OTC-19811-MS-P

OTC 19811 New Inflow Control Device Reduces Fluid Viscosity Sensitivity and Maintains Erosion Resistance Martin P. Coronado, SPE, Luis A. Garcia, SPE, Ronnie D. Russell, SPE, Gonzalo A. Garcia, SPE, and Elmer R. Peterson, Baker Hughes Incorporated

Copyright 2009, Offshore Technology Conference This paper was prepared for presentation at the 2009 Offshore Technology Conference held in Houston, Texas, USA, 4–7 May 2009. This paper was selected for presentation by an OTC program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been reviewed by the Offshore Technology Conference and are subject to correction by the author(s). The material does not necessarily reflect any position of the Offshore Technology Conference, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Offshore Technology Conference is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of OTC copyright.

Abstract In long horizontal wells, production rate is typically higher at the heel of the well than at the toe. The resulting imbalanced production profile may cause early water or gas breakthrough into the wellbore. Once coning occurs, well production may be severely decreased due to limited flow contribution from the toe. To eliminate this imbalance, inflow control devices (ICDs) are placed in each screen joint to balance the production influx profile across the entire lateral length and to compensate for permeability variation. Pressure drop in an ICD is created through either restriction or friction mechanisms. Restriction mechanisms rely on a contraction of the fluid flow path to generate an instantaneous pressure drop, resulting in higher velocities, and are thus more prone to long-term erosion damage as well as plugging during mud flowback. A restriction device, however, is less sensitive to viscosity properties of the fluid. A frictional device, which creates a pressure drop over a distributed length, is less likely to erode due to lower fluid velocities, but is more sensitive to viscosity changes. Viscosity insensitivity is desired to minimize preferential water flow whenever water breaks through into the well. This paper will detail the development of a new hybrid design concept that uses the best features of the restricting and friction designs, while minimizing the less desirable characteristics. Because these ICDs are permanent downhole components, their long-term reliability is imperative, and these new developments will improve their resistance to erosion and their ability to effectively balance inflow. Conceptual fluid dynamics analysis was used extensively to characterize the new design, along with actual full-scale flow testing. Introduction The purpose of inflow control devices (ICDs) is to effectively balance well production throughout the entire operational life of the completion to optimize hydrocarbon recovery. Since a typical well with ICDs can be in production from 5 to >20 years, the long-term reliability of such a device is crucial to the well’s overall success. The significant factor in the reliability of an ICD is its ability to maintain a uniform influx over the well life. If an ICD is not able to maintain a uniform flux rate, increased localized production rates will occur and the well will become unbalanced. This will render the ICD ineffective, leading to premature water and/or gas breakthrough and possible loss of sand control. At some stage in a well’s life, water may break through into the wellbore in certain sections due to heterogeneity of the formation and/or vertical fractures. Ideally, once this occurs, flow contribution from these water-producing zones should not be greater than the oil-producing sections. In production wells with higher-viscosity oil (>10 cp.), ICD type selection becomes a more critical factor due to the larger difference in viscosity between the oil and produced water. The pressure reduction mechanism in an ICD in this situation must have the lowest sensitivity to viscosity to maintain an even flow profile across the entire lateral wellbore. A restrictive-type ICD thus will provide best results in this regard due to its lower sensitivity to viscosity. This type of ICD however, has a greater potential for long-term erosion and lower plugging resistance. The ideal solution is to provide the lower viscosity sensitivity of the restrictive device with the lower erosion and higher plugging resistance of the frictional design. This means using the restrictive pressure loss mechanism while limiting the fluid velocity through the device below the critical level which will cause erosion. Limiting the fluid velocity also can result in increased minimum flow area if configured efficiently. The primary factor in maintaining a uniform influx is the ability of the device to resist erosion from fluid-borne particles that pass through the screen. Screens are not designed to prevent 100% blockage of all particles from the formation. During

