Recent Work in Well Performance Analysis for Tight Gas Sands and Gas Shales

Transcription

1 06 November 2009 Houston, TX Recent Work in Well Performance Analysis for Tight Gas Sands and Gas Shales Tom BLASINGAME Department of Petroleum Engineering Texas A&M University College Station, TX (USA) Slide 1/63

2 SPE Evaluating the Impact of Waterfrac Technologies on Gas Recovery Efficiency: Case Studies Using Elliptical Flow Production Data Analysis D. Ilk,* Texas A&M U. J.A. Rushing, Anadarko Petroleum Corp. R.B. Sullivan, Anadarko Petroleum Corp. and T.A. Blasingame, Texas A&M U. *Department of Petroleum Engineering Texas A&M University College Station, TX Slide 2/63

3 Elliptical Flow Type Curves: [Amini et al (2007)] Type curve for a fractured well centered in a closed (homogeneous) elliptical reservoir F E =1, various ξ 0 -values; q D functions versus t DA format. Type curve for a fractured well centered in a closed (homogeneous) elliptical reservoir F E =10, various ξ 0 -values; q D functions versus t DA format. Discussion: Elliptical Flow Type Curves Equivalent constant rate type curves shown in q D versus t DA format. Elliptical fracture conductivity: F E = 1, 10, 100, Elliptical geometry parameter: ξ 0 = 0.25, 0.50, 0.75, 1, 1.50, 1.75, 2, 3, 4, 5. Slide 3/63

4 Elliptical Flow Type Curves: [Amini et al (2007)] Properties that can be estimated using production data analysis based on elliptical flow: Effective gas permeability, k, md Effective fracture half-length, x f, ft Effective fracture conductivity, F E Contacted gas-in-place, G, BSCF Effective drainage area, A, acres The characteristic elliptical boundary parameter (ξ 0 ) correlates all of the aspects of the drainage area and the fracture half-length: 1 sinh(2ξ Discussion: Elliptical Flow Type Curves Objective: To demonstrate the elliptical boundary model as a diagnostic to establish the elliptical (transition) flow regime using production data. We utilize type curve solutions in terms of the equivalent constant rate case presented in "decline" form (q D and t DA ). x f = 2A π 0 ) 0.5 Slide 4/63

5 Production Analysis: Overview Procedure: Step 1: Data Overview: "Production History Plot" General data quality/correlation. Step 2: Data Correlation: "p wf (or p tf ) vs. Rate Plot" Crude comparison used only for general trends. Step 3: Clean/Edit Data for Clarity: Log-Log Data Plot(s) Remove spurious data from base data trends (e.g., q/δp). Step 4: Identify Flow Regimes (Diagnostics): Log-Log Data Plot(s) Characteristic flow regimes from normalized PI-style plots. Step 5: Compare Data to Reservoir Model: "Type Curve" matching (data onto a reservoir model). Step 6: Refine Match/Model: Improve match of model parameters (k, s, x f, F cd,...) using individual type curves, simulation models, and/or regression methods. Step 7: Summary History Match: Final "history match" of model and raw data (p wf and q). Slide 5/63

6 Case Studies: Small Waterfrac 20/40 Proppant Production history plot (daily flowrate and pressure measurements). Diagnostic log-log plot (dimensionless rate decline and rate decline integral functions). Discussion: Diagnosis of the Well Performance Data A consistent character in the rate and pressure functions is observed q g and p wf data correlation appears to be good to very good. q g /Δp p diagnostic functions appear to be relevant earliest third of data appear off-trend (weak q g and p wf correlation at early time?). Treatment 7745 bbls of slick water; 63,000 lbs 20/40 proppant. Slide 6/63

