- PDF LinkChapter PDF
Palmer Ian, 2009. "Getting Natural Gas Out of Shales and Coals", Unconventional Energy Resources: Making the Unconventional Conventional, Tim Carr, Tony D’Agostino, William Ambrose, Jack Pashin, Norman C. Rosen
Download citation file:
This paper will discuss some established procedures and recent learnings in regard to well completions and production in both shale gas and coalbed methane reservoirs. The talk will address certain commonalities, peculiarities, and challenges of both. Some of the technical aspects will include the importance of natural fractures and permeability, examples of commercial production, and optimizing well stimulation. The approaches and learnings from coals and shales may be transferable to newer unconventional resources.
Methane gas from coalbeds (CBM) started a little before 1980, so it? been going on for about 30 years in the USA. Growth has been steady and robust (Fig. 1). The early years were motivated by government R&D funding via the Department of Energy, and by tax credits for gas production. On the other hand, shale gas really only took off in the late 1990s (Figs. 2 and 3), when firstly new technology and secondly gas prices were the motivators. Both of these unconventional resources were difficult to understand initially. For example, in CBM we had to learn that the permeability of the coal was the key. In shale gas, we discovered that we had to create our own reservoir. Other intriguing aspects of natural gas from coals and shales reveal a creative history of production from these resources, and here we outline just a few of these.
Coalbed methane (CBM) has become a resource of global significance, with the emergence of active CBM plays in Canada, Australia, India, and China, for example (Palmer, 2008). Along with Russia, these countries each have more than 100 Tcf of CBM resources (and in Russia, China, and Canada possibly over 1,000 TCF). Total world-wide CBM resources have been estimated at 9,000 Tcf (Kawata and Fujita, 2001). Over the past decade solid CBM industries have been established in Australia, Canada, and China. In terms of industry maturity after the USA, Australia and Canada rank next, followed by China and India. At the other end of the spectrum, countries like England, France, Turkey, and Colombia have much smaller resources, but are actively trying to get a CBM industry started.
In the USA, the San Juan basin alone has CBM resources of ~50 Tcf (excluding Menefee coals), and this basin has supplied two-thirds of the CBM in the USA. Proven reserves and production numbers are given in Figure 1. By 2005, both reserves and production had reached almost 10 percent of total USA values, which is a rather significant threshold. One recent estimate is that there are now about 90,000 CBM wells in the USA.
Total world-wide shale gas resources have been estimated at 16,000 Tcf, about 130 Tcf, of which is in the USA (Kawata and Fujita, 2001). The USA and Canada are the dominant regions for shale gas activity (Fig. 2), which shot ahead in the Barnett shale after the late 1990s (Fig. 3). The Barnett technology has been transferred to other shale plays, such as the Fayettville, Marcellus, and Haynesville plays in the USA, and the Montney, and Horn River plays in Canada.
One characteristic of these unconventional resources is that they are hard to understand, which is why their exploitation was deferred until only fairly recently. As an engineer working CBM wells once said, “Everything about CBM is squirrelly,” which is slang for “perplexing.” First, most of the methane gas is adsorbed in the coal, not just freely sitting there in the rock? pore space, as it is in sandstone (if you lower the reservoir pressure though, it will come out of the coal?ike opening a bottle of soda). Second, stimulating CBM wells, which generally needs to be done to get the gas out fast enough, has been a mind-twisting experience. For example, one dramatic approach is to do a controlled well blowout, where a cavern up to 10 feet across is deliberately created in a coal seam at the bottom of a well (called a cavity completion). The more common type of completion is hydraulic fracturing, and even here we have discovered that fluids and chemicals normally injected into a reservoir (to create long thin fractures like those in the sidewalk) can be quite damaging to the coal permeability. Last, it has been found in the San Juan basin that the absolute coal permeability increases by 10 to 100 times as the reservoir depletes. The San Juan basin is a juggernaut for CBM production anyway, by virtue of the outstanding thickness, pressure, and permeability of the coals, but on top of that, the absolute permeability increases quite dramatically with time. How lucky can you be!