2

OTC 19811

production, formation fines that are produced through the screen also pass through the ICD. These fines can and will erode an ICD over time if the fluid velocity is high enough and fines are in the flow stream. The rate of erosion will depend on the following factors: particle size, particle concentration, and fluid velocity. The first two factors are dependent on well conditions, while the third is dependent on ICD geometry and design. Characteristics of Currect ICD Types Currently, there are two different types of ICD designs in the industry: orifice/nozzle-based (restrictive) and helicalchannel/labyrinth pathway (frictional). They use two different methods to achieve a uniform inflow profile. The orifice-based ICD uses fluid constriction to generate a differential pressure across the device. This method essentially forces the fluid from a larger area down through small diameter ports, creating a flow resistance. This overall change in pressure is what allows the ICD to function. The helical-channel (Fig. 1) and labyrinth pathway ICDs, however, use surface friction to generate a similar pressure drop. The helical channel design is one or more flow channels that are wrapped around the basepipe of the screen. The labyrinth design uses a tortuous pathway to create a pressure drop which makes the fluid change directions numerous times while transversing through the device. These designs provide for a distributed pressure drop over a relatively long area, versus the instantaneous loss using an orifice. Using friction to create a flow resistance allows the use of a channel with a larger cross-sectional area than an orifice-based ICD. When fluid flows through the channel or channels, fluid rheology and channel characteristics interact to create the designed pressure drop. Since typical wells using ICDs can be in production from 5 to more than 20 years, the long-term reliability of such a device is crucial to the well’s overall success. If an ICD is not able to maintain a uniform flow in a completion using multiple ICDs, increased localized production will occur and lead to premature water or gas breakthrough. The benefits of an orifice- or nozzle-based ICD is its simplified design and easier adjustment immediately before running in a well should real-time data collected during drilling the well indicate the need to change flow resistance. The disadvantage of the orifice- or nozzle-based ICD is the small diameter ports required to create flow resistance, which make it prone to both erosion from high-velocity fluid-borne particles during production and susceptible to plugging, especially during any period where mud flow back occurs. Because the larger cross-sectional flow area of the helical-channel ICD generates significantly lower fluid velocity than the ports of an orifice-based ICD with the same flow resistance, the helical-channel ICD is more resistant to erosion from fluid-borne particles and resistant to plugging during mud flow back operations. The disadvantage of the helical-channel ICD is its flow resistance is more viscosity-dependent than the orifice- or nozzle-based ICD. This characteristic could allow preferential water flow should premature water breakthrough occur. Hybrid ICD Design A new hybrid ICD design incorporates a series of flow slots in a maze pattern (Fig. 2). The primary pressure drop mechanism is restrictive, but in a distributive configuration. A series of bulkheads are incorporated in the design, each of which has two flow slots cut at 180° angular spacing. As with all ICD designs on the market, the ICD is located downsream of the filter unit and upstream of the inlet flow ports in the sand screen base pipe (Fig. 3). As the production flow passes each successive chamber that is formed by the bulkheads, a pressure drop is incurred. Pressure is reduced sequentially as flow passes through each section of the ICD. Each set of dual flow slots are staggered with the next set of slots with a 90° phase thus the flow must turn after passing through each set of slots. This prevents any jetting effect on the flowpath of the downstream set of slots which may induce turbulence. Without the need to generate the pressure drop instantaneously, as with an orifice, the flow areas through the slots are relatively large when compared to the orifice design, thus dramatically reducing erosion and plugging potential. Comparing the minimum flow area through the hybrid design to the existing helical design shows the hybrid design to be only slightly less than a three-channel ICD, but greater than a two-channel design. Since the amount of actual field experience with the helical design indicates no plugging problems during initial well cleanup and production, it is anticipated the hybrid design will also be plugging-resistant. Both the helical and hybrid designs have much larger minimum flow area than an orifice design with similar pressure drop. ICDs are distinguished by their flow resistance rating (FRR), which indicates the total amount of pressure drop through the device with a reference fluid property and flow rate. The design of the new hybrid ICD makes adjustment to different resistance ratings simple, since the cumulative pressure drop through the device is proportional to the number of bulkhead/slot sections used. This configuration is also easily scalable to different screen sizes as the pressure loss between slot pairs is negligible. This essentially makes the hybrid design screen size-independent. CFD Analysis and Modeling CFD Model Development Background. Flow affects most oilfield equipment due to the nature of the challenge to produce and inject fluids through the well and through formations. Full-scale testing has been and still is the final verification that a system works and meets all specifications. The cost of testing is dramatically offset through the use of computational simulation software.