7 Case Studies: Small Waterfrac 20/40 Proppant Elliptical boundary decline type curve match (moderate conductivity, near-circular drainage geometry). Production history plot with model match (excellent flowrate match, good pressure match). Discussion: Analysis Results Discrepancy at early times is observed probably an early-time artifact (shape of integral-derivative), possibly some modest fracture damage. Excellent match of the model and the rate history, good match of the pressure history "balance" of analysis between the diagnostic functions and the production history. k= md, x f =129 ft, F E =10, ξ o =2.0 (near-circular), G=1.28 BSCF. Slide 7/63

8 Case Studies: Large Waterfrac 20/40 Proppant Production history plot (daily flowrate and pressure measurements). Diagnostic log-log plot (dimensionless rate decline and rate decline integral functions). Discussion: Diagnosis of the Well Performance Data Erratic pressure and rate data observed from days (we were told a workover was performed to remedy a "tubing or packer leak"). Significant well clean-up effects in the pressure data during the first 2 months of production (poor correlation?). q g /Δp p diagnostic functions indicate a strong transient flow signature ( 1/2 high fracture conductivity), BDF is also observed. Treatment 7300 bbls of slick water; 247,500 lbs 20/40 proppant. Slide 8/63

9 Case Studies: Large Waterfrac 20/40 Proppant Elliptical boundary decline type curve match (very high conductivity, "fat" elliptical drainage geometry). Production history plot with model match (excellent flowrate match, fair pressure match). Discussion: Analysis Results Excellent type curve match obtained using an elliptical flow model with a high fracture conductivity (all flow regimes matched!). Outstanding match of the model and the rate history (q g ), good match of the pressure history (p wf ) surface pressure data is not representative of the bottomhole condition (up to 600 hr) due to "tubing/packer" leak. k= md, x f =184 ft, F E =1000, ξ o =1.50 ("fat" ellipse), G=3.31 BSCF. Slide 9/63

10 Case Studies: Hybrid Waterfracs Production history plot (daily flowrate and pressure measurements). Diagnostic log-log plot (dimensionless rate decline and rate decline integral functions). Discussion: Diagnosis of the Well Performance Data Erratic nature of the flowrate profile is seen (liquid loading). Major features of the liquid loading observed for the flowrate data are reasonably well-correlated with the calculated bottomhole pressures. Very clear diagnostic trends are obtained after judicious editing of the q g /Δp p function (1/2 slope during transient flow, 1/1 slope for BDF). Treatment 2082 bbls of slick water bbls of cross-link gel; 510,140 lbs 20/40 proppant. Slide 10/63

11 Case Studies: Hybrid Waterfracs Elliptical boundary decline type curve match (very high conductivity, "thin" elliptical drainage geometry). Production history plot with model match (very good flowrate match, acceptable pressure match). Discussion: Analysis Results Extraordinary "type curve match" is observed for this case the well is effectively stimulated and a low reservoir permeability is expected. Model flowrate response corresponds well with the observed flowrate data. The pressure match is not as poor as it might seem given the very erratic nature of the flowrate function. k= md, x f =200 ft, F E =1000, ξ o =1.00 ("thin" ellipse), G=1.60 BSCF. Slide 11/63

12 Case Studies: Discussion of Results Results correlation plot G versus k. Results correlation plot G versus x f. Discussion: Effect of the Contacted Gas-In-Place Expectation higher reservoir permeabilities, higher contacted gasin-place estimates. Most small waterfrac cases are off-trend (a single on-trend point with high permeability and high contacted gas-in-place estimate is noted). Excellent correlation of contacted gas-in-place with fracture halflength confirming that "the fracture defines the reserves." Slide 12/63

13 Case Studies: Discussion of Results Definitions: x f 0.5 2A 1 = π sinh(2ξ0 ) Or ξ 0 = A sinh π 2 x f Results correlation plot k versus ξ o. Results correlation plot x f versus ξ o. Discussion: Effect of the Drainage Aspect Ratio Clear correlation of increasing permeability with increasing drainage aspect ratio (ξ o ). Small waterfracs exhibit the highest values of permeability? Function of development strategy early wells targeted to best reservoir? Inverse correlation between x f and ξ o intuitive (see definition). Slide 13/63

