Engineering Aspects of Unconventional Oil and Gas Reservoirs Background (Production Analysis)

Transcription

1 27 April 2011 Stimulation of Unconventional Oil and Natural Gas Liquid Reservoirs Engineering Aspects of Unconventional Oil and Gas Reservoirs Background (Production Analysis) Tom BLASINGAME Department of Petroleum Engineering Texas A&M University College Station, TX (USA) Dilhan ILK DeGolyer and MacNaughton 5001 Spring Valley Road Suite 800 East Dallas, TX (USA) Slide 1/37

2 Starting Points: Engineering Aspects of UR What we REALLY know Tight Gas is Relatively Easy (Vertical Wells, HP/HT, PVT) Gas Shales are Technically Viable as a Resource (Economics?) We Can Consistency Deploy Horizontal Multi-Fracture Wells What we THINK know The fracture geometry is The phase behavior The p tf to p wf conversion(s) What we may NEVER know Distribution of natural fractures Transport mechanism for gas/liquids in shales (planar? complex?) (Can be extremely complex ) (what about the heavy water load?) What we should know in the near future Minimal/average duration of data required for EUR? Much better understanding of phase behavior. Optimal well spacing/orientation/placement. (impossible) (organic) Slide 2/37

3 Overview: Engineering Aspects of UR Where we are: (history lessons) Background Recent work Where we want to be: (or so we think) (Historical Perspective) (Blasingame/team) Fit for purpose stimulation. (oil/gas/condensate/geology) More effective reservoir monitoring (this is important!) Early EUR (prediction/correlation) Well Spacing (geology + PVT + modeling) How do we get there Better understanding of flowback/dewatering (optimization) Pressure-dependent properties (k, F cd, desorption?) Understanding of the pore-scale (what flows when/how) Petrophysics (conventional petrophysics not adequate) PVT (oil/gas/condensate/water HP/HT) Slide 3/37

4 Orientation: Engineering Aspects of UR Is "shale gas" here to stay? We know where it is We know how to get it (more/less ) We know how to produce it (sort of ) It all boils down to: Price/Timing/Technology What could be "disruptive" Being able to deliver significantly more stimulation energy. "Green energy" initiatives (unlikely in the short term). Supply pressure on oil (this is more likely than we think). Environmental constraints (e.g., limiters to drilling/stimulation). Slide 4/37

5 27 April 2011 Stimulation of Unconventional Oil and Natural Gas Liquid Reservoirs Engineering Aspects of Unconventional Oil and Gas Reservoirs Background (Production Analysis) Tom BLASINGAME Department of Petroleum Engineering Texas A&M University College Station, TX (USA) Dilhan ILK DeGolyer and MacNaughton 5001 Spring Valley Road Suite 800 East Dallas, TX (USA) Slide 5/37

6 Decline Analysis Arps Relations: Base Relations Loss Ratio: 1 qg D dq / dt Loss Ratio Derivative: b d dt g q dq q g q gi exp[ D t] g gi qg g / dt (1 bd ) (1/ b) it q i Trans., AIME (1945) 160, Analysis of Decline Curves J.J. Arps Question(s): How were the Arps' rate relations derived? The BASIS for the Arps' relations i.e., the behavior of the D- and b- parameters, is derived from OBSERVATIONS. These are empirical results. Case Rate-Time Relation Cumulative-Time Relation qg qgiexp[ Dit] qgi Gp [1 exp[ Dit]] D Exponential:(b=0) Hyperbolic: (0<b<1) Harmonic: (b=1) qgi qg (1 bd ) (1/ b) it qgi qg (1 Dit) G G p p i q (1 b) D q D gi i [1 (1 bd gi i ln(1 D t) i i t) 1(1/ b) ] Slide 6/37

