This paper was prepared for presentation at the Unconventional Resources Technology Conference held in San Antonio, Texas, USA, 1-3 August 2016.

Size: px

Start display at page:

Download "This paper was prepared for presentation at the Unconventional Resources Technology Conference held in San Antonio, Texas, USA, 1-3 August 2016."

Antonia Stone
6 years ago
Views:

1 URTeC: Managed Pressure Flowback in Unconventional Reservoirs: A Permian Basin Case Study Darryl Tompkins*, Robert Sieker, and David Koseluk, Halliburton; Hugo Cartaya, GMT Exploration Copyright 2016, Unconventional Resources Technology Conference (URTeC) DOI urtec This paper was prepared for presentation at the Unconventional Resources Technology Conference held in San Antonio, Texas, USA, 1-3 August The URTeC Technical Program Committee accepted this presentation on the basis of information contained in an abstract submitted by the author(s). The contents of this paper have not been reviewed by URTeC and URTeC does not warrant the accuracy, reliability, or timeliness of any information herein. All information is the responsibility of, and, is subject to corrections by the author(s). Any person or entity that relies on any information obtained from this paper does so at their own risk. The information herein does not necessarily reflect any position of URTeC. Any reproduction, distribution, or storage of any part of this paper without the written consent of URTeC is prohibited. Abstract Flowback programs on multi-stage horizontal wells (MSHW) in unconventional reservoirs are too often completed without a sufficient understanding of how the flowback could affect the long-term performance of the well. High initial production is a common performance indicator that drives flowback practices to focus primarily on flowing back the well as fast as possible to get it on production. Operators are becoming increasingly aware of how extremely sensitive unconventional completions and formations can be to aggressive flowback practices used to increase initial production. To date, the collection and analyses of data gathered during flowbacks have been infrequent and with little understanding of the inherent value of the data. If data is collected, it is usually low quality and unreliable. This paper focuses on the specific type of damage that can be created by aggressive flowback practices and how this damage can be mitigated by effectively monitoring and controlling initial rates and pressures. A Permian Basin case study is examined to demonstrate the benefits of optimizing choke schedules; this process effectively manages drawdown pressure to reduce damage to fracture conductivity and to increase cumulative production. Specialized diagnostic plots are used to demonstrate the potential consequences of conventional flowback practices on well deliverability. In addition, the paper presents information about the added value of using high-resolution surface pressure recorders and describes how the data can be used to optimize choke schedules in real time. The managed pressure flowback strategies and workflow established with this case study have proven to increase the performance of wells initially anticipated to have marginal production within the operator s asset area in the Permian Basin. Introduction The awareness of drawdown management during initial flowback and production of unconventional wells is increasing. This increased awareness can be partially attributed to a shift in focus away from obtaining only the highest initial production and toward maximizing total asset value. Flowback is the practice of controlling the rate and pressure of fluid being produced from a well during the early production period after the well has been stimulated. Early case studies have provided details about how aggressive drawdowns can damage fracture conductivity and reduce well performance (Robinson et al. 1988). Most of this damage is concentrated near the wellbore where fluid velocities and pressure gradients are greatest. This damage can be reduced by carefully controlling bottomhole fluid velocities before closure and by limiting bottomhole flowing pressure to prevent damage to the formation and proppant. Conservative drawdown strategies in low permeability formations and the elimination of shut-ins during early time flowback have been shown to positively affect well performance (Crafton 2008).

$URTeC: 2461207 2 Workflows for analyzing flowback data have been proposed that evaluate the changes in apparent system fracture half-length, skin, and permeability caused by transients introduced$

2 URTeC: Workflows for analyzing flowback data have been proposed that evaluate the changes in apparent system fracture half-length, skin, and permeability caused by transients introduced into the reservoir by choke changes at the surface (Crafton 1998). A recent case study by Deen et al. (2015) outlined damage mechanisms from aggressive drawdowns that could be controlled through choke management and demonstrated the successful implementation of choke management workflows that incorporate a diagnostic analysis to reduce damage to wells and to increase production. Figure 1: Regional extent of Bone Springs sand formation and stratigraphy. The case study presented in this paper focuses on two wells, Well A and Well B, in southeastern New Mexico; these wells were completed in the Delaware basin, Second Bone Springs sand, and are approximately 4.5 miles apart (Fig. 1). This formation is an overpressured, hydrocarbon-rich, unconventional reservoir with cyclic sedimentation of sands and carbonates. Well A was flowed back using the operator s conventional approach; Well B was flowed back using a drawdown strategy in which well performance derived from a diagnostic analysis was used to make operational decisions for choke management. Table 1 lists the completions and reservoir parameters for Well A and Well B. In addition, the density porosity was greater in Well A than in Well B. The porosity averaged 8.3% in Well A and 7.4% in Well B. The sand packages were also more contiguous in the Well A area than in the Well B area. In Well B, several carbonate stringer blocks isolated the sand package. In Well A, there were no carbonate stringers to limit the fracture height, and there were other nearby sand packages that could contribute to greater production from Well A. Well A Well B Total vertical depth (ft) 10,513 10,145 Lateral length (ft) 3,919 3,823 Tubing string 2 7/8 to 9571 ft 2 7/8 to 9500 ft Production casing 5.5in. 20lb/ft 5.5 in.20lb/ft Number of stages Proppant lb/ft 1,542 1,572 Total load fluid (bbl) 99, ,235 Estimated pore pressure (psia) 6,000 6,000 API gravity Water saturation (%) Table 1: Completion and reservoir parameters.