OTC 19811

3

Computational fluid dynamics (CFD) flow simulations for completions systems are an integral step in the design process for sand control system developments. Creating systems that are optimized for performance in a short time frame necessitates the ability to accurately simulate systems early in the conceptual stage of development. CFD modeling based upon the Navier-Stokes equation and various turbulence modeling approaches has been well established for years now. Flow predictions using CFD have led to fairly accurate predictions of flow coefficients. Flow coefficients have been simulated using CFD for flow control completion systems for about the last ten years within Baker Hughes and before that through experts from consulting firms and universities. Recent developments in accuracy of sand particle modeling and erosion prediction have led to further value achieving optimized solutions quickly with lower cost. The current economic climate and low price of oil add further drive to construct new simulation models to simulate the growing challenges of more aggressive well environments. A breakthrough in erosion CFD modeling was made in 2003 for simulating water injection and production tool lifetimes. Before 2003, erosion CFD modeling was primarily for qualitative purposes only. Full-scale verification erosion testing of downhole valves since 2003 has given additional confidence that our erosion CFD models now have good accuracy compared to year’s prior1,2,3. Additional erosion models have been developed in 2006 and 2008 and additional efforts are to make further improvements in accuracy with high sand concentrations and non-Newtonian fluids. CFD modeling of the new hybrid ICD was implemented extensively beginning in the conceptual design development for rapid optimization. This led to a first design that was highly optimized and exemplary in prototype testing. CFD in Hybrid ICD Design. Various ICD concepts and flow restrictor shapes were studied and modeled using CFD to determine effectiveness of each design. However, before using CFD, spreadsheet calculations or nodal analysis for combinations of pipe flow, orifice flow, and restrictor flows of varying types quickly narrow down ICD design concepts for the next step of detailed CFD flow analysis. Initial CFD models of a concept that appears promising give valuable information to facilitate design optimization of an ICD concept. CFD model results can determine flow coefficients to use in spreadsheet models or nodal analysis models to rapidly simulate and optimize a design to meet specified requirements. Design specifications for ICDs typically include size, pressure drop, flow rates, and viscosity ranges. Developing spreadsheet models of restrictor concepts allows more team members to participate in optimizing a design by trying their own design variations. Understanding velocity magnitudes and flow trajectories is crucial when optimizing an erosion- and plugging-resistant design. Figs. 4 and 5 display the velocity magnitude contours through the hybrid ICD for two cases using water and heavy oil (88cP). These figures show that the peak velocities are at the center of the flow slots, as expected, but then dissipate quickly before reaching the opposing face. The distance between the plates was refined to provide enough standoff for erosion resistance as well as a “self-cleaning” and anti-plugging design. Figs. 6 and 7 reflect the same two cases but display the flow velocity trajectories through the device. As shown, the flow slot configuration allows for a natural swirling effect in between stages which minimizes any stagnant areas where particle settling could occur. As mentioned, CFD modeling was used to support the spreadsheet calculations and nodal analysis since the conceptual stage. In the spreadsheet calculations, the number of stages is a factor for the total pressure drop across the device. Essentially, each stage has an equal contribution to the performance of the ICD. This is further discussed in the Model Verification section and is graphically displayed in Figs. 8 and 9. CFD Optimization and Automation. CFD automation can facilitate the study of flow characteristics of a design. Automation of CFD models added substantially to the hybrid ICD optimization process since CFD flow models are generally time-consuming if many runs are built and run manually step by step. Mesh and CFD model journal files driven from spreadsheet design inputs were created that automated CFD mesh creation and CFD flow runs for many flow rates and multiple fluids. Using this automation process for a particular hybrid ICD design allowed 80 CFD runs to be run in only a few days’ time. CFD automation led to the development of complete flow coefficient versus Reynolds number curves. These new flow coefficient curves were incorporated in spreadsheets to further improve hybrid ICD spreadsheet calculation models to give a nearly exact match to the CFD models for the entire flow range simulated. From this stage on, further optimization or new designs of the same type of hybrid ICD can be done entirely using these spreadsheet calculation programs. CFD Erosion Modeling. The hybrid ICD design benefits from features found in both restriction and friction designs allowing it to maintain the exceptional erosion resistance of proven frictional ICD designs and have the viscosity insensitivity of restriction designs. CFD erosion models were run of the hybrid ICD and showed that erosion rates would not be measurable for the entire flow range of the design specifications. Model Verification The mathematical model and hybrid ICD design were validated by a series of full-scale flow tests that covered a wide span of viscosities and flow rates. The test series used fluids with viscosities ranging from 1-200cP. Fluid types included a solvent, water, and mineral oils of varying viscosities. Fluids were circulated through the system up to 325 bpd (9.5 gpm).