7 Decline Analysis: EUR Plots (Arps' relations) Question(s): Graphical extrapolations of EUR? Family of "EUR plots" derived from the Arps' exponential and hyperbolic relations. Hyperbolic "EUR plot" requires a modular computing environment (e.g., a spreadsheet), as multiple variables are established simultaneously. SPE (2005) A Production-Based Method for Direct Estimation of Gas-in-Place and Reserves T.A. Blasingame, Texas A&M U. and J.A. Rushing, Anadarko Petroleum. Case Exponential: (b=0) qgi q g q gi D i G p G Di Hyperbolic: (0<b<1) 1 Gp q b g qgi 1 (1 ) G G Harmonic: (b=1) q g q gi D exp q i gi G p q gi (1 b) D i Plotting Function q g versusg p Gp log( qg) versuslog1 G log( q g) versusg p Slide 7/37

8 Decline Analysis: Blasingame-Rushing EUR Plot Question(s): Is there a distinctly unique mechanism for establishing the validity of the hyperbolic relation? Yes, the "hyperbolic" decline "type curve" plot yields straight-line trends. Gp qg qgi 1 G (1 b) G qgi (1 b) Di SPE (2005) A Production-Based Method for Direct Estimation of Gas-in-Place and Reserves T.A. Blasingame, Texas A&M U. and J.A. Rushing, Anadarko Petroleum. Hyperbolic Decline: (0<b<1) Exponential Decline: (b=0) 1 q g q gi D i G p G qgi Di a. "Hyperbolic Plot:" (log-log format) Provides a straight-line for ALL cases. b. "Hyperbolic Plot:" (Cartesian format) Provides a straight-line ONLY for b=0 case. Slide 8/37

9 Decline Analysis: Fetkovich-Carter Type Curve JPT (June 1980) Decline Curve Analysis Using Type Curves M.J. Fetkovich, Phillips Petroleum SPEJ (October 1985) Type Curves for Finite Radial and Linear Gas Flow Systems: Constant Terminal Pressure Case R.D. Carter, Amoco Production Variables for the Carter Decline Type Curve kt 1 tdd gictir 2 2 w 1 re re 1 1 ln 2 rw rwa 2 q( t) q s Dd rwa rwe kh ( pi pwf ) r. e gibgi ln rwa 2 a. Original format Fetkovich-Carter type curve most important observation is that 0<b<1. For cases where b>1; either transient flow OR external energy is being added to the reservoir system. Transient Stems: (left) Numerical or analytical model (p wf = constant). q(t) is concave up. Depletion Stems: (right) q(t) is concave down. b=0: p wf = con. b=1: q o = con. (q o /p). b>1: transient flow or external drive energy. : numerical gas flow solutions ( =f(p wf /(p i )). Reservoir Properties: k y-axis match. G x&y-axis matches. s r ed match. Slide 9/37

10 Decline Analysis: "Flowing Material Balance" Plot Question(s): What is the "Flowing Material Balance" plot? In simple terms, p wf (t) data are "converted" to p avg (t) data using the pseudosteady-state flow equation, then plotted as a straight-line extrapolation function and "solved" for gas-in-place. JCPT (June 1997), The 'Flowing' Gas Material Balance L. Mattar and R. McNeil, Fekete Associates "Flowing Material Balance" Plot: Theory: Palacio and Blasingame [1993] Mattar and McNeil [1997] Agarwal et al [1999] a. The "Flowing Material Balance" (Normalized Rate-Cumulative Function Plot) formulation is derived using the solution for the diffusivity equation during boundary-dominated flow regime. This formulation provides a direct estimate of the contacted gas-in-place using time, flowing wellbore pressure, and flowrate data. Advantages: Straightforward and intuitive. Shut-in pressures NOT required. Direct estimation of contacted gas-in-place. Limitations: Boundary-dominated flow regime must exist. Slide 10/37

11 Decline Analysis: Palacio Material Balance Time Can the well-reservoir model be inferred from such data? Yes. Is diagnosis sufficient? No, we must also be able to model/history match data with a model (complete process). qg qgi (1 bdi t ) (1 / b) Decline Curve Analysis Using Type Curves Analysis of Gas Well Production Data J.C. Palacio and T. Blasingame, Texas A&M U. Boundary-Dominated Flow Transient Flow? m( p) q g a. Raw (daily) rate and pressure data bottomhole pressures are calculated, note the effect of liquid loading. SPE (1993) Gp m g, pss trans q g (1/4) m( p) q g Gp m g, pss bdf q g (1) b. "Transformed" data shows fractured well response at early times, very strong evidence of closed system at late times. Slide 11/37 Question(s):