14 SPE Exponential vs. Hyperbolic Decline in Tight Gas Sands Understanding the Origin and Implications for Reserve Estimates Using Arps' Decline Curves D. Ilk, Texas A&M University A.D. Perego, Anadarko Petroleum Corp. J.A. Rushing, Anadarko Petroleum Corp. T.A. Blasingame, Texas A&M University Department of Petroleum Engineering Texas A&M University College Station, TX Slide 14/63

15 Rationale For This Work ASSUMPTION: The Arps decline parameter, b, defines the decline behavior when tight gas sand reserves are assessed. REALITY: Difficult to identify the correct b-parameter during the early decline period selection of the wrong b- parameter greatly impacts reserve estimates. a. (Semilog plot) Production forecast of a tight gas well. b. (Log-log plot) Production forecast of a tight gas well. Slide 15/63

16 Overview: Loss Ratio (Definition and Behavior) Loss Ratio: (basis for exponential rate decline) 1 qg q D dq / dt g exp[ D t] Loss Ratio Derivative: (basis for hyperbolic rate decline) q q d 1 d g gi b qg = dt D dt dqg / dt (1+ bd ) it g = q gi i (1/ b) [From: Johnson, R.H. and Bollens, A.L.: "The Loss Ratio Method of Extrapolating Oil Well Decline Curves," Trans., AIME (1927) 77, 771.] Slide 16/63

17 Overview: Arps' Rate Decline Functions Case Rate-Time Relation Cumulative-Time Relation Exponential:(b=0) qg = qgiexp[ Dit] Hyperbolic: (0<b<1) qgi qg = (1/ b) (1+ bdit) qgi Harmonic: (b=1) qg = (1+ Dit) Gp qgi = Di [1 exp[ Dit]] qgi G [1 (1 ) 1 (1/ b) p = + bdit ] (1 b) Di qgi Gp = ln(1+ Dit) Di [From: Arps, J.J: "Analysis of Decline Curves," Trans., AIME (1945) 160, ] Slide 17/63

18 New Rate Equation: Ilk et al [2008] Observed Behavior of Decline Parameter (D): D dq 1 ˆ (1 n) D + nd i t q dt Solving for Flowrate: q = qˆ exp[ D i t Dˆ i t n ] Solving for the b-parameter: b = [ ndˆ i ndˆ + i D (1 t n) (1 n) ] 2 t n Slide 18/63

19 Field Examples: Small WF Gas Well (SWF2) "q-d-b" Plot: SWF2 D-parameter data trend exhibits a power-law behavior essentially a straight line. b-parameter data trend is not constant (contrary to hyperbolic formulation). Computation of the b- parameter is severely affected by noise. Discussion: Small Waterfrac Gas Well (SWF2) Liquid loading effects are obvious in the latter portion of the flowrate data. The onset of the boundary-dominated flow regime is observed. D is set to 0 initially, then tuned to the latest data we obtain a very good match of the D-parameter data trend with the power-law models. We observe a very good match of the flowrate data with the "base" powerlaw exponential model (i.e., D =0). Slide 19/63

20 Field Examples: Small WF Gas Well (SWF2) a. Semi-log plot empirical matches are shown using power-law exponential and hyperbolic models. b. Log-log plot empirical matches are shown using power-law exponential and hyperbolic models. Discussion: Small Waterfrac Gas Well (SWF2) The hyperbolic rate relation (b=1) yields the highest reserves estimate. Excellent matches of data are achieved using the power-law exponential model for both the D =0 and the D 0 cases. The lower bound for the reserves estimate is 2.3 BSCF, which is consistent with our results from the model-based PA study (SPE ). Slide 20/63