4

OTC 19811

For comparative purposes, two orifice-type ICDs were also tested under the same conditions, one axial flowing and one radial flowing. The orifice-type devices will be referred to as orifice plate and orifice plugs respectively. Test objectives included: mathematical model validation by comparing test data points to the models calculated pressure drop values, quantify the viscous (friction) sensitivity of the hybrid ICD and orifice-type devices, and to validate the new hybrid design. As mentioned previously, the hybrid ICD’s total pressure drop is directly proportional to the number of plates in the design. Flow areas through and between the plates are kept constant so that the pressure drop rating can be increased or decreased simply by adding or removing plates. This is a key difference between the hybrid design and standard orifice designs. Standard orifice designs have to vary the minimum flow area through the device to alter the pressure drop while the hybrid design maintains a consistent flow area. This allows for low-velocity, plugging-resistant solutions. Pressure transducers were placed at various points along the hybrid ICD to measure uniformity of the pressure drop along the device. The test assembly consisted of a full-scale 5-1/2-in. hybrid ICD (modular) mounted on a premium screen. The screen joint had a 3-foot-long screen cartridge and included production geometries such as the base pipe undercuts and adapter ring (Figs. 10 and 11). For proper comparison of the devices, they were each designed to have an equivalent FRR. The FRR is essentially a pressure drop rating at a given set of fluid viscosity, fluid density and flow rate conditions. Each of the orifice devices had an FRR of 0.8. The hybrid ICD pressure drop was measured at various points along the device which corresponded to 0.4, 0.8, 1.2 and 1.6 FRR. Fig. 12 compares the change in pressure drop relative to water as a function of viscosity. The orifice and hybrid ICD designs exhibit a decrease in pressure as the viscosity is increased. This is due to the lower densities of the oils and supports that, in some instances, higher-viscosity fluids would have a preferential flow over water. The point where the curves cross the X-axis is where the higher-viscosity fluid has the same pressure drop as water at a constant flow rate of 188 bpd. The orifice plate crosses the X-axis at the 8cP mark, the orifice plugs at 25cP and the hybrid ICD at 60cP. This indicates that for the designs tested, the hybrid ICD design is the most “insensitive” to viscosity variations. In a well scenario, the hybrid ICD would exhibit an increase in resistance to flow if premature water breakthrough were to occur. Fig. 13 compares the pressure drop predictions for the 1.6-FRR hybrid ICD as a function of fluid viscosity and density and is overlaid with the test data points. For the fluids and rates tested, all the mathematical predictions fall within an experimental error band of ten percent. Erosion Testing Understanding the erosion resistance of different types of ICDs is crucial when planning a completion. The completion must be robust enough to offer years of service without compromising pressure drop due to erosion of the ICDs, which would lead to non-uniform influx into the well. Orifice-type ICDs are most susceptible to this type of failure due to the high velocities required to create the instantaneous pressure drop across the device. Formation fines that pass through the screen media can decrease the performance of the ICD over a period of time by eroding the geometry of the restriction, thus increasing the flow area. The hybrid design was tested under the same conditions as orifice-type ICDs as described in a previous lab study4. In this test, silicon carbide (SiC) particles were flowed through the hybrid ICD at a constant rate of 188 bpd (5.5 gpm), until a total of 500 lb had passed through the device. This volume and size of the sand, as well as the flow rate, was determined based on anticipated total sand production through a single screen joint referenced in previous papers5,6. The erosion resistance of the device is displayed by comparing a pressure profile with water from before and after the test. Test System Design. A full-scale hybrid ICD was mounted on a 6-5/8-in. base pipe in the same fashion as a production unit. The device was rated at 1.6 FRR and was placed in a flow fixture as shown in the test schematic (Fig. 14). The fixture had a 30° tapered entrance, which directed the slurry into the ICD. Upon passing through the ICD, the slurry then traveled through the perforations in the base pipe and into the return line. A differential pressure transducer was placed across the inlet and outlet of the ICD fixture to measure the pressure drop across the device. Since a typical well may be in production for over 20 years, it is imperative to test the ICDs in a way that will reflect long-term exposure to formation fines. The particle concentration and procedure used in this test is equivalent to an ICD Erosion Test Program conducted at Southwest Research Institute7. A slurry consisting of water and SiC particles was used and circulated through the system at a concentration of 2,500 ppm (parts per million). The slurry concentration was monitored and kept constant throughout the duration of the test, adding additional SiC when needed to accommodate for any settling in the flow lines or equipment. The particle size was between 53-63 μm and is representative of what could be seen during production, even when screens are in place. Test Procedure. A FMC triplex bean pump was used to pump the SiC slurry through the ICD. Due to the abrasiveness of the SiC particles, the test was paused on two occasions to redress the steel plungers and packing elements. A pulsation dampener was placed on the outlet of the pump to reduce pressure fluctuations typically seen in triplex applications. The flow rate used for the test was 5.5 gpm. Before starting the test, clean water was pumped through the fixture in a “ramp” method. Starting at 1.0 gpm, the flow rate was increased in 1.0-gpm increments holding for one minute at each rate. This data was used to create a pressure profile curve for the device using water and would later be compared to a post-test curve to help detect any signs of erosion. The test was run until a total of 500 lb of SiC had passed through the hybrid ICD. During the test an additional data point was captured at the 400-lb mark.

