Plugging-induced screen erosion difficult to prevent

Alex Procyk
Pall Corp.
Houston
Michael Whitlock
Pall Corp.
Cortland, N.Y.
Syed Ali
Chevron USA Production Co.
New Orleans

• The screen shows plugging-induced erosion. From the left are the outer cage, outer drainage mesh, first layer PMM, second and third layer of PMM, inner drainage mesh, base pipe, and base pipe opposite erosion site (Fig. 2) [10,998 bytes]
• After erosion tests, the high-void (60X magnification) PMF-II media experienced minimal erosion (Fig. 6a) [17,275 bytes]
• After erosion tests, low-void PMF-II media (60X magnification) shows erosion (Fig. 6b) [16,573 bytes]
• The tests produced various erosion patterns of the media. The media are as follows: (A) New media, (B) Media without shroud, (C) Media under perforated shroud, (D) Media under louvered shroud, and (E) Media under experimental shroud (shroud represented by blank disk) (Fig. 7) [11,796 bytes]
• For all samples, the bottom layer was more eroded than the top. The top photo shows the top media layer. The bottom photo is the underside of bottom media layer with the top media layer visible through the bottom layer (Fig. 8) [20,848 bytes]

Plugging-induced erosion of downhole screens is very difficult to prevent.

New materials and designs can improve screen-erosion resistance to sandblasting, but the most common screen-erosion failure mechanism (at least for open hole bare-screen completions where nominal flow rates are very low) may be plugging-induced erosion.

Laboratory testing can simulate such failures, but the test conditions assume high velocities that are not representative of actual well conditions, except under conditions of extreme plugging.

Such plugging is not the norm, and there is a danger that screen designs that optimize erosion resistance for a specific laboratory test may compromise other important screen design functions.

At present, the best approach to minimize erosion is to use solids-tolerant (high void) screens, ensure that the screen media pore size is chosen to retain most of the formation sand, and ensure that the completion fluid and formation do not contain material that can permanently plug screens.

Needless to say, this requires close cooperation between the operator, service company, and screen supplier.

Screen erosion

Erosion is a major concern in the oil and gas industry, and downhole screen erosion in particular has increased in importance with the development of expensive, long reach, deepwater completions.

Erosion commonly occurs in cased and perforated high-rate gas wells in which faulty completion practices lead to massive formation damage that limits the entire flow to only a few perforations.

This leads to high localized velocities and, when combined with sand production, produces a very effective sand-blast nozzle that typically erodes holes in screens and tubulars.¹

Although high-rate impingement erosion is typically only a problem in cased and perforated gas wells, a second type of erosion has been suspected as contributing to horizontal open hole well failures.

Open hole completions theoretically have very low drawdowns and flux rates through the screens because the production is spread across the entire completion interval. But open hole completions can be damaging to screens.

The wells are typically drilled with drill-in fluids containing bridging solids to form filter cakes. Usually, the residual drill-in fluid is swapped out to a clear system before running the screens, but this is not always the case.

After the screens are run, the solids need to be removed from the well and formation. Ensuring 100% solids removal (either by dissolution, backflowing through the screen, or backwashing off the screens) is an impossible task. Even though some fluids have high-efficiency breaker systems, the insoluble drill solids produced by the drilling action will still be left behind to damage the screens once production starts.

If the screen is extensively plugged by the completion techniques, highly localized flow velocities can occur in the remaining open areas of the screen, even though the total flow per completion length is relatively low. This high-velocity sand flow is suspected of causing horizontal well failures.

Erosion-resistant screens have been introduced into the marketplace with little more than an intuitive design for erosion resistance, although studies have just begun to fundamentally understand if and how this type of erosion can occur and what mechanisms govern it.

These designs force the sand to take a tortuous path to the screen media. This path theoretically decreases particle velocity below erosive thresholds and limits direct particle impact on the media. However, many factors govern erosion, and designs maximizing resistance to direct impingement erosion do not necessarily prevent plugging-induced erosion.

Therefore, a closer look at the mechanisms of erosion and screen design is in order.

Basic erosion mechanisms

Erosion is governed by several mechanisms. The main elements are velocity, particle size, and angle of incidence. Erosion is most dependent upon velocity. Erosion rates can be expressed as:

q = Kvⁿ (1)

where:

• q = Erosion rate
• K = Proportionality constant that accounts for particle hardness and shape
• v = Particle velocity
• n = Constant.

Considering kinetic energy only, the constant n = 2. In practice, however, n appears to vary between 2.3 and 2.7 for ductile metals. This indicates that high-velocity impact causes additional mechanisms to contribute to the overall erosion rate. ²

Erosion increases with particle size. This is also a kinetic energy effect.

