Point well, which had a relatively small interval of low permeability pay and had been scheduled for abandonment, has been producing between 200 and 250 Mcfd since the job was completed last year. A post-fracture tracer log showed that the treatment had been placed in the zone as designed.

“This is not rocket science. We simply developed some algorithms that allow us to control the mixing of the two fluid mixtures downhole by varying the tubing and casing pump rates at the surface,” explained Scott. “The value of this approach is two-fold: lower friction pressures mean less hydraulic horsepower and thus less cost, and the surface control of the downhole mixing permits a degree of real time influence over the treatment that isn’t possible when you’re mixing the treatment fluids at the surface. This can mean the difference between success and failure, particularly in wells that have a tendency to screen out prematurely.” Changes in stimulation pressures monitored at the surface also allow an operator to know if the fracture is being created as planned. If necessary, the operator can change the fluid mixture to ensure that a fracture goes in its intended direction. “There are other possibilities that become apparent,” says Scott. “For example, an ultra-clean borate gel fracturing fluid cannot be coupled with CO₂ under ordinary circumstances, so the benefits of these two different elements, a non-damaging fluid and a more effective clean-up process, are mutually exclusive. But if we mix the gel and the CO₂ downhole at the perfs, it may be possible that they can be used together.” Further field testing is needed to confirm this concept.

RealTimeZone can also combine this downhole mixing methodology with a downhole, real-time, surface readout fracture monitoring system to give an even more accurate picture of where the fracturing fluids are going. Working with Halliburton Energy Services to incorporate their gamma-ray Spectrascan log, RTZ has performed a treatment in another Eddy Co., NM well completed in the Willow Lake Delaware oil reservoir, and has more scheduled for the near future. Spectrascan utilizes distinctive radioactive tags (typically encapsulated tracers provided by CoreLab’s ProTechnics) on both proppant and fluid to reveal the relative distribution of pumped material. By monitoring the distribution of fluids during the treatment, a picture of where the fluid and proppant are going and what might need to be done to alter their direction is possible. This approach currently employs a wireline tool positioned above the perforations. RTZ is working with Halliburton on the experimental design of a procedure employing a second gamma-ray tool that will use wireless telemetry to monitor the below-zone region. This should permit a complete, real-time picture of the entire treatment interval.

The Energy Department's National Energy Technology Laboratory (NETL)– the primary field office for the department's Fossil Energy research program – began working with RealTimeZone on the hydraulic fracturing project in May 1999. RTZ developed a real time stimulation diagnostic system in conjunction with Halliburton (with some minor help from Schlumberger) several years ago. “The DOE’s National Energy Technology Lab (NETL) was familiar with the patent we co-authored with Halliburton for real time stimulation monitoring and contacted us a little over two years ago," Scott said, "and they basically wanted to fund continued development of the system.”

With a total project cost of $1.3 million, and with the federal government contributing $922,000, the project is now in its last two phases involving field testing the downhole mixing and real-time monitoring methodologies. The project is scheduled to end in June 2002. DOE is looking for interested operators with wells to be treated and will provide co-funding dollars.

The Permian Basin is a high-potential area for implementation of the downhole-mixing methodology and real-time monitoring. First of all, there are a lot of older wells. Also, many water zones occur in close proximity to producing zones and larger fracture stimulation treatments can easily fracture right into the water, ruining the well. In deep formations such as the Morrow, engineers are restricted on what can be pumped due to the extreme pressures, and any method that decreases treating pressures can be valuable. “We have a number of jobs scheduled, and our ability to successfully carry out these jobs should continue to develop over the next two years,” says Scott. “Eventually, we feel this technique will allow a large number of otherwise untreatable wells to be successfully fractured. If operators can fracture deep wells more economically, they’ll have more money as well as more candidates for fracturing.”