OTC 19811

5

Test Results. Upon completion of the test, the system was flushed and cleaned. Water was then circulated as mentioned previously and a post-test pressure profile generated. All three data points (pretest, 400 lb and 500 lb) are compared in Fig. 15. No measurable difference in pressure drop can be made at 5.5 gpm for the hybrid ICD. Different types of orifices tested under the same conditions resulted in differential pressure decreases between 7.8%-14.9% at 5.5 gpm. The high erosionresistance level of the hybrid ICD is attributed to its enlarged flow areas and slot configuration, which reduce velocities and friction losses through the device. A post-test inspection of the hybrid ICD (Figs. 16-18) showed no visual signs of erosion or geometry changes. The surface finish appeared to be visibly smoother in areas where changes in flow direction occur, but since the hybrid ICD is not a friction-based unit, this has no impact on its pressure drop or performance. As noted in the previous ICD erosion study, at comparable flow rates orifices showed a significant amount of material removal at the entrances, which resulted in a decrease in performance. ICD Flow Performance Characterization During the past decade, different ICD geometries (helical channel, labyrinth, orifice and recently the hybrid design) have been designed to improve horizontal-well flow performance as well as reservoir behavior. The ICD geometry design and flow-resistance rating selection have been based on the pressure drop models that describe the flow performance through each ICD. Those pressure drop models were developed based on analytical, numerical laboratory test data, or a combination thereof. In order to improve the current ICD knowledge in terms of pressure drops, as well as generated and unify comparison criteria, four ICD geometries were flow-tested as described previously. The data gathered during these laboratory tests show the pressure drop ICD performance is a function of the fluid viscosity8. The ICDs evaluated were designed to have the same FRR at 26 sm3/d water flow rate (1 specific gravity and 1 cP). However, as is observed, not all the geometries have the same flow performance characteristic. Therefore, it is not straight forward to determine which ICD geometry would offer a better performance under specific operational conditions. Fig. 19 shows the pressure drop versus flow rate for the ICDs tested. Each ICD was tested with different fluid viscosities (0.85, 18, 53, 82 and 200 cP). As can be observed, the hybrid design is less fluid-viscosity-independent than the other ICDs tested. Other key performance characteristics also need to be considered when selecting the proper ICD configuration for an application. The flow area plays a very important role in the ICD selection for high-viscous fluid since it would affect the plugging probability (Fig. 20). Also, some ICD designs (e.g. orifice design) can produce less pressure drop for high viscosity but the flow velocity could be four times higher than the helical or hybrid design (Fig. 21). A comprehensive and systematic approach9 was implemented to determine the flow characteristic of each ICD. Fundamentally, the approach used is based on the estimation of the pressure loss coefficient (called K or flow resistance) for each flow test measured at the laboratory. The pressure drop along pipeline accessories (e.g. valves, elbows) can be described with Eq. 1. Laboratory test and data analysis have validated that the K values for ICDs can also be described as a function of Reynolds Number (Fig. 4).

Δpliq

2 ⎞ ⎛ vliq ⎟ ⎜ = K ρl ⎜ 2 g 144 ⎟ ⎠ ⎝ c

Equation 1

The pressure loss coefficients, for each ICD and flow test, were determined using Eq. 2 which is derived from Eq. 1. It means that for each flow test (geometry, fluid properties and operational conditions) there is a K value that describes the equipment tested.

⎛ Δpliq K =⎜ 2 ⎜ q ⎝ liq

⎞⎛ 1 ⎟⎜ ⎟⎜ ρ ⎠⎝ l

⎞ 2 ⎟⎟ A (2 g c 144) ⎠

Equation 2

Fig. 22 shows the relationship of K versus Reynolds number for four ICDs tested (green points). As can be observed the K values follows a trend as a function of the Reynolds number. This means that the pressure loss coefficient can be described as a function of the Reynolds number and ICD geometry type. Different correlations10 were evaluated to describe the K performance and the best to fit the data gathered is presented in Eq. 3.

