Since hydraulic fracturing became a standard oilfield stimulation technique in the late 1940s, the industry has been working to improve the technique with new methods, technologies and materials. By the mid-1950s, “everyone” knew that a proppant pack’s flow capacity increased with fracture width, e.g., increasing numbers of proppant layers.
In a 1959 turnabout, however, Darin & Huitt (SPE 1291) demonstrated the theory of proppant partial monolayers. In a full proppant monolayer, grains are tightly packed; a partial monolayer has vacant areas around and between proppant particles. Darin & Huitt calculated that at some areal proppant concentrations below that of a full monolayer, conductivity far exceeds that of a monolayer. In fact, conductivity of partial monolayer remains superior to proppant layers up to about 10 to 12 layers, which corresponds to a sand concentration of about 3 lb/sq ft. Few oilfield applications reach such high proppant concentrations.
The theory suggested we could pump far less proppant and achieve similar — or even better — results than typical monolayer or multilayer proppant packs. For 40 years, experts searched fruitlessly for a means of achieving this Holy Grail of hydraulic fracturing, ultimately surmising that four factors prevented practical success:
· Inability to uniformly place proppant for complete coverage of the fracture with a partial monolayer;
· Insufficient proppant strength to support the load;
· Loss of fracture width due to proppant embedment; and
· Potentially deleterious non-Darcy flow effects in the relatively narrow propped fracture.
Forty years of frustration at trying to achieve a partial monolayer in the field led experts to declare them “virtually impossible to achieve.”
The advent and success of ultra-lightweight proppants (ULWPs) in the 2000s, however, suggests the experts were wrong. Careful examination of the highly conductive fractures associated with very low concentrations of these proppants provides ample evidence that partial monolayers are not only possible but — these days — frequent occurrences.
Theory to practice
Introduced in 2003, these proppants are ideally suited to slickwater fracturing treatments. One of these, BJ Services’ LiteProp 125, is composed of resin-coated, chemically modified walnut hulls. This successful proppant has a specific gravity of just 1.25, making it nearly neutrally buoyant in typical field brines. The material also has surprising strength, making it useful in reservoirs with closure pressures up to 6,000 psi and bottomhole temperature up to 225°F (107°C).
Initial success with these proppants led to refinements in placement technique and reductions in proppant concentrations, often achieving stimulated well productivity far beyond historical expectation. We surmised that the ULWP might be creating partial monolayers, but given the historical “impossibility,” we returned to the laboratory for some verification.
Darin & Huitt's equation for the permeability of propped fractures predicts that sands achieve an ideal partial monolayer concentration between 0.03 and 0.06 lb/sq ft and a full monolayer at 0.20 lb/sq ft, whereas a more typical field proppant concentration would comprise multiple layers at 1.0 lb/sq ft. Standard industry tests are designed to assess proppant concentrations between 1.0 and 2.0 lb/sq ft. Thus, to test ultra-lightweight proppants for partial monolayers, we developed new procedures and apparatuses based on API recommended practices.
The laboratory testing (Brannon et al, 2004, SPE 90698) found that partial monolayers of the ultra-lightweight proppants behaved similarly to equivalent volumes of sand at various temperature and pressure conditions, and the fracture conductivities observed were consistent with Darin & Huitt’s predictions.
Large-scale slot flow testing at the
The lab results restored the theoretical possibility that ULWPs could achieve partial monolayers in real-world fractures. Some 300 real-world fractures, meanwhile, were producing compelling anecdotal evidence that we had reduced theory to practice.
In
· In the mid- to late-1990s, the typical treatment was to pump 25,000 to 30,000 lb of 16/30-mesh
· In 2001, we tried to improve fracture length without increasing height growth by pumping 55,000 to 80,000 lb of 16/30-mesh
· In 2003, to achieve better proppant placement across the productive zone while containing fractures in zone, we tried a new tactic. We pumped about 1,800 bbl of slickened produced water with 7,000 lb of 20/40-mesh ULWP-1.25 proppant followed by a tail-in with 2,000 lb of 20/40-mesh Brady sand. For additional fractures in this area, we changed the treatment to use 2,200 bbl of slick 10-ppg brine with 12,000 lb of 14/30-mesh ULWP-1.25 followed by the Brady sand tail-in. (The customer has not reported a flowback issue, despite the lack of resin-coated sand.)
