Once considered marginally feasible, coalbed natural gas projects are benefiting from game-changing technology applications.
Not too many years ago, weak prices were driving most everyone out of the coalbed natural gas (CBNG) business except for a very few "plain vanilla" fields. The economics and risk simply didn't justify investing in technology to improve production or dewatering efficiency. However, over the past decade there has been a dramatic increase in CBNG drilling, and supported by stable prices in the $4-$5/Mcf range, technology is helping operators improve producibility and cut costs.
With improving communication to the coal seam as a goal, new stimulation techniques are paying off in the US Rocky Mountain region. Propped hydraulic fracturing has been the traditional method of accessing the coal seam, both to provide channels for efficient dewatering and for subsequent gas production. But often the techniques that work in traditional oil and gas reservoirs fail to deliver in the coal. Coalbed fractures have been found to under perform those in clastic reservoirs in spite of high stimulation volumes and relatively shallow depths. Effective fracture lengths are rarely documented beyond 200 ft (61 m). Some have estimated that recovery could be tripled if stimulation results could be brought on par with those experienced in typical sandstone reservoirs.
Three factors dominate in producing effective propped fracture systems: azimuthal orientation of the fracture relative to the prevailing horizontal stressfield, optimal placement of the fracture in the stratigraphic column and the use of compatible fracturing fluids.
In coalbeds, cleats or natural fractures form over time during the coalification process. The first to form are the face cleats, which are generally contiguous and, since cleating normally occurs during tectonically benign periods, are unlikely to be stress-oriented. Later, butt cleats form that are perpendicular to the face cleats and usually only propagate between two adjacent face cleats in relatively random patterns. This causes a natural fracture network that is tortuous, and that can impair the efficient propagation of subsequent hydraulic fractures. When considering hydraulic fracturing, it is useful to know the relative orientation of regional horizontal stressfield with the cleating system since fracture efficiency can be greatly affected. The smaller the angle between the cleats and the horizontal stressfield, the better the frac will perform. Permeability and permeability anisotropy are strongly dependent on horizontal stress orientation and magnitude. Unfortunately, although the prevailing tectonic stress may be well known for a particular basin, regional stresses can occur due to post depositional faulting that create an overriding local stressfield that can affect the direction of fracture propagation in the zone. The dipole shear sonic log can identify the presence and azimuth of these local stressfields and their relationship to the cleating system. Armed with this information, an operator can make decisions on completion techniques with higher likelihood of success.
Depending on the operator's understanding of the various stresses and the orientation of cleats and natural fractures, completion alternatives can be evaluated. There are three principle types of completion in common use today: natural completions in a vertical borehole - either perforated or barefoot; horizontal and/or multilateral completions in the coal seams; and propped hydraulic fractured completions. For the latter, elastic tensile fracturing is the most efficient way to create long fracture lengths. Unfortunately, coal is generally softer than the surrounding rock, and has a much higher fracture gradient. As a result, it becomes unstable when subjected to hydraulic pressure causing highly complex fracturing, either in a zig-zag pattern following an alternating path from face cleat to butt cleat, or a dispersed path among several adjacent cleats. Either way, fracture half-length is dramatically suppressed and proppant placement is inefficient. The difference is significant. For a given volume of frac fluid, a 2.6:1 ratio of fracture lengths was observed depending upon whether the coal failed in tension along a plane, or degraded into a complex fracture pattern among the cleats.
Alternatively, when high fracturing pressures are employed, shear slippage can occur along cleat planes that accomplishes nothing. In addition, when coal fails in shear, closure stress is lost, so there is nothing to hold proppant in place after pressure is released. This can lead to unacceptable proppant flowback during production. The occurrence of shear failure in a particular area has been an argument for performing slickwater fracs, and these have had moderate success.
Gelled fracture fluid can cause significant damage when it gets out of the pressure-induced fracture. Leakoff into the cleating system is difficult to control and almost impossible to remediate. The presence of unflushed frac fluid in the cleats can cause microscopic swelling of the coal that kills porosity and permeability. Even when aggressive breakers are used, wells have experienced a 5- to 10-fold reduction in permeability. Swelling damage is irreversible. A production simulation showed that damage from polymer based frac fluid will cut gas production more than 50% in the first two years after treatment.