Waterfrac Success Depends on Reservoir Selection

Ten or more years ago, as U.S. natural gas producers began to look seriously at unconventional gas resources (tight sands, gas shales, coalbed methane) as a means to replace gas reserves, the well-stimulation industry was focused on designing more cost effective ways for carrying and placing proppant into hydraulic fractures. Conventional fracturing technology was employed in early efforts to exploit low permeability tight gas sands and fractured gas shales, but in many cases the high cost of fracturing treatments made the reserves uneconomic. In the mid 1990s, several cost-conscious operators tried lowering costs by reducing the proppant concentrations in large treatments and using less expensive, lower viscosity fluids. In several basins the results were impressive: significantly lower costs and as good or better well productivity. The term “waterfrac” (also called light sand fracs) was applied to a treatment similar to what had been years before termed a “slick water” treatment. While waterfracs are not applicable in all situations, proper candidate reservoir selection can help determine where they are appropriate.

A typical waterfrac includes:

• large volume of water (1,000 to 2,500 bbl/foot of gross pay) treated with friction reducers, surfactants and clay stabilizers

• 50% pad with constant 0.5 ppg sand concentration

• tail-in with 0.5 to 2 ppg for last 1% to 5% of treatment, to ensure good communication between the fracture and wellbore area.

Variations include: use of linear gel, sand concentrations as low as 0.25 to 1 ppg, “sweeps” without sand, different pad sizes, and no tail-in with higher concentrations.

According to Mike Mayerhofer, Staff Engineer with Pinnacle Technologies in Houston, “Waterfracs appear to work through a combination of mechanisms. The walls of the fracture are rough and when the surfaces are slightly offset by shear forces there is a “self-propping” phenonenon that may be taking place. Rock debris in brittle formations can enhance this behavior.” Mayerhofer adds that with low proppant concentrations, proppant can settle out in vertical offsets, forming proppant bridges that act to “pin” open the fracture, and that clean-up problems that can be associated with conventional high-viscosity fluids are absent.

The performance of waterfraced wells has been reported on by a number of operators: Mitchell Energy in the Barnett shale of the Fort Worth Basin; Union Pacific Resources and others in the Cotton Valley of the East Texas Basin and in the Austin chalk of South Texas; and Anadarko in the Bossier sand of East Texas (Table 2).

Table 2: Recent Applications of Waterfracs

Experience has shown that there are some general guidelines that can be followed when considering waterfracs. According to Mayerhofer, “Operators should look to apply waterfracs in the most marginal, lowest permeability areas first and then proceed to better areas. Naturally fractured, “stiff” rocks (with a high Young’s Modulus) in normal stress environments are good candidates.” When designing a waterfrac, Mayerhofer suggests that, “Optimization with respect to fluid volume, injection rate, pad size, proppant concentration, etc. will be field specific. When testing to see if waterfracs make sense, be sure to set up an “apples to apples” test that allows a good comparison with a conventional treatment, and consider quantitative fracture diagnostic techniques to measure fracture geometry and quality.” 

Not all recent applications of waterfracs have been successful. Attempts to use waterfracs in the Travis Peak sands in East Texas were considered to be unsuccessful due to the higher permeability of that formation relative to the Cotton Valley (Brister, 2000). However, their success in particular situations indicates that operators who can find the right “niche” for this approach can profit.

Damage Removal Using Microbiological Tools

The same rheological characteristics that allow fracturing fluids to carry proppant into a fracture can, in some cases, cause damage to a producing formation by decreasing permeability and blocking flow. In severe cases the damage can kill what would otherwise have been a productive well. One new approach to remediating this damage is the use of biologically-generated chemicals.

Water-based fracturing fluids are most commonly composed of borate crosslinked guar gels. An important part of any fracture treatment is reduction of the gelled fluid’s viscosity after the frac treatment has been performed so that produced fluids can flow through the proppant matrix. Chemical or enzymatic agents that degrade the polymer structure or cross links in the gel (termed “breakers”) are used to accomplish this.

In a significant number of fracturing operations, the breaking process is incomplete, resulting in less than optimal flow. A new strategy for repairing such damage is the use of biological culture products specifically targeted to degrade the gel polymeric structure. One such product, produced by Micro-Bac International of Round Rock, TX, and called Gum-Bac™, is specifically designed to degrade the carbohydrate backbone of guar gels. This degradation reduces viscosity, increases polymer fragment solubility and promotes removal of the gel from the fracture matrix.