6

OTC 19811

K = K lowR +

K highR

(K

highR

+ K lowR ) c

⎛ Re ⎞ (1 + ⎜ ⎟ ) d ⎝ t ⎠ b1 = a1 Re

Equation 3

K lowR = a2 Re b2 Eq. 3 is a function of Reynolds number and seven parameters (a1, b1, a2, b2, c, t and d). Each ICD geometry can be described with the correspond parameters. A regression analysis is required to determine the seven parameters that describe the K correlation. The red lines observed in Fig. 22 represent the fitting curve of the laboratory data which was obtained using Eq. 3. As can be observed, all the geometries evaluated are a function of the Reynolds number, and therefore the fluid viscosity. All ICDs have some sensitivity to fluid viscosity. Some of them are less fluid-viscosity-insensitive but cannot offer the FRR required to balance the influx along the horizontal section. One important observation can be made. The hybrid design offers more flow resistance as soon as the Reynolds number achieves the value of 104. This means that as soon as water breaks through into the well the hybrid design would be able to provide more flow resistance. This approach offers a common ICD dimensionless performance curve11, a better flow behavior understanding, and options to optimize the flow performance as well as selection of the proper ICD. Can we talk the same language when comparing ICD functioning and performance? Yes, we can. It would be an excellent opportunity for the oil and gas community to have a uniform, comprehensive, and systematic approach to describe the ICD in a dimensionless form. This would aid in standardizing ICD flow performance flow characteristics as well as their implementation in third-parties software. The estimation of ICD benefits can be determined from three points of view: uniform influx, maximize recovery, or water/gas delay. The uniform influx can be analyzed with the discretion of the well completion and applying a downhole network analysis, but the recovery and water delay require coupling the downhole network analysis with a reservoir simulator. Well Completion Effects on Reservoir Behavior The ICD flow performance characteristic, described previously, was implemented in a third-party reservoir simulator called Reveal to quantify the geometry effect on the production forecast. The well completion description in terms of annulus flow, gravel pack, screen basepipe, packers, blank liner, and ICD were included in the modeling of the well completion effects. All those components would determine the best well completion accessories to be implemented in order to have better management of the production fluids. In order to quantify the well completion effects in terms of ICD geometry12 a theoretical reservoir model was defined. The control volume is shown in Fig. 23. A uniform block size and uniform reservoir properties (porosity – 29% and permeability – 8 Darcy) were used in order to simplify the modeling13,14. The bock size was 160 x 200 x 10 ft. An aquifer was considered at the bottom of the control volume. Fig. 23 considers two control volumes with similar reservoir and fluid properties as well as operational condition. The well completion types evaluated were: standard screen without ICD (negligible pressure drop), orifice design ICD and hybrid design ICD. The hole size was 8 ½-in., 5-in. base pipe & 6.54-in. OD screen. The pressure drop flow performance characteristic used to describe the orifice and hybrid design is shown in Fig. 22 (red line). The simulation results are shown in the Figs. 24-27. Fig. 24 shows the fluid movements in the reservoir into the wellbore after 16.5 days of production. Early water breakthrough at the heel of the well with the orifice radial design ICD can be clearly observed. The hybrid design ICD is able to delay the water breakthrough as well as provide a uniform influx along the horizontal section. Fig. 25 shows several cross-section plots (heel, close heel, close toe and toe) along the horizontal well after 186 days of production. Fluid production at the heel is observed in both cases with less fluid into the well at the toe through the orifice radial design ICD. Both ICD designs have the same FRR but the orifice design does not have enough pressure loss coefficient to control the flow at the heel (see Fig. 22). Oil and water production forecast for standard screen, orifice, and hybrid designs can be observed in Fig. 26. The main observation obtained from this figure is the highest oil production and lowest water production, as well as increased water delay using the hybrid design. This reservoir fluid behavior is due mainly to the flow performance characteristic shown in Fig. 22, which illustrates that at higher Reynolds number the slope is increased, adding more flow resistance to this operational condition. The fluid viscosity evaluated in this case was 4cP. The hybrid flow effect on the reservoir behavior is independent of the fluid viscosity. Fig. 27 shows the cumulative oil and water production for three screen types and different fluid viscosities (1, 4, 20 and 80cP). The results shown the hybrid

OTC 19811

7

design would be able to produce higher cumulative oil production and lower cumulative water production with any of the fluid viscosities evaluated, with the difference reducing as fluid viscosity increases. Conclusions • A new ICD design has been developed and tested which provides improved fluid viscosity insensitivity over current designs and also maintains low erosion potential (lower flow velocities) and high plugging resistance (large flow area). • Testing data has been developed which allows direct comparison of existing frictional (helical channel) and restrictive (orifice) performance at various flow rates and fluid viscosities. • CFD models have been constructed and verified to predict accurate performance of ICDs under varying conditions. • A new comprehensive and systematic approach was developed to describe ICD characteristics. • ICD performance can be described mathematically where the geometries are characterized with seven parameters. • The new hybrid design offers a particular flow performance characteristic with a reduced pressure loss coefficient at lower Reynolds numbers and increasing flow resistance at higher Reynolds numbers. Therefore, once water breaks through into the well, the hybrid design will add more flow resistance. • The reservoir flow behavior observed in the reservoir simulation with the hybrid design is independent of the fluid viscosity (more oil and less water production). Acknowledgments The authors would like to thank the management of Baker Oil Tools, Baker Hughes Inc., for their support and permission to publish this paper and Petroleum Expert for their cooperation and continuous support in the ICD integration in their reservoir simulator. SI Metric Conversion Factors psi × 6.894757 E+00 = lbf × 4.448222 E+00 = in × 2.54* E+00 =

kPa N cm

*conversion factor is exact.