Shale gas too is hard to understand. Some of the gas is also adsorbed on the shale. But a large part is also free gas, which should readily flow out, except for two be found in Jenkins and Boyer, 2008). Even after the things: (1) the pore throats are about the size of a molecule of methane gas, and (2) the permeability (flowability) is often impossibly low. In the Barnett shale, the matrix permeability is ~0.1 ?D (0.0001 mD), which is 10,000 times lower than a typical 1 mD for standard conventional reservoirs that produce oil or gas. Although natural fractures exist in the Barnett shale, which should boost the reservoir permeability by a lot (e.g., ~100 times), they are all closed: sealed up by cement deposits that accrued over millions of years! How unlucky can you be!
To make matters worse, both CBM and shale gas are statistical plays, as Figure 4 attests (CBM proof can be found in Jenkins and Boyer, 2008). Even after the new technology has been brought in, the spread of gas rates is huge: around the year 2005 in the figure well rates vary from 0 to 4.5 Mcfd, and there are many wells that produce very little gas. This cannot be due to well drilling and completion techniques, because these have been more-or-less standardized (although they can still be tweaked). It must instead be due to (1) the presence or absence of the sealed natural fractures, which are the conduits for the rock-cracking process, or (2) variations in crackability of the rock itself, or (3) differing access by the gas into the cracks, along which the gas flows to the well.
Keys to Understanding CBM and Shales
The path to understanding is tortuous in unconventional plays. Here we give a few examples, which will lead us to some key principles.
One large company pursuing CBM had the vision from the earliest days. They set up a group of staff for R&D. They stuck with a well that performed poorly in the San Juan basin, trying several different types of well completions, until they made it work (they later became one of the biggest players in the San Juan). The company also had the vision for CBM in other basins. The vision apparently came through one particular VIP at the company.
Although the learning was obvious, we uncovered a delightful expression for it: “As a way to arrive at the truth, exactitude and methodology are, in the end, far inferior to vision and apotheosis” (Helprin, 1991).
We discovered cavity completions quite by accident. As we heard the story, one company in the San Juan basin sweet-spot (called the fairway) had to deal with a well blowout, which is a dangerous thing. After they wrestled for control of the well, they discovered it made gas at a faster rate than offset wells which had been hydraulically fractured (the normal well completion technique). So they developed a controlled well blowout technique. By quickly turning a valve at the surface, they allowed a well to blow down, and to belch up gas and fragmented coal particles for a short time.
Then they let the reservoir pressure build up again and repeated the process. Eventually they found a cavern had been created up to 10 ft across in the coal seams (Fig. 5). Cavity completions work in only a few CBM plays (Palmer and Cameron, 2003), but where they do work, they work well (other successes include Fairview and Spring Gully in the Bowen basin of Australia).
The learning was to recognize a serendipitous event, and to adapt it to a useful purpose. Serendipity can only occur if we are moving forward, trying to make things happen.
Permeability-based completion bands
This finding came about after many years of CBM production in many basins. CBM well data was available after almost 30 years of production in the USA, and it could be integrated with new well data from Australia and Canada. Out of this analysis we found that most CBM well completions could be assigned (ahead of development) if we knew the approximate permeability of each part of a CBM play (Fig. 6). This was an important synthesis. It immediately told us that if the coal permeability were 1 md or less, vertical wells that were fracture stimulated (the standard CBM completion) would not be successful. Despite this knowledge, we still run across operators who are having difficulty getting wells to produce. When asked what the permeability is, the answer comes back “We don’t know, or it’s around 1 md.” If the permeability is around 1 md or lower, the only proven completion is multi-lateral wells, such as a quadrilateral or a fishbone (pinnate) well pattern (Fig. 6).
The learning was to find someone to analyze the well data, to integrate it with other available data (e.g., from the literature), and to synthesize the results into useful field applications. Analysis is a left-brain process (linear, logical), while synthesis is a right-brain process (holistic, lateral-thinking), so someone (or a group) with both mental capacities is needed.
Permeability increase with depletion
The final CBM example is the prospect of a permeability increase with depletion in the San Juan basin, as a coal reservoir is drawn down with continuing production. The source of the permeability increase would be shrinkage of the coal matrix as gas desorbs from the coal. Any matrix shrinkage would imply that the coal cleats (fractures) would open wider, thus increasing the permeability. The concept is very unconventional. The effect had been measured on cores in the lab, so we knew it is real. But it remained hidden in the field, partly because we felt that compression of the cleats by increasing effective stress would cancel it out. Nevertheless, we outlined a theory which included both effects, and showed that under certain conditions matrix shrinkage could be the dominant effect. One manager in the San Juan basin, after hearing the theory, said “I hope you are correct, because we have a layoff in 3 months time, and I want to save all my CBM staff jobs.” Everyone laughed.