“The results of Gum-Bac applications have been very encouraging,” says Brian Cummings with Micro-Bac International. Cummings cites a well in southwest Texas that had been treated with an acid fracture stimulation fluid incorporating a complex copolymer, where after treatment production could not be restored. Micro-Bac treated the well with Gum-Bac to degrade the copolymer and a large amount of polymer material was immediately freed up. “The well produced over 14 MMcf during the first month, declined steadily over seven months to about 3 MMcf per month, then dropped further as it began to produce “slick” water (indicative of polymer),” added Cummings.

There are specific benefits to choosing biologicals over chemical or enzymatic agents. For example, the microbes used are motile and thus can migrate within the formation to degrade gel in regions well beyond that reached by simple fluid displacement. Also, the continuous in-situ production of polymer-degrading substances by the microbes allows for a greater likelihood that a maximum amount of residual occluding gel will be contacted, degraded and removed.

Micro-Bac International’s Para-Bac, introduced in 1986, was the first biological product to successfully improve production in paraffin-clogged wells. The action of Para-Bac is the result of natural by-products (biosurfactants, organic acids, ketones, and alcohols) produced in-situ by the organisms’ metabolism. By combining bacterial strains that work in harmony, the treatment is able to accomplish specific goals. Today, many operators are using microbiological alternatives to conventional paraffin treatment methodologies such as hot-oiling, steam and chemicals. The application of biologicals to similar problems with fracturing polymers is a logical extension of this technology. Micro-Bac has recently introduced the Custim™ product line, the company’s second generation of oilfield biological products for well stimulation.

Restimulation Candidate Selection Methodology To Be Tested

“If it ain’t broke don’t fix it” may be a good adage in most circumstances, but when it comes to fracturing it may be the reason the industry has such a poor perception of restimulation as a method for improving well performance. Operators are reluctant to restimulate a well that is producing reasonably well, even if it stands the best chance of significant incremental production gains. According to work carried out by the Gas Technology Institute, many operators make restimulation decisions without benefit of downhole data and base their decisions only on production information, which may not be a good indicator. The industry tends to restimulate only the “broke” (worst performing) wells and not those that have the most potential for improvement. Gas Technology Institute (GTI) has found that refracs are only 2-3% of total US fracture stimulations, and most of these involve gas wells in the US Midcontinent, Rocky Mountains, and South Texas regions. There may be significant potential, however, if operators can apply a successful methodology for identifying candidate wells.

GTI is finishing up a two and a half year study on the best way to select restimulation candidates. The study focused on developing an efficient candidate-selection methodology and included demonstration tests in three regionally dispersed tight gas fields (Reeves, 1999). Results indicate that roughly 15% of the wells in a field represents 85% of the restimulation potential. After analyzing 200-300 wells in each area, GTI refraced two wells in the Mesaverde formation in the Piceance basin of Colorado, four wells in the Frontier formation of the Green River basin of Wyoming, and three wells in the Cotton Valley formation in the East Texas basin.

According to Scott Reeves, Vice President of Advanced Resources International, a GTI contractor managing the analysis, “Seven of the nine restimulations were considered successful, and the program did determine that refracs could succeed if operators pay attention to procedures.”

GTI developed and tested three distinct processes for selecting restimulation candidates. Process I looked at production data and selected wells that were underperforming relative to their offsets. Process II selected wells where “less-than-best” practices were employed using detailed well data, pattern-recognition and neural networks. Process III employed type-curve matching to select wells with the greatest potential. Each method yielded different candidate wells and GTI did a benchmark study on set of synthetic wells to gain insight into each method's effectiveness for ranking candidates.

One conclusion was that none of these methods accurately estimated incremental reserves. GTI found that Process I identified underperforming wells relative to offsets but overlooked highly productive wells that could be superior restimulation candidates. Although GTI found a good correlation between the best 12 months of production and ultimate recovery, the method was ineffective for selecting restimulation candidates. Process II provided some insights but needs more development to be reliable. Process III identified highly productive wells with restimulation potential but was found to require a significant amount of engineering effort. GTI found this method to be the most reliable. GTI found this method to be the most reliable for identifying high-potential restimulation candidates in multilayered tight gas reservoirs.