References 1. McCasland, M., Barrilleaux, M., Gai, H., Russell, R., Schneider, D. and Luce, T., “Predicting and Mitigating Erosion of Downhole Flow-Control Equipment in Water-Injector Completions,” paper SPE 90179 prepared for the SPE Annual Technical Conference and Exhibition held in Houston, Texas, U.S.A., September 2004. 2. Russell, R., Barrilleaux, M., Gai, H., Macrae, J. and Luce, T., “Design, Analysis, and Full-Scale Erosion Testing of a Downhole Flow Control Device for High Rate Water Injection Wells,” paper SPE 90759 prepared for the SPE Annual Technical Conference and Exhibition held in Houston, Texas, U.S.A., September 2004. 3. Russell, R., Shirazi, S. and Macrae, J., “A New Computational Fluid-Dynamics Model to Predict Flow Profiles and Erosion Rates in Downhole Completion Equipment,” paper SPE 90734 prepared for the SPE Annual Technical Conference and Exhibition held in Houston, Texas, U.S.A., September 2004. 4. Visosky, J., Clem, N., Coronado, M. and Peterson, E.: “Examining Erosion Potential of Various Inflow Control Devices to Determine Duration of Performance,” paper SPE 110667 presented at the 2007 SPE Annual Technical Conference and Exhibition, Anaheim, CA, Nov. 11-14. 5. Vazir, H., Allam, R., Kidd, G., Bennett, C., Grose, T. Robinson, P., and Malyn, J.: “Sanding: A Rigorous Examination of the Interplay Between Drawdown, Depletion, Start-up Frequency and Water Cut”, paper SPE 89895 presented at the 2004 SPE Annual Technical Conference and Exhibition, Houston, TX, Sept. 26-29. 6. Svedman, S., and Bennett, C.: “Sand Control Screen Erosion Industry Joint Venture,” Southwest Research Institute Project 04-8560, San Antonio, May 1998. 7. Siebenaler, S.: “Erosion Testing of Inflow Control Devices,”Southwest Research Institute Project 18.12922, San Antonio, April 2007. 8. Garcia Luis. Test Report No. 1A1-166-02C. “Inflow Control Device (ICD) Performance Testing”. November 2008. 9. Garcia Gonzalo. “ICD Flow Characterization”. October 2008. Houston TX. Technical Service Sand Control Report. 10. F.Garcia, R. Garcia, et al. “Power law and composite power law friction factor correlations for laminar and turbulent gas-liquid flow in horizontal pipelines”. International Journal of Multiphase Flow 29 (2003) 1605-1624. 11. Gavioli Paolo, Garcia Gonzalo and Garcia Luis. “Evaluating Four Types of Passive Inflow Control Devices ICD Performance Curves”. Passive Inflow Control Technology Forum. 24th October 2008. Perth, Australia. 12. D.S. Qudalhy et al. ”Improving Horizontal-Well Productivity Using Novel Technology and Optimization of Drilling Fluid”. 2005 SPE Drilling & Completion. September 2005. 13. Augustine, J. ”An Investigation of the Economic Benefit of Inflow Control on Horizontal Well Completions Using a ReservoirWellbore Coupled Model”, paper SPE 78293 presented at the 13th European Petroleum Conference, Aberdeen, Scctland, October 2931, 2002. 14. Augustine, J. and Ratterman, E.: “Advanced Completion Technology Creates a New Reality for Common Oil Field Myths”, paper SPE 100316 presented at the 2006 Europec/EAGE Annual Conference and Exhibition, Vienna, Austria, June 12-15.

8

Figure1: Existing helical-channel ICD design creates the desired flowing pressure loss via friction as fluid passes through the channel(s).

OTC 19811

Figure 2: New hybrid ICD design uses a distributive restrictive geometry which is less sensitive to erosion and maintains the plugging-resistance flow area of the helical design. Patent pending design.

Figure 3: Hybrid ICD shown incorporated into the screen design (translucent housing over the ICD). The ICD is located between the screen jacket and inflow ports into the liner.

OTC 19811

9

Figure 4: CFD velocity magnitude contour output of hybrid ICD with water flow at 187.4 bpd flow rate.

Figure 5: CFD velocity magnitude contour output of hybrid ICD with heavy oil flow (88cp) at 187.4 bpd flow rate.

Figure 6: CFD flow line velocity contour output of hybrid ICD with water flow at 187.4 bpd flow rate.

Figure 7: CFD flow line velocity contour output of hybrid ICD with heavy oil flow (88cp) at 187.4 bpd flow rate.

Figure 8: Generic flowing static pressure contour through the hybrid ICD showing distributive pressure loss similar to the frictional ICD.

Figure 9: Static pressure graph showing pressure loss occurs at the slots with negligible frictional loss between the slots (between bulkheads).

10

OTC 19811

Figure 10: Test setup used to quantify performance characteristics of the new hybrid ICD design.

Figure 11: Test fixture used to quantify performance characteristics of the new hybrid ICD design. Multiple pressure transducers located between each bulkhead in the ICD perform transient data collection and analysis.

OTC 19811

11

Pressure Change Sensitivity

Pressure Drop Comparisons f/ 1.6FRR Hybrid ICD @ 5.5 GPM

relative to Water - 0.8FRR @ 5.5 GPM 50

100%

45

Actual dP Calc dP

80%

60%

0.866 g/cc

35 0.997 g/cc

detlaP (PSI)

% deltaP Change (relative to water)

40

40%

20%

30

0.814 g/cc

0.856 g/cc 0.835 g/cc

0.858 g/cc

25

20

15 0% 0

20

40

60

80

100

-20%

120

140

160

180

200

%change Orifice Plate %change Orifice Plugs %change Hybrid ICD

-40%

Fluid Viscosity (cP)

Figure 12: Chart depicting viscosity sensitivity for hybrid ICD and orifices. Shows the hybrid ICD to be the most insensitive to viscosity variations.