12 27 April 2011 Stimulation of Unconventional Oil and Natural Gas Liquid Reservoirs Engineering Aspects of Unconventional Oil and Gas Reservoirs Historical Work (Blasingame Team) Tom BLASINGAME Department of Petroleum Engineering Texas A&M University College Station, TX (USA) Dilhan ILK DeGolyer and MacNaughton 5001 Spring Valley Road Suite 800 East Dallas, TX (USA) Slide 12/37

13 Decline Analysis: Tight Gas Systems x X Pressure Monitoring Point No. 2 SPE (2007) 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 13/37

14 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 14/37

15 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 15/37

16 Low k Tight Gas Sands: Petrophysics/Permeability Gas Flow v y v x SPE Improved Permeability-Prediction Relations for in Low Permeability Sands F.A. Florence, Occidental Petroleum Corp., J.A. Rushing, Anadarko Petroleum Corp., K.E. Newsham, Apache Corp., and T.A. Blasingame, Texas A&M U. a. Gas Slippage Kundt, A. and Warburg, E.: "Über Reibung und Wärmeleitung verdünnter Gase, " Poggendorfs Annalen der Physik und Che-mie (1875), 155, 337. b. Knudsen "microflow" model (Modified from Karniadakis and Beskok, 2002). k k a tan 4a0 1 k p m 0.4 a 1 a2 a1 a2 1 11/ a c. Microflow model and correlation, "fully implicit" formulation. a 0 1 k p m k p m a1 a2 Slide 16/37

17 Low k Reservoirs: Characteristic Behavior SPE The Characteristic Flow Behavior of Low- Permeability Reservoir Systems T.A. Blasingame, Texas A&M U. A B a. Systematic "mapping" of the inter-relation of petro-physical properties. Note that Archie observed that permeability was "connected" to saturation, porosity, and electrical properties but the relationship was vague, as it remains today. (Formation Resistivity Factor) (Permeability, md) b. Crossplot of formation (resistivity) factor versus permeability (F = A/k B ). C Legend: SEM Micrographs A. (240X) Grains with clay overgrowths. B. (2000X) Microporosity formed by illite clay filaments. C. (600X) Microporosity and clay filling. D. (1400X) Rosettes of chlorite (note illite deposition). c. Severe influence of clay minerals in this reservoir system production shown to be uniquely tied to reservoir quality and effectiveness of well stimulation. D Slide 17/37

18 Permeability: Characteristic Behavior SPE Towards a Characteristic Equation for Permeability A.A. Siddiqui, D. Ilk, T.A. Blasingame, Texas A&M U. k c exp[ ] k xa b (1 x ) exp[ ] Slide 18/37 k a ( c ) b

19 Production Analysis: Elliptical Flow 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 and R.B. Sullivan, Anadarko Petroleum Corp., and T.A. Blasingame, Texas A&M U. [DI] b. Production history plot with model match (very good flowrate match, acceptable pressure match). c. Results correlation plot G versus k. Slide 19/37 a. Elliptical boundary decline type curve match (very high conductivity, "thin" elliptical drainage geometry).

20 Production Analysis: Power-law Exponential Decline 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 U., J.A. Rushing and A.D. Perego, Anadarko Petroleum Corp., and T.A. Blasingame, Texas A&M U. [DI] b. Derivation of the power-law exponential relation. c. "qdb" plot Data matched using power-law exponential rate decline relation. Slide 20/37 a. Production forecast of a tight gas well using Arps' hyperbolic decline relation.