Fig. 1a [186,284 bytes] shows data on steel coupon erosion with quartz as the abrasive material.³ A maximum erosion rate was reached at a particle size of about 100 microns. Although kinetic energy effects should result in increased erosion at larger particle sizes, for these tests some other mechanism, such as sand fracturing, dissipated the energy that would have otherwise increased erosion rates.

A third main effect on erosion is the angle of incidence. Sand-abrasion studies on metal coupons with gas/solids particle streams yield results of which Fig. 1b is a typical example.^{2 4-6}

While the most efficient energy transfer from the particle to the surface would appear to be at angles normal (90°) to the surface, in fact, angles of about 25-30° to the surface are the most erosive. Analysis of eroded surfaces revealed that erosion for ductile materials occurs as a stepwise process of extrusion, work-hardening, and fracturing.⁴

Although 90° incidence angles can impart all of the particle's kinetic energy to the surface, there is no direct mechanism to extrude the metal. The initial extrusion mechanism is maximized at more oblique angles.

For liquid/solids stream erosion the effect is reversed. Most erosion occurs at 90° angles to the metal surface.⁴

In liquid carrier fluids, boundary lubrication effects protect the metal from erosion at oblique angles. Even so, studies on steel coupons have shown that a secondary maximum occurs at 40-45° angles.⁴ Thus erosion is minimized only in a narrow range from 50-70° and <30°.

Screen erosion mechanisms

In oil and gas well completions, screen erosion can occur from two main mechanisms:

1. Direct impact (sandblasting) erosion
2. "Wormhole" erosion as particles pass through the screen.

The sandblasting effect has been observed many times in serviced screens where a round hole has been cut into the screen.1 In most cases erosion is through the base pipe itself. This effect can occur in cased and nongravel packed completions in which dirty completion practices leave very few perforations open for flow.

The wormhole effect has been observed in both oil and gas production and is caused by highly localized fluid velocities through the screen media. This type of erosion requires very high flow velocities; therefore, it usually is only seen during high-rate water pack or frac packs that have pumping rates in the order of 15-30 bbl/min.

Hamid and Ali studied these types of failures and found that dirty completion fluids containing significant amounts of fines can erode wire-wrap screens during the end of the completion when fluid velocities through the screen are at a maximum.⁷

This type of erosion has also been observed during production in which the high velocities could have only been caused by plugged screens.

Fig. 2 [10,998 bytes] shows an erosion failure during production. The screen was the sole sand-control device for a thru-tubing recompletion in a gas well. Severe plugging by rust, scale, and clay during the workover caused the screen erosion. The screen had failed within 2 weeks after being placed back on production.

The screen appears to have the opposite type of damage than observed from sandblasting. The screen's protective cage (or shroud) had only slightly eroded, but a large section of the media and even the base pipe under the cage had completely eroded. In this sample, the rate of the gas exiting the erosion site was so high that areas of the base pipe opposite the erosion site had started to pit from direct impingement after the initial breach.

This erosion pattern indicates a relatively low particle velocity at the cage that rapidly accelerated through the media as the pressure drop increased. In this case, the small open area remaining in the plugged screen acted as a venturi to produce high particle velocities.

This effect can be qualitatively modeled using the following equation to predict pressure drop though the Stratapac screen with PMM media as a function of nominal flow rate:⁸

DP = 3.55 * 10^-4mQ₀ + 3.14 3 10^-6sp gr Q₀² (2)

where:

DP (psi) = Pressure drop across the screen m (cp) = Oil viscosity

sp gr = Oil specific gravity

Q₀ (b/d/sq ft) = Oil flux through the screen.

For illustration, assume that a horizontal well is producing 15,000 b/d through 2,000 ft of 3.5-in. screen (for simplicity, m = 6, sp gr = 1). Fig. 3 [70,144 bytes] plots the flux and DP against the relative amount of screen plugging at constant 15,000 b/d production. The clean (unplugged) flow rate per unit area is 23 b/d/sq ft at 0.04 psi differential pressure (DP). If the screen is plugged, however, there is less open area available to yield the 15,000 b/d production. Thus, both the flux (flow per unit area) and the differential pressure must increase.

For example, if the screen is 99.8% plugged, then the flux must increase to 11,500 b/d/sq ft to achieve 15,000 b/d, and the pressure drop increases to 458 psi.

The flux through the remaining open areas of the screen only significantly increases for an extensively plugged screen (99%). Not coincidentally, screens returned from the field with plugging-induced erosion effects (such as in Fig. 2) have been plugged to (and beyond) this extent.