10

5

0 0

20

40

60

80

100

120

140

160

180

200

Fluid Viscosity (cP)

Figure 13: Chart comparing calculated pressure drop versus measured pressure drop for the hybrid ICD as a function of viscosity.

Figure 14: Erosion test fixture setup used to quantify any change in pressure loss attributed to erosion from SiC particles in the flow stream.

12

OTC 19811

Pressure Profile Comparisons 6-5/8" Hybrid ICD 1.6FRR Fluid Medium: Water, Test Flow Rate: 5.5 GPM

90

deltaP Pretest (PSI) deltaP 400lbs (PSI) deltaP 500lbs (PSI)

80 70

deltaP (PSI)

60 50 40 30 20 10 0 0

1

2

3

4

5

6

7

8

9

10

Flow Rate (GPM)

Figure 15: Pressure profile comparisons with water pre-test, after 400-lb and after 500-lb of exposure to SiC. Performance of the hybrid ICD remains unchanged.

Figure 16: Full view of the hybrid ICD after completion of erosion testing.

Figure 17: View of the inlet side of a flow slot (post-test). No visible geometry changes to flow slot or opposing bulkhead face.

Figure 18: View of the discharge side of a flow slot (post-test). No visible geometry changes. Some surface finish variance but machine marks still visible from manufacturing.

OTC 19811

13

80

80

Helical Channel

Axial Orifice

Pressure Drop psi

Pressure Drop psi

HL

60

HL

60

40

20

0

40

Black – Water Brown – 4 cP Green – 18 cP Red – 53 cP Blue – 82 cP Purple 200 cP

20

0

2

4

HL

6

8

0 0

10

2

6

8

10

8

10

Flow Rate gpm

80

80

Radial Orifice

Hybrid Design

Pressure Drop psi

HL

60

HL

60 Pressure Drop psi

HL

4

Flow Rate gpm

40

20

0

40

20

0

2

4

HL

6

8

0 0

10

2

HL

4

Flow Rate gpm

6

Flow Rate gpm

Figure 19: Comparison pressure loss data through various ICD types and designs with varying flow rates and fluid viscosities.

0.16 80

80

0.14

Orifice Axial (psi) Equalizer Helical Channel (psi)

70

Orifice Radial (psi) Orifice Axial (ft/sec)

0.08 0.06 0.04

60

Equalizer Helical Channel (ft/sec)

0.1

Orifice Radial (ft/sec)

50

50

Hybrid Design Hybrid Design

40

40

30

30

20

20

10

10

Flow Velocity (ft/sec)

60 Pressure Drop (psi)

Min. Flow Area (in2)

70

0.12

0.02 0 Hybrid Design

Equalizer Helical Channel

Orifice Axial Flow

Orifice Radial Flow

ICD Geometries

Figure 20: Comparison of minimum flow areas through the various ICD designs (sized to generate the same pressure loss).

0

0 0

1

2

3

4

5

6

7

8

9

10

Flow Rate (gpm)

Figure 21: Pressure drop and flow velocities in the various ICD designs (water flow).

14

OTC 19811

500

Pressure Loss Coeficcient

Helical Channel Design

100 50 Hybrid Design

10

50 ft

5

1 10

Orifice Radial Design

Orifice Axial Design

100

8000 ft

104

1000

105

Reynold Number

3200 ft

Figure 22: Flow performance characteristics of the various ICD designs.

Figure 23: Reservoir control volume, horizontal well types and aquifer.

Hybrid

Hybrid Design

Hybrid Orifice Radial

Orifice Radial Design

Orifice Radial

Hybrid

Hybrid Orifice Radial

Figure 24: Water flow in the reservoir comparing orifice and hybrid designs.

Orifice Radial

Figure 25: Fluid movement along horizontal section.

18 1 cP

Cum u lative Produ ction

4 cP

20 cP

80 cP

16 Oil - STB

Cummulative Production (MMSTB)

Oil - MMSTB W ater - M MSTB Standard S creen 3.22 14 .3 Orifice Radial 3.95 13 .6 Hybrid 4.20 13 .1

14

W ater - STB

12 10 8 6 4 2

PICD's type & Standard Screen

Figure 26: Fluid production forecast model output.

Figure 27: Cumulative fluid production as a function of fluid viscosity and ICD geometry.

Hybrid

Orifice Radial

Standard Screen

Hybrid

Orifice Radial

Standard Screen

Hybrid

Orifice Radial

Standard Screen

Hybrid

Orifice Radial

Standard Screen

0