21 Production Analysis: Continuous EUR q qˆi exp[ D t Dˆ i t n ] SPE Continuous Estimation of Ultimate Recovery S.M. Currie., D. Ilk, and T.A. Blasingame, Texas A&M U. [DI] b. Continuous EUR workflow including Hyperbolic, Power-Law Exponential, and straight-line extrapolation models. c. Summary "Continuous EUR" plot note all model results are plotted in time. Slide 21/37 a. Base data plot ("qdb-plot"), data matched using Power Law Exponential and Hyperbolic models.

22 Production Correlation: Flowback Analysis SPE A Comprehensive Workflow for Early Analysis and Interpretation of Flowback Data from Wells in Tight Gas/Shale Reservoir Systems D. Ilk and S.M. Currie, Texas A&M U., D. Symmons, Consultant, J.A. Rushing, Anadarko Petroleum, N.J. Broussard, El Paso, and T.A. Blasingame, Texas A&M U. c. Crossplot All wells: p2/qg versus t. [DI] b. Crossplot All wells: pcf versus t. Slide 22/37 a. Crossplot All wells: qg versus qw.

23 Production Analysis: Time-Rate Diagnostics SPE q,cp (t ) Hybrid Rate-Decline Models for the Analysis of Production Performance in Unconventional Reservoirs d ln( q) t dq d ln(t ) q dt D. Ilk and S.M. Currie, Texas A&M U., D. Symmons, Consultant, J.A. Rushing, Apache, and T.A. Blasingame, Texas A&M U. q,cp (t ) d ln( q ) t dq d ln(t ) q dt a. -derivative Holly Branch Wells. d ln( q) t dq d ln(t ) q dt [DI] b. -derivative "Shale Gas Field C" Wells. c. -derivative All Models (Holly Branch Well). Slide 23/37 q,cp (t )

24 Production Analysis: Integration of Results SPE Integration of Production Analysis and Rate-time Analysis via Parametric Correlations Theoretical Considerations and Practical Applications D. Ilk, DeGolyer and MacNaughton, J.A. Rushing, Apache, and T.A. Blasingame, Texas A&M U. k a n b Dˆ c i qˆ d i "Shale Gas Field C" Horizontal well with multiple vertical fractures: a. Correlation plot k cal versus k meas. [DI] EUR ˆ n ˆ "Shale Gas Field C" b. Correlation plot EUR cal versus EUR meas. Power-Law Exponential Relations: 1 dq ˆ (1 n) D( t) D nd i q dt t ( ) ˆ exp[ ˆ n q t qi D t Di t ] Slide 24/37

25 Production Analysis: Workflow/Unconv. Reservoirs SPE Production Analysis in Unconventional Reservoirs Diagnostics, Challenges, and Methodologies D. Ilk, C.D. Jenkins, DeGolyer and MacNaughton, and T.A. Blasingame, Texas A&M U. [DI] b. History match plot (selected well) q and pwf vs. t. c. q/ p data (all wells) with constant rate solution and Gp vs. t for sensitivity analysis for various parameters. Slide 25/37 a. Diagnostic plots (selected) q/ p vs. t and D-parameter vs. t.

26 27 April 2011 Stimulation of Unconventional Oil and Natural Gas Liquid Reservoirs Engineering Aspects of Unconventional Oil and Gas Reservoirs Well Performance Analysis for Tight/Shale Oil Reservoir Systems [DI] Tom BLASINGAME Department of Petroleum Engineering Texas A&M University College Station, TX (USA) Dilhan ILK DeGolyer and MacNaughton 5001 Spring Valley Road Suite 800 East Dallas, TX (USA) Slide 26/37

27 Field Example: Bakken Oil Well [DI] Discussion: Bakken Oil Well Change in production trend after almost 120 days. Off-trend data are removed prior to analysis. Slide 27/37

28 Field Example: Bakken Oil Well [DI] Discussion: Continuous EUR using Arps' Hyperbolic Decline b-parameter value ranges between 3.7 (earliest interval) and 1.5 (all data) for the subsets of data. No strong evidence of boundary-dominated flow is observed. Slide 28/37

29 Field Example: Bakken Oil Well [DI] Discussion: Continuous EUR using Power-law Exponential Model Regression is used to match the data with the model at specified intervals. Variability in the production trend yields different matches. Slide 29/37