3 URTeC: Methods/Theory The flowback program for Well B was designed to prevent damage from proppant and fines migration caused by excessive fluid rates during flowback. With this in mind, the initial rates on Well B were maintained at a lower level than Well A. The intent was to limit the mobilization of proppant in fractures and to maintain equal drawdown along the lateral. Wellbore flow models constructed for Well B indicated that the initial rates observed on Well A could potentially draw the majority of fluid from the heel section. The initial flow rates for Well B were reduced to enable proportional drawdown from toe to heel. This reduction also enabled hydrocarbons to break through more evenly throughout the fractured system and increase load fluid recovery. The diagnostic analysis outlined by Crafton (1998) was used for Well B to assess the effect of choke changes during flowback and to evaluate stimulation effectiveness. This analysis method incorporates the use the reciprocal productivity index (RPI), Argarwal-Gringarten type curve, and pseudo-steady state plot as the primary diagnostic tools for evaluating early time flowback data. The RPI plot is created by calculating the pseudo-potential corrected to bottomhole conditions divided by the total mass flow rate. The RPI allows for non-ideal fluid properties while accounting for the total fluid removed from the well. This plot enables the identification of changes in system connectivity (apparent system fracture half-length) and far-field permeability thickness. The merged Agarwal-Gringarten type curve is used to determine early time fracture or near-wellbore dominated behavior and the transitioning to late time pseudo-steady state depletion behavior. The use of multiple independent diagnostic plots will help to ensure interpretation consistency and uniqueness (Crafton 1997). Surface equipment selection and operations conducive to optimum reservoir performance were incorporated in the flowback design for Well B. This incorporated small 1/64 in. choke increases, positive chokes for stable production, adjustable chokes for bringing the well online, and the use of dual flow paths on critical equipment to eliminate shut-ins early in the flowback. It has been demonstrated that shut-ins early in the flowback can reduce well performance (Crafton 2008); consequently, every effort should be made to reduce and/or eliminate the need to shut the well in during post stimulation flowback. In addition, high-resolution, dual quartz surface pressure transducers were used for consistent, high-quality data acquisition. Examples Fig. 2 provides a summary of the flowbacks performed on Well A and Well B. Notable differences include the initial choke sizes, flow rates, and pressure declines. Well A was flowed back for 26 days, with an initial shut-in wellhead pressure (WHP) of 2,750 psi, average initial fluid rate of approximately 1600 BWPD (67 BWPH), and average initial pressure decline of 140 psi/day. The choke was increased from 12/64 to 15/64 in. early to maintain the rate and was held on a 15/64 in. until hydrocarbons broke through and water rates began to decline. After hydrocarbons broke through, the choke was brought up in several 2/64 in. increments to a final choke size of 23/64 in. and held there until the end of the flowback. Total fluid recovered was 35,264 bbl, with 16,710 bbl recovered before the beginning of oil production. The final rates were 838 MCFD, 648 BOPD, and 840 BWPD with 775 psi at the wellhead. The reservoir pressure was estimated to be 6,019 psia (0.57 psi/ft.) using the method described by Jones et al. (2014).

URTeC: 2461207 4 Figure 2: Well A and Well B flowback summary. Well B was flowed back for a total of 32 days with an initial shut-in WHP of 2,300 psi.