21 Field Examples: Large WF Gas Well (LWF2) "q-d-b" Plot: LWF2 D-parameter data trend exhibits a power-law behavior essentially a straight line. b-parameter data trend is not constant (contrary to hyperbolic formulation). Computation of the b- parameter is significantly affected by noise. Discussion: Large Waterfrac Gas Well (LWF2) Erratic rate behavior caused by liquid loading is seen in the latter portion of the rate data. The behavior of the computed D- and b-parameters is almost identical to the previous case suggesting (to some degree) consistency of the data. Outstanding matches of the computed D- and b-parameters with the power-law exponential model are observed. Slide 21/63

22 Field Examples: Large WF Gas Well (LWF2) a. Semi-log plot empirical matches are shown using power-law exponential and hyperbolic models. b. Log-log plot empirical matches are shown using power-law exponential and hyperbolic models. Discussion: Large Waterfrac Gas Well (LWF2) Estimation of the reserves using the hyperbolic rate decline relation is almost five times higher than the contacted gas-in-place predicted previously using a model-based match (SPE ). Outstanding matches of the data are obtained using the power-law exponential model, both the D =0 and the D 0 cases. The D 0 model provides the most conservative estimate of reserves. Slide 22/63

23 Field Examples: Hybrid WF Gas Well (HWF1) "q-d-b" Plot: HWF1 D-parameter data trend exhibits a power-law behavior essentially a straight line. b-parameter data trend is not constant (contrary to hyperbolic formulation). Computation of the b- parameter is severely affected by noise. Discussion: Hybrid Waterfrac Gas Well (HWF1) Severe liquid loading is observed at late times. The computed D- and b-parameters reflect the effects of liquid loading however; the D-parameter data trend is essentially power-law. Good matches of the computed D- and b-parameters are obtained using the power-law exponential model. Slide 23/63

24 Field Examples: Hybrid WF Gas Well (HWF1) a. Semi-log plot empirical matches are shown using power-law exponential and hyperbolic models. b. Log-log plot empirical matches are shown using power-law exponential and hyperbolic models. Discussion: Hybrid Waterfrac Gas Well (HWF1) The hyperbolic rate relation (b=1) yields the highest reserve estimate. Reasonable matches of the rate data are obtained using the power-law exponential model for both the D =0 and the D 0 cases. The power-law exponential model applied using D 0 provides the most conservative estimate of reserves (as is expected) this result is quite comparable to the model-based results obtained in SPE Slide 24/63

25 SPE Decline Curve Analysis for HP/HT Gas Wells: Theory and Applications D. Ilk, Texas A&M University J.A. Rushing, Anadarko Petroleum Corp. T.A. Blasingame, Texas A&M University Department of Petroleum Engineering Texas A&M University College Station, TX Slide 25/63

26 Orientation: Rate-Time Relation From: Knowles R.S Development and Verification of New Semi-Analytical Methods for the Analysis and Prediction of Gas Well Performance. M.S Thesis, Texas A&M University, College Station, Texas. Ansah, J., Knowles, R.S., and Blasingame, T.A A Semi-Analytic (p/z) Rate- Time Relation for the Analysis andprediction of Gas Well Performance. SPEREE. 3 (6): Rate-Time Relation: Discussion: Rate-Time Gas Flow Relation (Knowles et al) Basis is the linearization of the nonlinear "μ g c g " term (Ansah, et al). D-function and b-function are formulated using the definitions for lossratio and the derivative of the loss-ratio. q D Dd D D D = ((1 + p 1 = q = Dd wd wd dq dt 4 p 2 wd ) (1 p Dd Dd exp[ p ) exp[ p wd Dimensionless D-function (D D ): pwd (1 pwd + (1 + ( p 1+ (1 + p wd wd t Dd ] wd pwd ) exp[ p ) exp[ p wd t wd t Dd t Dd ]) Dd ]) 2 ["Loss-Ratio"] b-function (b): ["Derivative of Loss-Ratio"] b = d dt Dd wd qdd ( dqdd / dtdd ) 2exp[ pwd t b = (1 p + (1 + p Dd wd ] (1 p )exp[ p 2 wd wd ) t Dd ]) 2 ]) Slide 26/63