Later, the paper was presented at a national meeting, and a colleague who had heard the paper called me excitedly. He said he had read the paper on a plane right after the meeting, and applied it the next few days to successfully match gas rate histories of CBM wells he had never been able to match before. The permeability increase due to matrix shrinkage was confirmed to be a strong effect. The same colleague wrote a paper that was published soon afterwards. The original company changed their history-matching approach, and also began to measure permeabilities in the field. Some results are shown in Figure 7, where very strong increases of permeability associated with depletion are seen. This work changed forever our thinking about permeability in coal seams, and matrix shrinkage has boosted the CBM economics there. Recent theoretical work has made further progress in this area (Palmer, 2009b and c; Shi and Durucan, 2009; Clarkson et al., 2009). The final answer is not yet in, because such permeability increases have not been reported in other CBM plays, and that is a puzzle. If we could confirm the permeability increase in other plays, this would reinforce the predictability of the theory for a new play, which could turn marginal plays into economic successes. The theory (in reverse) has also been adapted to predict the swelling and loss of permeability during sequestration in coals of greenhouse gases like CO2, so it has had quite a far-ranging impact.
The learning was to encourage theoretical development, because that can enlighten previously unexplained field data, and can be used to make predictions that once could only be guessed at. Theory and experimental (field) development is strongest when it moves forward hand in hand.
Persistent technological innovation
Shifting now to shale gas, we were unsure the gas molecules could even break out of the tiny pores of the rock, and we also knew that in many cases the flow channels to the well were impossibly tight (i.e., very low permeability). Apart from a couple of low-level shale gas plays that seemed like a non-starter, and it was, except for one company (Mitchell) in the Barnett shale who since 1981 had continued to believe (Fig. 8). Frac treatments were needed, of course, to let the gas flow to the well faster, and the original fracs used gelled fluids to carry the proppant (e.g., “first MHF” in Fig. 8). Breakthrough #1 came in 1997, when the first slick-water frac was applied, which had two advantages: (1) not damaging the sensitive fractures as much, and (2) creating more fracture branches: i.e., cracking open more of the tight shale rock (Fig. 9).
Breakthrough #2 came in 2003 with the first commercial horizontal well. Breakthrough #3 occurred simultaneously: fracturing separate segments along the length of a horizontal well (typically 6 segments are now fractured in a 3000 ft well in the Barnett, but some wells had up to 20 segments). If a well is oriented correctly, the segment fractures will be transverse to the well, and access the largest volume of reservoir gas. Yet another innovation occurred in 2004 when two or more horizontal wells were fractured simultaneously (Fig. 10). This led to fracture branches and cracks in the rock that are denser (as evidenced by micro-seismic bursts), and therefore, allows the gas to flow even more readily.
The learning was that persistent technological innovation was key to success in the Barnett shale. Innovation will be accompanied by failure, and a company needs to embrace a culture of “failure-is-okay” for this to work. Although rare, a persistent innovative culture has been exemplified by one company (Moorman, 2008).
Creating your own reservoir
We knew long ago that the tightness of the shale reservoirs meant we had to fracture hydraulically the wells to get the gas to flow. What we discovered was that fracturing by segments (up to 20) with slick-water fluid in long horizontal wells led to a broad expanse of cracked rock (fracture network) as illustrated by Figures 9 and 10. The flow permeability in the cracked region was much greater than in the virgin shale. In a real sense, we were creating our own reservoir, because the flow contribution from the virgin shale was very small. Therefore, the larger the fracture network, the larger was the gas recovery (see Fig. 11). And the denser the rock cracking in the fracture network, the higher was the gas rate (Mayerhofer et al., 2006). In consequence, the onus was on the well stimulation, and the goal became to create a fracture network as large as economically possible.
The learning was a paradigm shift: to realize we could create our own reservoir, rather than remain handcuffed by the limited permeability and flow of the virgin shale. As far as we know, this has no precedent in the long history of the oil and gas industry.