4 URTeC: Figure 2: Well A and Well B flowback summary. Well B was flowed back for a total of 32 days with an initial shut-in WHP of 2,300 psi. The beginning choke size was selected to maintain an initial water flow rate of 715 BWPD (30 BWPH) and an average initial pressure decline of 100 psi/day. The initial choke size of 11/64 in. was maintained for the first three days of flowback. The wellbore flow models for Well B indicated that flow rates of less than 50 BWPH would not create large pressure gradients along the lateral and in the fractures, which would enable a more balanced flow contribution and drawdown per stage. The choke schedule was designed to be conservative with single 1/64 in. choke changes every 24 hours. This schedule would enable sufficient data to be collected for evaluation using RPI without increasing the flow rate so quickly that a decrease in well performance could not be reversed. The total fluid recovered was 34,474 bbl, with 13,758 bbl recovered before the beginning of oil production. At the end of the 32 days of flowback, the well was flowing 792 MCFD, 696 BOPD, and 552 BWPD through a 34/64 in. choke with 325 psi at the wellhead. The reservoir pressure was estimated to be 5,326 psi (0.52 psi/ft.) using the same method as for Well A. Table 2 provides production ratios for Well A and Well B for comparison. Production Ratio Well A Well B GOR (MCFD/BOPD) GWR (MCFD/BWPD) OWR (BOPD/BWPD) Table 2: Production ratio comparison. High quality surface pressure transducers were used to collect one-second pressure data on Well B. This level of resolution illuminated the surface operating practices that have the potential to cause damage if not properly mitigated. During the flowback of Well B, some pressure cycling events were noticed, and appropriate action was taken to reduce and eliminate these practices.

URTeC: 2461207 5 Figure 3: Pressure cycling attributable to surface operations. Fig. 3 shows a comparison of the high-resolution, one-second pressure data in blue and the hourly pressure data in red on plot 1.

5 URTeC: Figure 3: Pressure cycling attributable to surface operations. Fig. 3 shows a comparison of the high-resolution, one-second pressure data in blue and the hourly pressure data in red on plot 1. The high-resolution data captured pressure cycling events at the surface that were not visible in the hourly data recorded onsite. Upon further investigation, the pressure cycling was found to be the result of a common industry practice of rocking the choke, in which the adjustable choke is turned in and out to clear debris. This process causes a surging effect that can reduce well performance. After this effect was discovered, every effort was made to eliminate the need for this practice. This awareness of potentially harmful surface practices showed added value in using high quality surface pressure transducers with one-second sampling. In addition, it helped in planning future equipment configurations and oriented crew training toward operating the equipment to maintain completion effectiveness. Results Using the workflow outlined by Crafton (1998), the log-log Argarwal-Gringarten type curve is first examined. The two curves are merged to enable the evaluation of early time, near-wellbore behavior and late time, pseudo-steady state behavior on a single plot. The Argarwal portion represents dimensionless fracture conductivity (F CD ) and the Gringarten portion represents the ratio of the bounded reservoir length to fracture length. The red line indicates the computed dimensionless pressure vs. dimensionless time of the data collected during the flowback. During the early production life of the well throughout the flowback, the data should initially be analyzed on the Argarwal-Gringarten type curve for any half slopes or unit slopes that would indicate linear flow or boundarydominated behavior.

6 URTeC: Figure 4: Well B Agarwal-Gringarten type curve. Fig. 4 shows the data plotted on the Agarwal-Gringarten type curve for Well B. The very early portion of the data that does not completely fall on the curve indicates cleanup; the remaining data falls on the infinite conductivity type curve with no indication of linear flow or boundary effects. Next, the modified MDH plot is analyzed. The graph shown in Fig. 5 illustrates the bottomhole RPI values vs. the log of time plotted in red for Well B. The blue line is the model match of the data. This plot is used to determine the middle time reservoir-dominated portion of the well performance. The slope of the linear portion of the data is used to calculate the total far-field permeability thickness (kh). The y-axis intercept is used to calculate the apparent skin or, in the case of a fractured well, the apparent system fracture half-length. The apparent system fracture half-length determined from this analysis is not any one specific fracture half-length or an average of all the fracture halflengths; it is used here as a qualitative metric for assessing relative completion effectiveness among similar wells completed in the same region and formation. Figure 5: Well B modified MDH plot for RPI analysis. Well B showed a relatively consistent linear trend throughout the flowback. Fig. 6 shows the comparison of the RPI plot for Well A (blue line) and Well B (red line). The permeability thicknesses for Well A and Well B were determined to be 15.9 md-ft and 13.1 md-ft, respectively. The final apparent system fracture half-lengths of Well A and Well B were 279 ft and 761 ft, respectively.

URTeC: 2461207 7 The data sets for both wells compare very well early on, and the slope of the RPI is similar for each, which indicates a similar performance.

7 URTeC: The data sets for both wells compare very well early on, and the slope of the RPI is similar for each, which indicates a similar performance. However, in each data set, there is an increase in the slope of the trend line, indicating the onset of pseudo-steady state or a decrease in connectivity caused by a reduction in the apparent system fracture halflength. Because the onset of pseudo-steady state during the flowback is unlikely (though not impossible), it is reasonable to conclude that a decrease in apparent system fracture half-length is responsible for the decrease in well performance. Given this, further interpretation was performed to identify possible causes of the decrease in apparent system fracture half-length. Figure 6: RPI comparison of Well A and Well B. By analyzing most significant transients created during the flowback, the change in permeability thickness and apparent system fracture half-length was observed to change with time. Fig. 7 shows the most significant transients created near the time the slope began to increase on the RPI plot for Well B; this indicates that the apparent system fracture half-length was decreasing in magnitude, but the far-field permeability thickness remained constant. Figure 7: RPI analysis of individual transients.

