Fixing a Hole and Filling the Cracks (Part 1)

ear-wellbore screen-outs have gone from significant operational problems in some hydraulically fractured wells to an almost non-existent operational issue. Premature screen-outs have gone mostly into hiding downhole, where their partial impact on perforation cluster blockage can impact shale well production.

30 Years ago, around the time when I first learned about frac’ing, premature screen-outs could be prevalent, especially as the North American shale industry started to drill deviated and then horizontal wells.?It could take extra days on location for surging, flowing back or running coil.?Most likely, the frac crew would be released.?We would implement stepdown tests prior to pumping proppant to if a potential flow restriction was due to perforation or near-wellbore friction; and then assess screen-out risk and decide on pumping proppant slugs, increase gel loading, change perforation strategy for a next well, etc., to minimize premature screen-out potential on future wells.

The prevalence of screen-out problems triggered research at Delft University of Technology into the near-wellbore fracture geometry led by my thesis advisor Dr. Hans de Pater.?As drilling outside the (vertical) preferred fracture plane was done more frequently, the near-wellbore fracture was expected to become more complex as fractures turned from the local stress concentration near the well toward the far-field preferred fracture plane.?

A scaled experiment of this problem became my world for a few years, as I worked with a group of researchers and technical personnel at the Dietz Lab to evaluate what would happen with fractures growing from small holes in concrete cubes. Here are a few example photos of the near-wellbore geometry in deviated and horizontal wells.

First, going back to the photo at the top of this article, I show a cubic concrete block, about a foot on each side, with a deviated wellbore at a 45-degree azimuth with the preferred (vertical) frac plane and at a 45-degree inclination with the vertical.?This wellbore has (pre-cast) perforations on the top of the wellbore, at 0-degree phasing. As shown in the drawing below, when the plane of perforations is misaligned with the principal stresses, starter fractures may not link up along the perforations into an idealized single fracture.?Instead, misalignment can cause growth of multiple fractures.?

Many of these starter fractures grow out into multiple individual fractures, from which you see partial “peeled onion” portions on the top side of the test block above.?On the left side of the photo, the outermost of these fracture takes over and starts reorienting toward the preferred fracture plane.?

Multiple fracture growth behavior is not necessarily driven by the fact there are multiple perforations. The photo below is from an experiment in a horizontal well with a 45-degree azimuth with the preferred fracture plane, but without any perforations and with a small openhole section instead.?While the initial fracture along the wellbore is rather smooth, multiple fractures develop as the system propagates upward and downward beyond a few wellbore diameters away from the well. So, multiple fractures simply appear due to misalignment of the stress field and the well orientation.

The photo below shows that this behavior of multiple fracture growth is more abrupt, with a lesser impact of a fracture along the wellbore, when perforations are present.?The hoop stress around the perforations wins out over the hoop stress of the full wellbore and fractures start propagating in the preferred plane from the start.

The graph below, shows – in the big picture – maximum test pressure as a function of injection rate and viscosity.?In addition, it shows maps of the fracture geometry in relation to the well orientation.?Note that the minimum principal stress is left to right, so the preferred far-field frac orientation is from top to bottom.?The block diagram with fractures in black shows the horizontal well orientation in grey, with the openhole pressurized section of each well in white.?

Here are a few of the main conclusions from this complex summary graph.?First, if the well is perpendicular to the preferred plane, initiation of axial fracture (along the well) is preferred at high pressurization rates while initiation of transverse fractures (perpendicular to the well) is preferred at lower pressurization rates.?Second, the near-wellbore frac geometry is MOST complex with multiple fractures for oblique wellbore azimuths, as indicated by the area highlight in yellow.?Coincidentally, that is the most prevalent orientation between well and far-field frac in shale plays.?While this map of fracture geometries does not predict screen-outs, the more complex near-wellbore geometry in these oblique wells is expected to show higher near-wellbore friction, as well as a higher potential for obstruction once proppant particles are present in the flow stream.

For example, N-S oriented wells in North Dakota are at an oblique orientation with the N50°E preferred fracture plane. In the Liberty’s early Bakken days, our screen-out rate was a few % of all jobs.?Early fluid viscosity problems – we ran a bit too hot on biocides that broke FRs – exacerbated screen-outs. Also, we were still learning how much overflush was needed to eliminate all unwanted proppant between stages in less than perfectly drilled laterals.?We learned fast – that is why I am still near the top of the screen-out ranking for Liberty Field Engineers, as my 25% screen-out rate in the FE seat represented the first stages we pumped at Liberty.

Above is one of "my" Bakken screen-outs.?Notice how the full screen-out likely started with a partial screened-out of one of the perforation clusters at 1:20. While we lowered rate and started flush, it was not fast enough to avoid screening out. Sadly, Christmas break came early in 2011 and sent us back home and to the drawing board.?It was a tough day for a startup frac company.

Despite the oblique orientation challenge in many shale wells, premature screen-outs at Liberty happen about once every 150 stages (0.67%). According to Liberty Ninja Tyler Starn, that number has been roughly constant for Liberty history beyond our first two months of frac’ing.

Many aspects of today’s routine shale frac designs are geared toward screen-out minimization:

• Smaller sand grains don’t get stuck so easily - 100 mesh sand is now the dominant proppant in shale wells
• Lower proppant concentrations.?Average proppant concentrations in shale wells are about just over 1 PPG
• Higher rates, albeit with thinner fluids.?Velocity and viscosity are your friends in near-wellbore proppant transport.?Despite their low viscosity, FR fluids have higher elasticity, which can help with proppant transport at very low shear rates
• Shooting shorter perf intervals and shooting perfs in a vertical plane

In addition to the screen-out rate reduction, what has changed is that screen-outs have very little impact on frac operations today.?With multiple well ops now common in most shale basins, if a well needs to be flowed back or cleaned out with coil following a screen-out, the problem well is just temporarily taken out of the lineup until it is ready while frac and wireline continue their musical chairs around remaining wells.

Most screen-outs are cleaned out when flowing back two wellbore volumes. For hard screen-outs, like that one in 2011, a costly coil run may be necessary. But it will not shut down frac operations for 2 weeks.

This change was also at display at a near-wellbore workshop organized by SPE earlier this month.?There were no talks about jobs screening out. That is a testament to our industry’s efforts to clean up the near-wellbore neighborhood, at least from common view.

So, what was discussed at a conference dedicated to the near-wellbore geometry? Screen-outs hidden from view downhole.?Partial screen-outs of perforation clusters that prevent a more equal distribution of proppant and fluid throughout the reservoir.?While screen-outs have gone into hiding at the surface where we work, new diagnostic tools allow their observation in their new partial form.?And this awareness can then steer us toward design choices that help achieve a more equal distribution of fluid and proppant – to obtain production equally from everywhere in the reservoir.

As advised by Paul McCartney, we are fixing a hole and filling the cracks – but we have work to do to fill them properly.?There were lots of great presentations that our live panel summarized at the end of the conference that show some progress toward filling cracks.?A brief summary will follow in part 2.??