Experimental apparatus was designed and used to determine post-treatment permeability of a crushed coal sample after a borate crosslinked fluid and a linear gel fluid were injected. The technique was sufficient to yield qualitative comparisons. The sample's baseline permeability was determined by injecting 2% KCl water at a predetermined volumetric rate. Then the tested fluids were pumped back through the sample at the same rate until one pore volume was injected through the coal. Regained permeability was determined by once again pumping 2% KCl back through the sample. Pressure and time required to achieve good cleanup were recorded. In the case of the borate crosslinked fluid cleanup permeability was only 33% of original permeability. In addition, by analyzing the differential pressures required to initiate cleanup it was estimated that pre-cleanup permeability was far less, on the order of 2.2% for the borate fluid and 4.4% for the linear gel.
A coal-compatible solids-free frac fluid was designed using viscoelastic surfactant technology (VES). This fluid produced regained permeability of 76% and required only one-eighth of the time/pressure combination to achieve full cleanup. Final post-treatment permeability was 230% of that produced by the borate fluid and the linear gel (Fig. 1).
Developing Fracture Models
Encouraged by their results, the investigators set out to model coal fracture propagation so predictive algorithms could be derived. Unfortunately, the technology to model complex fracture development in coalbeds does not produce quantitative results. Nevertheless fracture simulators can be useful in a qualitative sense to predict fracture height growth and coverage as well as for pressure response interpretation. An extensive study was undertaken using two commercial simulators in a controlled environment provided by reservoirs with rich, comprehensive datasets. Details of this study can be found in SPE 84122. The ultimate results of the tests were somewhat inconclusive, meaning that neither simulator was adequate to the task. However, the tests established a clear need for further study.
Having determined what doesn't work, the investigators started thinking outside the box - literally looking for solutions outside the coal seams - and were rewarded. Using a technique that was developed and perfected in the North Sea called Indirect Vertical Fracture Connectivity (IVFC), fractures were initiated in adjacent clastic strata and allowed to propagate into the coal beds. There are numerous advantages. Firstly, it is possible to treat two adjacent coals simultaneously by initiating a fracture in the intervening sandstone. Not only does this produce much greater fracture half-lengths, but when the fracture reaches the sand/coal interface it maintains its geometry, propagating a clean planar fracture into the coal seam. Because vertical permeability is higher than horizontal permeability in the coal, the fracture does not need to penetrate the coal seam deeply to be effective - length is more important than depth in this case. In addition, proppant placement in the coal is very efficient and clean-up back into the porous, permeable sand is more efficient and thorough. The reason this technique works is that the relatively lower fracture gradient in the sandstone ensures that as the fracture propagates into the coal seam it maintains its geometry and connectivity.
IVFC Case Histories
Two outstanding examples of the application of IVFC in the Rocky Mountain region provide evidence of its ability to deliver positive results. In Utah, a comprehensive study was made of 31 wells with fracturing performed in the coal only and in 42 wells in which fracturing was initiated in adjacent sandstones. Production rates across the frequency spectrum of the 42 wells were double those of the wells fractured in the coal only. In addition average treating pressures were reduced by 1,000 psi and screenouts were significantly reduced.
In the Central Rockies, 11 wells were perforated and treated in the coal only, followed by 4 wells that were perforated in both the coals and in the adjacent sandstones. In the former group of wells screenout rate was 65%, but with the latter group screenout rate was zero. As with the Utah example, average production rates in the wells perforated and treated in both coal and sand beds were double those of the 11 wells treated in the coalbed only (Fig. 2).
When low-damage VES fracturing fluid technology is combined with IVFC stimulation techniques effective fracture penetration is greatly enhanced, and a step-change in low-permeability coalbed production economics can be realized. Implications are high. For example, in the prolific Powder River basin, the southern half is characterized by two very thick coal deposits; however, in the northern half, at least eleven narrow coal beds have been identified. This new technology has the potential to affect development economics significantly.