27 Orientation: Rate-Cumulative Production Relation Rate-Cumulative Production Relation: qdd α 2 = 1 α GpD + G 2 pd α = 2 / (1 2 p wd ) Dimensionless D-function (D D ): D D dq = dg Dd pd DD = α ( 1 GpD) ["Loss-Ratio"] From: Knowles R.S Development and Verification of New Semi-Analytical Methods for the Analysis and Prediction of Gas Well Performance. M.S Thesis, Texas A&M University, College Station, Texas. Ansah, J., Knowles, R.S., and Blasingame, T.A A Semi-Analytic (p/z) Rate- Time Relation for the Analysis andprediction of Gas Well Performance. SPEREE. 3 (6): b-function (b): ["Derivative of Loss-Ratio"] b = qdd d 1 dg pd ( dqdd / dg pd ) 2 2α G 2 pd + α G pd b = 2α ( G 1) 2 pd Discussion: Rate-Cumulative Gas Flow Relation The definition of the loss-ratio can be re-cast in terms of rate and cumulative production. A quadratic relationship exists between rate and cumulative production. Slide 27/63

28 Orientation: Analysis Methodology Discussion: Methodology The main goal is to match the data with the model using the definitions for the q-d-b functions during the boundary-dominated flow regime. b-function 0.5 for high drawdown cases (almost constant behavior). Slide 28/63

29 Field Example: HP/HT Tight Gas Well p i = psia and T R = 260 o F Field Example: Application of the Methodology 3.5 years of daily data are available for a hydraulically fractured well completed in a HP/HT gas reservoir. Well clean-up effects, liquid-loading, and operational changes are observed in the data trends. The flowrate data are reviewed prior to analysis; and any erroneous/ redundant data points are removed. The half-slope trend is evident in the rate-integral derivative function. Slide 29/63

30 Field Example: HP/HT Tight Gas Well a. q g versus G p (Cartesian plot). b. D-function versus t (Cartesian plot). c. b-function versus t (Cartesian plot). d. q g versus t (Semilog plot). e. D-function versus t (Semilog plot). f. b-function versus t (Semilog plot). Field Example: Application of the Methodology For the computation of D- and b-parameter data functions we remove the outlying data points; then we perform the numerical differentiation. Our analysis provides a gas-in-place estimate of approximately 8.0 BSCF. Slide 30/63

31 Field Example: HP/HT Tight Gas Well Field Example: Application of the Methodology Reasonable matches of the D-function with the data using the semianalytical model is achieved (post-transient flow only). The matches of the b-function data with the semi-analytical model are problematic data indicate no unique characteristic behavior. Computation of the b-parameter data function is severely affected by factors such as liquid loading. Slide 31/63

32 Field Example: HP/HT Tight Gas Well Field Example: Application of the Methodology We observe a good match of the flowrate data with the model (except for the early time data affected by "cleanup"). The "power-law exponential" model yields G p,max 8.0 BSCF. Gas-in-place estimates are consistent comparing the methods we used. Slide 32/63

33 Field Example: HP/HT Tight Gas Well Field Example: Application of the Methodology An analytical reservoir model (vertical well with a finite conductivity vertical fracture) is used to confirm the analysis results. Excellent match of the flowrate data and a reasonable match of the bottomhole pressure history are obtained. All of the models yield consistent estimates of reserves/gas-in-place. Slide 33/63