Screen erosion tests

To study erosion mechanisms, tests were conducted on various media and shroud configurations. Direct-impingement studies used gas as the carrier fluid. Gas is ideal for direct-impingement tests because at ambient pressure gas is a low density, low-viscosity fluid that minimizes particle transport around media through fluid streamlines.

Gas also minimizes particle "cushioning" against the media from viscous boundary layers.

Plugging-induced erosion tests used oil as the carrying fluid because the particle-transport efficiency in oil is much higher than in gas, leading to better transport through the media. Also, the oil boundary layer minimizes erosion due to direct impact. This allows the tests to focus almost exclusively on erosion mechanisms of high-velocity particles through the media.

Direct impingement

Stratapac screen samples were tested with both the PMM (60 micron) media and PMF2040 (125 micron) media. The screen samples were prepared per manufacturer's design. Four layers of PMM media and two layers of PMF2040 media were sandwiched between upstream and downstream drainage meshes and covered with a perforated outer shroud.

Additional samples were prepared with four layers of PMM media and an upstream mesh under a louvered shroud, and four layers of PMM media and an upstream mesh under a proprietary experimental erosion-resistant shroud.

A base pipe was omitted because screen erosion was defined as a breach of the media.

The screens were prepared by cutting 3.5-in. discs from each layer and installing them into the test set-up (Fig. 4 [117,943 bytes]).

Each sample was impinged with sand introduced into a high-velocity air stream. The actual particle velocity was measured with a rotating shutter device. The sand nozzle was aimed directly at a perforation in the outer shroud of the screen to allow the jet to directly impinge on the drainage mesh and media.

For tests with the louvered shroud, the jet was aimed at an interface between the louvers. The jet did not have direct line-of-sight access to the media until the edges of the louvers eroded away.

A 0.0005-in. thick foil-grid sensor was positioned on the back side of each specimen and was instantly cut (opening the control circuit and shutting down the test stand) as soon as sand was produced through the test section.

The test parameters were as follows:

• Jet (nozzle) diameter: 0.25 in.
• Particle velocity: 70 fps (10 psi gun pressure)
• Contaminant: AFS 60 silica sand
• Challenge rate: 6-7 g/sec
• Impingement angle: 90°
• Nozzle to target distance: 1.0 in.

Four trials included Stratapac screen with PMM media and the PMM under the louvered shroud. Three tests were with the screen and PMF-II media, and one test included an experimental shroud.

Fig. 5 [85,078 bytes] shows the average test results. These data represent the time at which sand was produced through the test section.

The relative deviation was ±2.5 hr for the screen with PMM media and PMM media under the louvered cage, and ±7 hr for the screen with PMF2040 media.

The direct impingement erosion tests show that the PMM media has one-third more erosion resistance than the PMF2040 media. This is because the PMM is more dense than the PMF2040 (the PMF2040 media has 70% porosity compared to 50% for the PMM).

When the standard perforated shroud was replaced with a louvered shroud, the erosion rate increased by about 40%. The deflector shroud was designed to keep the particles from directly impinging upon the media. However, Fig. 1b shows that erosion is very dependent on angle of incidence.

In these tests, the deflector shield failed to protect the media because it redirected the particles from 90° to the more damaging 25-30° angle of incidence. In fact, studies on the erosion rate of steel have shown that the erosion rate at 90° impingement was about one-half the erosion rate at 25°, similar to the results observed in these tests.

Plugging-induced erosion

Plugging-induced erosion studies were conducted with an oil slurry. A high-velocity oil slurry simulated localized high-velocity oil production through the open areas of an extensively plugged screen.

Prepared media samples of relatively low-void material were compared to higher void samples to study effects of media porosity on erosion. Tests were conducted on two 47-mm disks of media in series (as per production screens).

The media was placed on a support plate, which had a similar perforation pattern to a production-screen base pipe. The screen sections were placed in a high-pressure housing and challenged with oil slurry at high flux rates. The screens were placed at 90° to the direction of flow.

To keep protective sand bridges from forming on the screens, sand was sieved from commercially available grit to remove particles greater than 45 micron (325 mesh) in size. The screens were tested in a multipass test stand.

To test the effects of shroud design on erosion, samples were prepared of two layers of media under a perforated shroud, a louvered shroud, and the experimental shroud described previously.

