Waterfracs: A New Perspective Based on Field Experience
John Ely: Ely and Associates
Almost since the inception of hydraulic fracturing low viscosity treatments have been conducted. We would define low viscosity treatments as treatments using water and no significant amount of thickening agent where surface viscosities at ambient temperatures are less than 10 centipoise measured at 511 reciprocal seconds shear. In fact in the San Juan basin slick water treatments were the prevailing type of treatment to be utilized, pre 1968, and have continued to be so today. Prior to the introduction of crosslinked gels in 1968 low viscosity treatments were a very large segment of fracture treatments. With the development of crosslinked fracturing fluids and all their attributes, low viscosity fracs were considered low technology and became a small segment of treatments pumped. Over the past 9 years plus there has been a huge movement toward non-viscosified fluids containing small proppants. This was led initially by the early efforts of UPRC with their paper on the lack of need for proppants and early treatments in tight sandstone reservoirs such as the Cotton Valley, Travis Peak, Bossier, and many tight sands in the Rocky mountain areas. With the advent of Source Rock stimulation in the Barnett and other shale plays the use of water fracs has virtually exploded. The success of these types of treatments has shaken fracturing theory to the core. Tremendously successful stimulation has been achieved using propping agents which based upon conventional thinking have virtually no conductivity at down hole conditions. Remarkably, the use of larger, more conductive, proppants in this 1 centipoise fluid has not only not improved stimulation results but has been shown to be detrimental far beyond the problem of potential screenouts due to poor transport.
Over flushing has become the norm and pumping alternating sand laden and neat fluid has improved results rather than given the expected decrease due to lack of effective conductivity. Conventional requirements for net pressure gain to achieve good conductivity has been replaced with the need to show no net pressure gain on properly designed treatments. Net pressure gain is indicative of packing proppant and it is our belief that the small proppants used in water fracs do not function in a pack but rather act as bridging and diverting agents and or function as a mechanism to hold the fracture open akin to partial monolayer theory. This translates into using low concentrations of proppant and designing in sweeps to be sure no packing occurs.
The utilization of viscous fluids in the vast majority of shale’s has been counterproductive. It is believed that the high viscosity fluids tend to create a dominate hydraulic fracture which is counterproductive in naturally fractured reservoirs. Where production comes mainly from natural fractures one does not want to parallel the fracture systems which is the natural course of events when using viscous fluids. We have watched this phenomenon during microseismic work where we have a multitude of seismic events occurring at great distances from the wellbore while pumping thin fluid. Upon commencing to pump viscous fluid, such as a hybrid treatment, all seismic events cease away from the wellbore and a dominate narrow fracture pattern is generated near wellbore. It is felt that the low viscosity fluid tends to follow the natural fracture plane allowing for much improved stimulation compared to a dominate fracture paralleling these same fractures. Some authors have hypothesized that the success of water fracs in stacked sand shale sequences such as the granite wash, cotton valley, Olmos etc. are due to differential width between the sands and shale and the small proppant bridges holding open infinitely conductive fractures. Another mechanism, perhaps more acceptable, is that many of the so called microdarcy formations are dominated by natural fractures in the reservoir and in fact the matrix permeability is too low to produce hydrocarbons in geologic time. This in fact is borne out by the huge success of waterfracs in many reservoirs where crosslink gels have been mostly unsuccessful.
What has been extremely interesting has been the fact that more and more “conventional reservoirs” have been found to achieve better results with waterfracs compared to the conventional packed proppant pack treatments utilizing crosslinked gels. Examples of these reservoirs are the Cleveland formation in the Texas Panhandle, the deep Morrow formation in the Texas Panhandle, the Olmos formation in South Texas, The cotton Valley and Travis Peak formation in North East Texas, The Mesaverde formation in Colorado and many others. Obviously we have had great success with crosslinked gel systems in many reservoirs but the question comes to mind what a properly designed waterfrac would do in these same reservoirs. Almost all of the deep high temperature reservoirs were being treated with slick water, or certainly systems with no stable viscosity, prior to 1968 but the treatments were typically small and utilized large proppants.