34 Summary and Conclusions: Summary: We utilize a semi-analytical rate relation given by Knowles [1999] and generalized by Ansah [2000] for the direct estimation of gas-in-place. The semi-analytical relation is formulated in terms of the Arps' D- and b-functions for diagnosis and analysis. "Power-law exponential" rate decline relation is used to augment and validate our other analyses. Conclusions: For high drawdown cases, the value of the Arps' b-parameter should converge to approximately 0.5 during boundarydominated flow regime. The straight line linearization scheme is validated including application to cases of HP/HT gas reservoirs. The use of the D- and b-data functions provides a unique insight into flow regime identification and the simultaneous matching of the data. We conclude that this approach is a robust mechanism for estimating gas reserves. Model-based analysis (if applicable) should be performed to validate the results of the semi-analytical/empirical methods addressed in this work. Slide 34/63

35 Reserves Estimation in Tight Gas/Shale Gas Reservoirs using the Continuous EUR Concept (manuscript in preparation for presentation at the 2010 SPE Unconventional Gas Conference) S.M. Currie, Texas A&M University D. Symmons, Consultant D. Ilk, Texas A&M University T.A. Blasingame, Texas A&M University Department of Petroleum Engineering Texas A&M University College Station, TX Slide 35/63

36 Continuous EUR: Simulated Case Discussion: Continuous EUR using Arps' Hyperbolic Decline Subsets of data are matched using the Arps' hyperbolic decline relation. EUR of each subset is estimated (q gi is fixed). Slide 36/63

37 Continuous EUR: Simulated Case Discussion: Continuous EUR using Power-Law Exponential Model EUR of each subset of data is estimated progressively. We do not fix any model parameter for the power-law exponential model. EUR stabilizes when boundary-dominated flow regime is established. Slide 37/63

38 Continuous EUR: Simulated Case Discussion: Continuous EUR using Arps' Hyperbolic Decline Hyperbolic b-parameter value decreases with time. EUR estimate decreases over time and stabilizes at late times. EUR estimate using the power-law exponential model is more conservative at early times. Slide 38/63

39 Continuous EUR: Field Case Discussion: Continuous EUR using Arps' Hyperbolic Decline Almost 2000 days of daily production data are available. b-parameter value ranges between 0.99 (all data) and 2.2 (earliest portion 50 days) for the subsets of data. Slide 39/63

40 Continuous EUR: Field Case Discussion: Continuous EUR using Power-Law Exponential Model All of the matches for each part of the data set are almost outstanding. Additional data constrain the EUR estimate. Boundary-dominated flow regime effects are being established at late times. Slide 40/63

41 Continuous EUR: Field Case Discussion: Continuous EUR using Arps' Hyperbolic Decline Almost 50 percent decrease is observed in the EUR estimate when Arps' hyperbolic relation is used. We advise caution when Arps' hyperbolic decline is used for estimating reserves in particular, at early times. Slide 41/63

42 SPE A Numerical Study of Tight Gas and Shale Gas Reservoir Systems C.M. Freeman, Texas A&M University G.J. Moridis, Lawrence Berkeley National Lab. D. Ilk, Texas A&M University T.A. Blasingame, Texas A&M University Department of Petroleum Engineering Texas A&M University College Station, TX matt.freeman@pe.tamu.edu Slide 42/63

43 Numerical Simulation Model: Concept a. Formation linear flow. b. Compound linear flow. c. Pseudo-elliptical flow. Discussion: van Kruysdijk and Dullaert [1989] Flow Regime Concept: Complex pressure profile behavior evolves due to fracture interference. Onset of compound linear period marked by rate decline. Slide 43/63

44 Numerical Simulation Model: Concept Discussion: van Kruysdijk and Dullaert [1989] Flow Regime Concept: Complex pressure profile behavior evolves due to fracture interference. Onset of compound linear period marked by rate decline. Slide 44/63

45 Numerical Simulation Model: Modeling Discussion: Linear (Planar) Fracture Repetitive Element Repetitive element used to represent flow of individual fractures. Fracture interference modeled by a no-flow boundary. Slide 45/63

46 Flowrate Behavior: Effects of Completion Discussion: Effects of Fracture Spacing Rate and auxiliary functions merge during compound linear flow. Smaller fracture spacing causes earlier fracture interference effect. Slide 46/63