Normalized cumulative production per well (Figure 1) suggests that the considerably lower concentration of ULW proppant produced tremendous improvements in oil production compared with the much larger sand volumes used previously.
For another operator in
The Big Sandy field of eastern
To increase the effective fracture length, minimize the frac fluid impact on the formation and reduce the logistical burden of hauling so much sand, we turned to ultra-lightweight proppants (Kendrick et al, 2005, SPE 98006). We designed a treatment using 48,000 gal of nitrogen foam (6,850 gal of liquid fracturing fluid) with 6,000 lb of 14/30-mesh ULWP followed by 4,000 lb of 8/12-mesh per stage — just 20,000 lb ULWP in all.
After 16 successful fracturing treatments using this formula, the operator provided extended production data for five wells. Of the five, extended production volumes for four were equal to or higher than normalized production volumes from comparable offsets stimulated with conventional proppants. In the 6-month cumulative production comparisons for the various sets of wells, the ULWP-fractured wells outperformed the offsets by 30 to 70%.
As a bonus, the ULWP treatments cost about 8% less than typical conventional treatments because they eliminated considerable hauling fees, man hours, dozer expenses and pumping services. Furthermore, the operator reported significant reduction in wellbore cleanup time (possibly because fracturing fluid volumes dropped) and elimination of post-frac proppant migration that occurred on some conventional fracture treatments.
Validation in shale
The problem with absolutely confirming a partial monolayer is that it's difficult (impossible) to quantify particles in a fracture 2,500 ft (762 m) under the surface and 500 ft (152 m) away from the well bore of a producing (or injection) well. However, if we can obtain the fracture geometry and other treatment parameters, we can use basic math to develop generalizations about proppant behavior and fracture results.
For example, we were invited to participate in an engineering study of effective fracturing practices in the Diamond M field southwest of
The conventional fracture treatment on the lower interval used 88,500 lb of brown sand in 21,000 gal of borate gel, resulting in a propped half-length in Well 1 of 296 ft (90 m) and height of 354 ft (108 m). The comparable ULWP treatment used 15,500 lb of 14/30-mesh ULWP-1.25 in 71,000 gal of 10-lb brine, resulting in propped half length in Well 2 of 598 ft (182 m) and height of 324 ft (99 m) (Figure 2).
(We also fractured a second interval, which is not detailed in this article because of space limitations.)
Using the lengths and heights from the microseismic mapping results, we calculated rough total fracture face surface areas for the wells: 690,000 sq ft (64,103 sq m) for Well 1 and 1.39 million sq ft (129,135 sq m) for Well 2.
Using basic geometry, we then calculated surface areas for one face of the fractures and divided by pounds of proppant pumped to show proppant concentration of 0.047 lb/sq ft in the ULWP-treated well and 0.45 lb/sq ft in the conventionally treated well. (Recall that the ideal partial monolayer concentration for sand is between 0.03 and 0.06 lb/sq ft.) Using Darin & Huitt's equation with these concentrations and the estimated 2000-psi closure stress in these wells, we can calculate the ULWP-treated well conductivity as 10,000 mD-ft and the conventionally-treated well at just over 1,000 mD-ft.
Based on these results, we determined that conventional sand and fluid could probably achieve comparable proppant half-length and larger fracture surface area with four times the liquid and proppant — at a cost of 50% more than the ULWP-125 fracture.
Later, we evaluated results from 21 area wells fractured with brown sand against eight fractured with ULW proppants. For each type of well, we compared the oil production immediately following the treatment and the cumulative production a year later. The wells fractured with ULW proppants produced 22% more oil, cumulatively, because production declined more slowly over the course of a year.
Without microscopes that can travel into rock formations and provide absolute proof of proppant location, we must rely on mathematical assurances that these treatments achieve partial monolayers of proppant (Brannon et al, 2006, SPE 102758). Laboratory and field results from more than 1300 ULWP fracture treatments point to just one conclusion: Partial monolayers are not “virtually impossible,” but, rather, virtually ubiquitous with the right technology.