Rationale for small and in some cases substandard propping agents
In the early stages of the “Waterfrac Boom” operators were having screenout problems while using 20/40 or larger proppants. To alleviate the screenouts operators switched to smaller 40/70 proppant and were able to place the proppants at concentrations exceeding 2 pounds per gallon. The surprising thing was that not only were they able to place the proppant, they achieved better stimulation results compared to wells where larger proppant had been placed. Following the same thinking operators in the Barnett shale started using 100 mesh sand as the primary proppant. With the huge number of treatments and the extraordinary volumes of proppant pumped there was simply not enough Ottawa quality sand available and operators in the Barnett started pumping much lower quality 100 mesh sand. This proppant was available very close to the Barnett shale play and remarkably the wells responded equally well as they had with the Ottawa 100 mesh sand and also Ottawa 40/70 sand. We have noted success in utilizing sand as proppant in deeper shale such as the Haynesville, Woodford, and Eagleford where based on the frac gradient we should be using ceramic proppants. The initial thought to explain the phenomena was speculating that since the shale in many cases does not have a dominate stress that the closure would be less. Perhaps a more believable explanation for the shale is that since there is virtually no permeability in the matrix of the rock there is no mechanism for the pressures to be drawn down to crush the proppant. Needless to say this use of weak proppant should not necessarily be transferred to conventional reservoirs where there is measurable permeability in the matrix. I would note that one operator has reported success in water fracs in a conventional reservoir using 100 mesh Ottawa sand. The depths are within the range of Ottawa but belie the need for a conductive proppant pack in the tight reservoir.
Using or not using surfactants in waterfracs
More than 9 years ago a very frugal customer took it upon himself to remove surfactants from his fracture treatment in the Cotton Valley Sand. Those of us who have both learned at the feet of countless frac Gurus and also have taught Hydraulic Fracturing were aghast at this omission. We were sure that the well having high surface tension fluids would not clean up and perhaps we would be blamed for the potential problem. We closely monitored the well and were surprised to find that the well cleaned up very well and appeared to be better than offsets. After pondering this apparent anomaly for some time it was hypothesized that perhaps it is not a good idea to pump low surface tension fluids into very low permeability reservoirs allowing water to penetrate into tiny pore throats and hairline fractures where the surfactant would inevitably plate out leaving high surface tension fluid in the pore spaces creating damage. For over 6 years we have not recommended the use of surfactants in water fracs and have seen no ill effects due to the omission.
There have been several water recovery products introduced to the industry over the history of fracturing and intuitively we felt that with better water recovery that we would achieve better production. In reality what we have found is that enhanced water recovery has little or no effect on production. Many extremely complex “scotch guarding” type materials and some relatively expensive fluorocarbon surfactants have been utilized in fracturing and have indeed yield enhanced flow back but with no real effect on productivity. We have noted some cases in extremely tight rock that particular surfactant formulations have in fact damaged the rock to the extent that an interval would not build up pressure after previous production. There is indeed a correlation between load recovery and productivity but the correlation is inverse i.e. production and the quality of the well is better with minimal load recovery and very poor wells recover very large percentages of the load.
The tremendous success of Waterfracs has indeed shaken the foundations of conventional hydraulic fracturing theory. It is extremely interesting to observe those that continue to utilize both conventional high viscosity gels and high strength proppant in unconventional rock and, even with obvious failure to achieve success compared to waterfracs, persist in their thinking and procedures. Although it is obvious that we do not totally understand the mechanisms of why waterfracs work so well, at the same time we cannot allow ourselves to overlook the obvious results. What is most encouraging for the future is that we have identified a process which appears to be unlocking heretofore non accessible hydrocarbon reserves and in fact have within our hands the ability to make our nation, for the first time in more than 60 years, self sufficient in hydrocarbons by utilizing natural gas as our primary energy medium.