URTeC: 2461207 8 Figure 8: Event causing lost performance on Well B. This result led to the conclusion that some loss in well performance occurred in the near-wellbore region.

8 points to the time directly before the slope increase on the RPI plot of Well B. This time corresponded to a short, three-hour shut-in to repair a leak in the lease equipment. The red arrow in Fig.

8 URTeC: Figure 8: Event causing lost performance on Well B. This result led to the conclusion that some loss in well performance occurred in the near-wellbore region. Additional investigation was required to determine the cause of this event and what could be done to eliminate it in the future. The red arrow in Fig. 8 points to the time directly before the slope increase on the RPI plot of Well B. This time corresponded to a short, three-hour shut-in to repair a leak in the lease equipment. The red arrow in Fig. 9 points to the time directly before the slope of the RPI plot increases on Well A. This point corresponds to the third of 2/64 in. increment choke changes, which were followed by flowing the well through the production facilities. A sharp increase in rates occurred at this point, followed by a sharp decrease in pressure. Although the well responded satisfactorily to the first two 2/64 in. choke increases, the third increase caused the reduction in performance. Figure 9: Event causing lost performance on Well A. These examples are provided demonstrate how the diagnostic analysis can be used to identify the sensitivity of unconventional well performance to surface operations. Although there was some loss in performance on Well A and Well B after nearly 200 days on production, Fig. 10 shows that Well B was producing more cumulative BOE than Well A.

9 URTeC: Figure 10: Well A and Well B cumulative BOE comparison. Conclusions The diagnostic analysis that incorporates the Argarwal-Gringarten type curve and RPI plot was used to analyze the flowback data of two wells with similar completion and reservoir parameters. This analysis helped to identify damage caused by surface operations and to guide decisions regarding choke management. In addition, although the analysis of Well A indicated it had a slightly greater permeability thickness, the total apparent system fracture halflength was considerably smaller. This result was attributed to a single 2/64 in. choke change. The initial production of Well B was greater than anticipated. The diagnostic analysis proved to be a valuable tool for evaluating post stimulation completion effectiveness. In addition, by using the diagnostic analysis during the flowback of Well B, operational decisions regarding choke changes were determined by the well performance. This analysis, combined with near real-time collaboration with the field well test personnel, enabled the identification/mitigation of damage on Well B and, ultimately, an increase in cumulative production. Acknowledgements Darryl Tompkins would like to acknowledge Dr. Jim Crafton of Performance Science Inc. for his help and support throughout this project. References Crafton, J.W Oil and Gas Well Evaluation Using the Reciprocal Productivity Index Method. Presented at SPE Production Operations Symposium Oklahoma City, Oklahoma, US March. Paper SPE Crafton, J.W Well Evaluation using Early Time Post Stimulation Flowback Data. Presented at SPE Annual Technology Conference and Exhibition, New Orleans, Louisiana, USA September. Paper SPE Crafton, J.W Modelling Flowback Behavior or Flowback Equals Slowback. Presented at SPE Shale Gas Production Conference, Fort Worth, Texas, USA November. Paper SPE Crafton, J.W Flowback Performance in Intensely Fracture Shale Gas Reservoirs. Presented at SPE Unconventional Gas Conference, Pittsburgh, Pennsylvania, USA February. Paper SPE

10 URTeC: Crafton, J.W. and Noe, S.L Factors Effecting Early Well Productivity in Six Shale Plays. Presented at SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana, USA. 30 September 30 2 October. Paper SPE Daneshy, A Proppant Distribution and Flowback in Off-Balance Hydraulic Fractures. Presented at SPE Annual Technical Conference and Exhibition, Houston, Texas, USA September. Paper SPE Deen, T., Daal, J., and Tucker, J Maximizing Well Deliverability in the Eagle Ford Shale Through Flowback Operations. Presented at SPE Annual Technology and Exhibition, Houston, Texas, USA September. Paper SPE Jones, S., Pownall, B., and Franke, J Estimating Reservoir Pressure from Early Flowback Data. Presented at the Unconventional Resource Technology Conference, Denver, Colorado, USA August Paper URTec Robinson, B.M., Holditch, S.A., and Whitehead, W.S Minimizing Damage in Propped Fracture by Controlled Flowback Procedures. Presented at SPE Unconventional Gas Technology Symposium, Louisville, Kentucky, USA May. Paper SPE

Successful Completion Optimization of the Eagle Ford Shale

Successful Completion Optimization of the Eagle Ford Shale Supplement to SPE Paper 170764 Presented at the Society of Petroleum Engineers Annual Technical Conference and Exhibition Amsterdam, The Netherlands