Media samples were prepared from coarse PMF1220 (200 micron) media to minimize sand bridging. All samples contained a drainage mesh between the shroud and media. Screen samples were tested under the following conditions:

• Flow rate: 40 gpm
• Oil viscosity: 16 cp
• Nominal face velocity: 7 fps
• Pore throat velocity: 9 fps
• Sand concentration: 10 mg/l. up to a final concentration of 500 mg/l.
• Test duration: 24 hr
• Contaminant: 325-mesh silica flour.

The screens were challenged for 24 hr. After the tests, the screens were removed and examined under a scanning electron microscope (SEM) or optical microscope.

SEM photomicrographs of the high and low-void samples are shown in Fig. 6a [17,275 bytes] and Fig. 6b [16,573 bytes]. Fig. 7 [11,796 bytes] shows the erosion pattern of PMF1220 media without any shroud, with a perforated shroud, with a louvered shroud, and with the experimental shroud.

Fig. 6 shows that the high-void sample experienced minimal erosion, but the low-void sample eroded. The high-void sample may have had relatively little erosion compared to the low-void sample because the high porosity and slightly less sand retention capability resulted in less plugging and lower pore velocity.

The low-void sample showed effects that can be attributed to "plugging-induced" erosion. This is opposite from the gas test studies where a higher density, lower-void media (PMM) had a lower erosion rate than the lower density, higher-void media (PMF-II). This illustrates the competing forces at work in screen erosion.

The shroud experiments show that a cage altered the erosion pattern. Fig. 7 shows, albeit barely discernable, that the pattern of the perforated or louvered cage is reflected in the media. This is because the reduced open area of the cages resulted in fluid velocity acceleration.

Although the screens contained drainage meshes to uniformly distribute flow across the media and slow the particles down, the test flow rates were too high for uniform drainage. At very high flow rates, flow through the drainage meshes enters the turbulent flow regime, whereas the flow remains laminar through the media. Therefore, the pressure drop increases through the drainage mesh relative to the media, resulting in concentrated flow under the perforations or louvers.

The overall erosion rate did not visually appear significantly different between the two shroud types.

The experimental erosion-resistant shroud that offered significant protection in the direct impingement gas tests offered no increased protection to the media in the high-velocity, liquid-flow tests. Again, the reduced open areas of the shroud resulted in high localized fluid velocities that increased the media erosion rate in selected areas.

It is interesting to note that none of the shrouds or the upstream drainage meshes experienced significant erosion even though the media underneath the shrouds had pronounced erosion. Some rough edges made by the stamping process remained in the shrouds.

Additionally, for all samples the bottom media layer was more eroded than the top media layer (Fig. 8 [20,848 bytes]). This indicated that none of the shroud designs minimized high pore velocities.

The liquid/solid phase erosion tests were run at unrealistically high fluxes to simulate localized, high-velocity effects of an almost totally plugged screen. However, although the tests show erosion effects due to high-velocity, sand-laden fluids, these results are difficult to correlate to downhole conditions.

Going back to the example presented previously, a well producing 15,000 b/d through 2,000 ft. of 3.5-in. screen has a nominal, clean flux of 23 b/d/sq ft or 0.0159 gpm/sq ft (assuming uniform flow).

The erosion test conditions are nominally 4,000 gpm/sq ft. Therefore, massive amounts of plugging are required to achieve erosion velocities.

In addition to the extremely high flux rate, the tests were run with a finely sieved contaminant specifically to prevent the screen from plugging and preventing flow. Tests with contaminants containing more-natural particle size distributions quickly ran the pressure up to 2,000 psi differential across the screen resulting in very little or no flow.

While tests have shown that coarse wire-wrap screen can produce sand amounts enough to cut the screens, it is still unclear how fine prepacks or composite metal screens can erode from sand without prior formation of protective sand bridges.

Recommendations

Although advanced shroud designs can protect the media from a persistent sand-blast jet, minimizing formation damage with nondamaging mud systems and 0.5 micron or 2 micron (bx 5,000) filtered completion fluids may prevent such conditions from occurring in the first place. The Beta ratio, bx, means that only one particle of size x from more than 5,000 will pass through the screen.

This same approach to cleanliness will also minimize the risk of plugging-induced screen erosion.

Proper hole cleaning and mud conditioning will minimize screen plugging, and proper media selection to match the formation sand size will minimize highly localized, high particle velocities through the media.

If these steps are followed, then during actual well production the reservoir sand should form a protective and permeable sand bridge long before the onset of erosion.

Based on the experimental tests and field experience, the following are recommendations for minimizing erosion risk:

• Avoid deflection shields-Direct impingement erosion is a function of the angle of incidence the sand makes with the screen. Deflection shields only serve to direct the sand into the most damaging angle of incidence for gas/solids streams, and show no enhanced protection in liquid/solids streams over perforated shrouds.