47 Flowrate Behavior: Effects of Desorption Nonlinear Depletion: Sorption surfaces near fracture are significantly more depleted. Pressure/Density/Sorptive Storage is not intuitively related. x y Flowrates with Various Langmuir Volumes. Dimensionless Desorption Map Dimensionless Density Map Slide 47/63

48 Flowrate Behavior: Effects of Medium Discussion: Effect of Matrix Permeability Matrix permeability affects the early time rate behavior in particular. Convergence of rate profiles are observed at late times. Slide 48/63

49 Conclusions and Remaining Issues: Conclusions: For TGSG reservoir systems, many factors affect performance. Numerical modeling must be robust and tied to flow physics. Desorption effects originate near fractures. The well completion controls early time behavior. Complex fractures (any type) substantially enhance early rates. x y Flowrates with Various Langmuir Volumes. Dimensionless Desorption Map Dimensionless Density Map Slide 49/63

50 Conclusions and Remaining Issues: Issues for Well Performance in TGSG Systems: Correct assessment of the fracture distribution must be made to ensure proper flow regime identification and forecasting. Desorption is highly nonlinear, but it defies direct assessment. Assessment of fracture conductivity is non-unique. Well completion issues are critical particularly well clean-up. Slide 50/63

51 06 November 2009 Houston, TX Tight Gas/Shale Gas Well Performance Analysis: Future View Tom BLASINGAME Department of Petroleum Engineering Texas A&M University College Station, TX (USA) Slide 51/63

52 Decline Analysis: Tight Gas Systems SPE (2007) x X Pressure Monitoring Point No. 2 Estimating Reserves in Tight Gas Sands at HP/HT Reservoir Conditions: Use and Misuse of an Arps Decline Curve Methodology J.A. Rushing, A.D. Perego, R.B. Sullivan, Anadarko Petroleum, and T.A. Blasingame, Texas A&M U. y Wellbore Pressure Monitoring Point No. 1 X Hydraulic Fracture Numerical Model Considers: Reservoir Layering. k v /k h ratio. Fracture Length, x f. Fracture Conductivity, F cd. Analysis/Validation Approach: Fit q(t) with Arps' hyperbolic relation. Compare reserves to model at 30 years. Slide 52/63

53 Vertical TG/SG Wells: Elliptical Flow Domination SPE (2007) Evaluation of the Elliptical Flow Period for Hydraulically-Fractured Wells in Tight Gas Sands Theoretical Aspects and Practical Considerations S. Amini, D. Ilk, and T. A. Blasingame, SPE, Texas A&M U. a. Elliptical flow type curve solution low fracture conductivity case. b. Elliptical flow type curve solution high fracture conductivity case. c. Elliptical boundary configurations (finite conductivity fracture case [Amini, et al (2007)]. Slide 53/63

54 Vertical TG/SG Wells: Elliptical Flow Domination Results Generated Using: Ecrin Product Suite, Kappa Engineering, Sophia- Antipolis, France (2008). a. Pressure profile at 0 year (0 hr). d. Pressure profile at 9.26 years (81,200 hr). b. Pressure profile at 1 year (8768 hr). e. Pressure profile at years (161,700 hr). c. Pressure profile at 5.59 years (49,010 hr). f. Pressure profile at years (386,600 hr). Slide 54/63

55 Horizontal TG/SG Wells: Compound Linear Flow Presented at the 2nd European Conference on the Mathematics of Oil Recovery, Cambridge, England (1989). A Boundary Element Solution of the Transient Pressure Response of Multiply Fractured Horizontal Wells C.P.J.W. van Kruysdijk and G.M. Dullaert, Shell a. Rate performance behavior for a horizontal well with 4 transverse fractures infinite-acting reservoir (analog to van Kruysdijk and Dullaert work). Fine-scale numerical model. b. Specialized derivative plot (ref: van Kruysdijk and Dullaert) for a horizontal well with 4 transverse fractures infinite- and finite-acting reservoir cases. Fine-scale numerical model. c. Schematic diagram for the "compound linear flow" concept [van Kruysdijk and Dullaert (1989)]. Slide 55/63