GTI has begun a final test of this selection methodology with Patina Oil & Gas Corp. in the Codell gas-condensate reservoir in the Wattenburg field area in Colorado’s Denver-Julesburg basin. In 1999 Patina restimulated 110 wells and added 30,500 boe/well to gross reserves. Patina has had considerable success in selecting restimulation candidates using an algorithm that considers a weighted average of such factors as formation porosity-feet, gas-oil ratio, peak production, cumulative production, expected ultimate recovery, and differences in the ultimate recovery from offset wells. Patina estimates that while more than 1000 wells have been refraced in the area, they represent less than a fifth of potential restimulation candidates.

“The Patina wells represent an excellent opportunity to test the methodology developed by GTI,” said Reeves. “We will apply the methodology using only pre-restimulation data from Patina’s wells and develop a list of ranked restimulation  candidates. Then we will compare our selections with the actual results of Patina’s program and, based on post-restimulation performance, see if we were successful in picking the best candidates.”

The analysis and selection is being done by a team that includes ARI, Schlumberger-Holditch, West Virginia University, and GTI. A final report on the entire project, including the results of the Patina well comparison, should be completed within the next few months, according to Scott.

The Strategic Center for Natural Gas at DOE’s National Energy Technology Laboratory (NETL) is funding similar analyses that aim to provide methodologies for producers wishing to select stripper gas wells for remediation work. Three projects have been initiated in the Appalachian and Midcontinent basins based on decline curve analysis, offset well performance comparisons, and type curve analysis. According to Gary Covatch, Project Manager at NETL, “We hope to offer producers a portfolio of candidate selection methods so that they can pick an approach that works for their particular area. The focus is on developing a low-cost method for determining the wells with the highest potential for productivity improvement.”

James Engineering, Inc. will be looking at ways to identify remediation candidates from decline curve abnormalities. This will result in troubleshooting software and a handy, step-by-step “triage spreadsheet” that pumpers can use to identify, measure and solve well performance problems. Their data set for refining the methodology will consist of 376 Clinton wells in southeastern Ohio.

Holditch-Reservoir Technologies, part of Schlumberger, Inc., will be looking at new ways to use offset production histories to identify high-potential remediation candidates. They will be using a 410-well Median/Whirlpool formation data set from a producing field in northeastern Pennsylvania.

Advanced Resources International will be applying some of the lessons learned in the fracture restimulation study to develop type curves for assessing production enhancement potential. They will also be developing a “virtual intelligence” assessment tool for determining the applicability of specific production enhancement solutions (waterfracs of unfraced intervals, siphon strings for liquid removal, and high energy gas fracing to remove skin damage). Results of all of these investigations should be available later this year.

New Technologies Focused on Squeezing More from Existing Wells

Most of the improvements outlined above are focused on finding ways to reduce the cost of fracturing gas wells or improving the performance of wells previously fractured (i.e., getting more gas out of the wells we have). This path will have to be followed if independent producers are to meet the strong and growing demand for domestically-produced natural gas in an environment of dwindling exploration acreage opportunities.

References

Brister, Brian S., Lammons, Lance, 2000. “Waterfracs prove successful in some Texas basins,” Oil & Gas Journal, March 20, pp. 74-76.

Fletcher, Sam, 2000. “Mitchell succeeds with light sand fracing in Barnett shale,” Oil & Gas Journal Online, July 19.

Mayerhofer, M.J., et al., 1997. “Proppants? We Don’t Need No Proppants,” SPE paper 38611 presented at the SPE Annual Technical Conference and Exhibition held in San Antonio, Texas, October.

Mayerhofer, M.J., 1998. “Waterfracs – Results from 50 Cotton Valley Wells,” SPE 49104 presented at the SPE Annual Technical Conference and Exhibition held in New Orleans, Louisiana