• Use multilayer screens-Direct impingement erosion is a function of the amount of metal in the screens. Therefore, dense or thick materials will retard erosion longer than light or thinner materials. However, field history on screens subjected to sandblasting have shown that even the base pipe can erode; therefore, heavy screen materials may only offer minimal increase in service life.

Dense media (such as wire-wrap screens) pose a serious drawback because they typically have low-void volumes. High-void volume is important to resist plugging and minimize the risk of plugging-induced erosion. Therefore, screen designs with multiple layers of high-void media provide the best combination to retard direct impingement erosion and resist plugging-induced erosion.

• Use high-void media-To minimizing plugging-induced erosion, significant screen areas must be open for flow. This will minimize particle velocity through the media.

High-void media, also refered to as high inflow area, will reduce the chances of plugging. Note that high-void materials must be used in multiple layer screens. Layering with low-void media results in very restrictive fluid pathways that can cause high internal particle velocities.

The same recommendation can be made for the protective cages. Cages with small open areas or narrow pathways speed up the fluid and particle velocities.

High-void media should not be confused with large pore-size screens that resist plugging solely due to a lack of solids control. In fact, if poorly matched to the reservoir sands, these screens can be more-easily eroded than screens having better sand-retention capability.

For example, an industry study of various screens showed that of all screens tested, only the screen with the largest slot size (over twice the average particle size) eroded because it failed to retain the majority of the test sand.⁹

This is also the erosion effect that Hamid and Ali studied for high-rate water pack and frac-pack completions.⁷

• Meticulously clean holes and use nondamaging drill-in fluids-Even the best screen design cannot compensate for dirty completion practices. The hole should be very clean before running the screens. This involves cleaning with chemicals and/or trips with casing brushes.

For open hole completions or completions where fluid-loss pills are spotted inside the screen, the fluid-loss material must not contain significant amounts of drill solids and must be removed as completely as possible.

The best fluids have internal breakers that can effectively contact the entire cake. Backflowed filter cakes, while operationally appealing, can plug screens if they come off the formation in aggregates.

• Gravel pack-Nothing prevents operational problems like a gravel pack, and high-rate, cased gas wells should always be gravel packed. However, the majority of thru-tubing workovers are not gravel packed.

For horizontal wells, a carefully executed, complete gravel pack can also avoid failures. However, a poor gravel pack has a higher probability for voids and cannot overcome poor well cleaning and completion techniques. In fact, the opposite is often true because a gravel pack can lock-in dirty fluids and unbroken drill-in fluid systems, leading to very high skins and poor productivity.

Therefore, the above recommendations should be followed even for gravel-packed completions.

Acknowledgments

The authors thank Clive Bennett of BPX, and Steve Svedeman at Southwest Research Institute for information on liquid-phase erosion testing. The authors also thank Pall Corp. and Chevron USA Production Co. for permission to publish this article.

References

1. Suman, G.O., Ellis, R.C. and Snyder, R.E., Sand Control Handbook, Gulf Publishing Co., Houston, 1985.
2. Birchenough, P.M., "Recognizing the Importance of Fluid Mechanics in the Understanding of Erosion Problems," IBC Multiphase Offshore Conference, London, 1991.
3. Tilly, G.P., "A Two Stage Mechanism of Ductile Erosion," Wear, Vol. 23, pp. 87-96, 1973.
4. Levy, Alan, Solid Particle Erosion and Erosion Corrosion of Materials, ASM International, 1995.
5. Finnie, I., "Erosion of Surfaces by Solid Particles," Wear, Vol. 3, 87-103, 1960.
6. Rabinowicz, E., "The Wear Equation for Erosion of Metals by Abrasive Particles," 5th International Conference on Erosion by Liquid and Solid Impact, 1979.
7. Hamid, S., and Ali, S.A., "Causes of Sand Control Screen Failures and Their Remedies," SPE Paper No. 38190, SPE European Formation Damage Conference, The Hague, Netherlands, June 2-3, 1997.
8. Lester, G.S., Malbrel, C.A., and Whitlock, M.B., "Field Application of a New Cleanable and Damage Tolerant Downhole Screen," Paper No. SPE 30132, SPE European Formation Damage Conference, The Hague, Netherlands, May 15-16, 1995.
9. Ali, S.A., and Dearing, H.L., "Sand Control Screens Exhibit Degrees of Plugging," Petroleum Engineer International, Vol. 69, No. 7, July 1996.

The Authors