When a new idea is offered in one area and a player adapts it to another area, this is what we term a crossover. The following example starts in CBM and ends in shale gas. In CBM, we learned from lab tests around 1990 that almost any chemical can be adsorbed by coal, damage the sensitive cleats (fractures), and hurt the permeability. When we published the results, it created a bit of an uproar, because the status quo for CBM well stimulation was a “Cadillac frac,” a frac treatment containing expensive gels and other chemicals deemed important to carry the proppant and protect the well from damage. Instead, we recommended a simple water frac, and verified that it gave better well performance. We later announced this finding to a group of engineers in our company who were engaged in drilling and completing wells in tight sands (similar to tight shales). We pointed out that expensive gelled frac fluids could also damage the natural fractures in tight sands, and that maybe slick-water fracs might do better. The room fell silent and not one person responded, nor seemed to appreciate the new knowledge that was placed before them. In part, it appeared this was due to the tunnel-vision attitude that ?BM is different.?In years to come, slickwater fracs would have good success in some tight sands. More to the point, they were first tried in 1997 in the Barnett shale (Fig. 8), and now almost all frac treatments are done in the Barnett using slickwater.
The learning was to be aware of new ideas beyond your immediate job focus, and to be proactive in pulling them in, exploring them, and testing them in your own field. Fresh eyes can come in and boost this process, because they are innocent enough to ask “Hey, why don’t we try this?”
As a post-script, we illustrate a technology transfer from shale gas to CBM. In Figure 10 each small yellow triangle reveals where a segment of the well was fractured. So we have three wells, each of which was fractured about 6 times, for a total of 18 frac jobs. The total volume of cracked rock is huge, and the total gas rate lay between 7–9 MMcfd for the first months (worth a lot of money when the gas price was high in 2006). Recently, this technology has been carried over to the dry Mannville coals in Canada, using a multi-lateral well design that looks something like Figure 12. Each of the open-hole laterals (all in one seam) was fractured between packers: quite a complicated well completion, but apparently commercially viable, with a reported gas flow rate of ~5 MMcfd (anecdotal).
First, about learning curves in unconventional gas plays. The learning curve for CBM, starting from around 1980, has been ~15 years. As Figure 13 shows, the learning curve in the Barnett shale was ~20 years. For armchair analysts and government planners, this says to expect 15-20 years to acquire the knowledge to exploit a new unconventional resource. For spinoffs however, the time is much less. Figure 13 reveals that exploitation in the Fayetteville shale, which followed in the footsteps of the Barnett shale, has been less than 4 years. This is good news: after the initial development, sister exploitations can occur much more rapidly.
Next we summarize the technology lessons we have discussed from the CBM and gas shale industries. The ?earnings?have been taken from above and rearranged. The hope is that these lessons might be transferable, and benefit the development of other unconventional resources:
A clear and robust vision is critical.
Persistent technological innovation is a key to success. Innovation will be accompanied by failure, and a company needs to embrace a culture of failure-is-okay for this to work.
Realize that via technology one can create their own “reservoir,” rather than remain handcuffed by the limited “reservoir” available. This requires creative thinking.
Find someone to analyze the data, to integrate it with other available data (e.g., from the literature or from other fields), and to synthesize the results into useful applications. Someone, or a group, with skills for both analysis and synthesis is needed. This also is creative thinking.
Encourage theoretical development, because that can enlighten previously unexplained field data, and can be used to make predictions that once could only be guessed at. Development is strongest when theory and experimental (field) work move forward hand in hand.
Be aware of new ideas beyond the immediate job focus, and be proactive in pulling them in, exploring them, and testing them in one? own field. Fresh eyes can boost this process, because they are innocent enough to ask “Hey, why don’t we try this?”
Recognize a serendipitous event, and adapt it to a useful purpose. Serendipity can only occur if one is moving forward, trying to make things happen.
These attributes are not meant to be comprehensive. Rather, they are just a sample that arose out of real-life situations which occurred during the development of the CBM and shale gas industries.
Finally, a fascinating post-script on the CBM industry. On the east-coast of Australia several companies, including multi-nationals, are planning five new LNG trains for CBM gas (an LNG train can cost several billion dollars). The driver is the price of LNG, for shipment to South-East Asia and Japan, which is projected to be around $10/Mcf. It is estimated that 10,000 to 30,000 new CBM wells in eastern Australia will be needed by 2014. Can you imagine? Although there is likely to be some amalgamation of LNG trains, the concept is both daring and daunting. Application of the above learnings may help to accelerate the achievement of these goals.