30 Field Example: Bakken Oil Well [DI] Discussion: N p,max using Straight Line Extrapolation Straight line extrapolation yields the minimum value for the reserves estimate as boundary-dominated flow assumption is made. N p,max (t) increases with time. Slide 30/37

31 Field Example: Bakken Oil Well [DI] Discussion: Conclusions EUR estimates from "hyperbolic" relation are greater than estimates from PLE especially at early times. EUR (at 30 years) should lie between 0.14 and 0.26 MMSTB. Slide 31/37

32 Field Example: Bakken Oil Well Log-log plot: Normalized rate functions versus material balance time (data). Log-log plot: Normalized rate functions versus material balance time (data and models). [DI] Discussion: Model-Based Production Analysis Good diagnostic character of the data functions. No strong evidence of linear flow is observed. Good match of the data with the analytical model solution. Model: Horizontal well with multiple transverse fractures. Slide 32/37

33 Field Example: Bakken Oil Well [DI] Discussion: Model-Based Production Analysis Very good match of the data with the analytical model. No volume results INFINITE-ACTING RESERVOIR. Production forecast at 30 years = 0.29 MMSTB. Consistent estimate with rate-time analysis. Slide 33/37

34 27 April 2011 Stimulation of Unconventional Oil and Natural Gas Liquid Reservoirs Engineering Aspects of Unconventional Oil and Gas Reservoirs Where we want to be and how do we get there Tom BLASINGAME Department of Petroleum Engineering Texas A&M University College Station, TX (USA) Dilhan ILK DeGolyer and MacNaughton 5001 Spring Valley Road Suite 800 East Dallas, TX (USA) Slide 34/37

35 Big Questions: Engineering Aspects of UR EUR? Early EUR? EUR = f(t)? Well Spacing? QUANTIFYING reservoir properties? Pressure Transient Analysis Production Analysis Petrophysical analysis Liquids-Rich Systems? Fluid-Flow Mechanisms PVT Improved Recovery Fit-for-Purpose Stimulation Artificial Lift Oil-Prone Systems? Recovery (can this be meaningful?) (how do we incorporate this?) (is this really the holy grail?) (ultra-low k?) (p tf may not be sufficient) (theory application) (what is really flowing where?) (near-critical fluids are not trivial) (we all know this is coming ) (higher F cd, more complexity) (fact of life ) (low to extremely low primary recovery) Slide 35/37

36 Perspectives: Engineering Aspects of UR Never-Ending Arguments SRV (What is it, really?) Desorption (Significance? Timing? Relevance?) Stimulation Fluids (Where does it go? Does it matter?) Microseismic (Crystal ball or roulette wheel?) Pressure-Dependent Whatever (So what?) Natural Fractures (If/when/why/what?) Dual Porosity/Dual Permeability (What about the physics?) Well Placement/Effect of Layering (When does it matter?) Things that SHOULD help Production Logs (But just a snapshot in time.) Optimal Proppant Design/Placement (Obvious, but ) Stimulation Stages/Perforation Clusters (Geology + logs) Things that DEFINITELY WOULD help Measured p wf Downhole Fluid Sampling Horizontal Core (Yes, this is my favorite song ) (Sooner or later ) (Why not?) Slide 36/37

37 27 April 2011 Stimulation of Unconventional Oil and Natural Gas Liquid Reservoirs Engineering Aspects of Unconventional Oil and Gas Reservoirs End of Presentation Tom BLASINGAME Department of Petroleum Engineering Texas A&M University College Station, TX (USA) Dilhan ILK DeGolyer and MacNaughton 5001 Spring Valley Road Suite 800 East Dallas, TX (USA) Slide 37/37

38 Nelson, P.H.: "Pore-throat sizes in sandstones, tight sandstones, and shales," AAPG Bulletin, 27 April 2011 Stimulation of Unconventional Oil and Natural Gas Liquid Reservoirs v. 93, no. 3 (March 2009), pp Slide 38/37 Point to Ponder: Nelson Pore/Molecule Size Chart