56 SPE : Schematic Model for Simulation Base Simulation Model for Horizontal Well with Multiple Hydraulic Fractures Top View Horizontal Well Multiple Vertical Fractures Slide 56/63

57 SPE : q-d-b Plot Numerical Simulation Finite-Acting Reservoir Case Discussion: Horizontal Wells with Transverse Fractures Very high resolution simulation case. Very good D-parameter and good b-parameter computed from results. Excellent rate match using the new model (all regimes). Slide 57/63

58 SPE : q-d-b Plot Numerical Simulation Infinite-Acting Reservoir Case Discussion: Horizontal Wells with Transverse Fractures Very high resolution simulation case. Transient D- and b-parameters at late times (quasi-radial flow). Impossible to predict reserves. Slide 58/63

59 Horizontal TG/SG Wells: Diagnostics? Q1.Compound Linear Flow Domination? (transient flow) A1. Possibilities for estimating reservoir properties: a. Just give up impossible to resolve anything. (default) b. "Lump" k, x f, and L well into a "parameter." ("mechanistic model") c. Develop testing practices to estimate properties. (maybe ) d. Other model concepts (e.g., propagating ellipse). (very tedious) Q2.Estimating Reserves? A2. Issues: a. Extremely long transition to boundary-dominated flow. (reality) b. Hyperbolic rate relation will overestimate reserves. (as always) c. Power-law/exponential rate relation? (more validation) Q3.Role of simulation/modeling? A3. In the short-term, simulation/modeling is the primary tool at our disposal the analogy of using a hammer in place of a screwdriver comes to mind (sometimes effective, but always a sub-optimal solution particularly in the hands of children). Slide 59/63

60 06 November 2009 Houston, TX Blasingame Current Projects/Interests Tom BLASINGAME Department of Petroleum Engineering Texas A&M University College Station, TX (USA) Slide 60/63

61 Current Work: Blasingame Projects: Status Simple Rate-Time Models for Shale Gas Systems (active) Diagnostics and Simplified Production Analysis (active) Analytical/Numerical Models for Shale Gas (active) IPR for Solution Gas-Drive Systems (nearing closure) Concepts: Production Mechanisms for Shale Gas Systems Status (start-up) Petrophysical Properties of Shales (2010) Continuous EUR (start-up) Focus Simple Rate-Time Models for Shale Gas Systems Analytical/Numerical Models for Shale Gas Continuous EUR Production Analysis for Shale Gas Systems Correlation/Quality Control for Data Analysis Petrophysical Properties of Shales/Tight Gas Sands Priority very high very high very high high high medium Slide 61/63

62 Students: Blasingame Boulis: Extended Hyperbolic Models for DCA (M.S./complete) Carballo: TBA (Gas Reservoir Eng.?) (Ph.D. coursework DL) Currie: Simple Rate-Time Models for Shale Gas Systems (M.S./active) Freeman: Numerical Models for Shale Gas Well Performance (M.S./active) Ilk: Production Analysis Tight Gas Systems (Ph.D./active) Jam: Analytical Models for Shale Gas Well Performance (M.S./active) Johnson: Simplified Production Analysis for Gas Wells (B.S./complete) Mendel: PTA/PA for Heavy Oil Production Systems (B.S./complete) Nass: IPR For Solution Gas-Drive Reservoirs (M.S./active DL) Olsen: Reserves Practices (M.S./start-up DL) Slide 62/63

63 Presentation at Texas A&M U. 17 November 2009 College Station, TX Recent Work in Well Performance Analysis for Tight Gas Sands and Gas Shales End of Presentation Tom BLASINGAME Department of Petroleum Engineering Texas A&M University College Station, TX (USA) Slide 63/63