Modeling of the Steam Chamber Growth During SAGD* Eric Bathellier1, Olivier Lerat2, Axelle Baroni2, and Gerard Renard2 Search and Discovery Article #41380 (2014)** Posted July 7, 2014 *Adapted from oral presentation given at Pacific Section AAPG, SEG and SEPM Joint Technical Conference, Bakersfield, California, April 27-30, 2014 **AAPG©2014 Serial rights given by author. For all other rights contact author directly. 1 CGG, Houston, TX ([email protected]) IFPEN, Rueil-Malmaison Cedex, France 2 Abstract This paper presents an integrated workflow for the interpretation of 4D seismic data to monitor the steam chamber growth during the steam-assisted gravity drainage recovery process (SAGD). Superimposed on reservoir heterogeneities of geological origin, many factors interact during thermal production of heavy oil and bitumen reservoirs, which complicate the interpretation of 4D seismic data: changes in oil viscosity, fluid saturations, pore pressure, and so on. The workflow is based on the generation of a geological model inspired by a real field case of the McMurray formation in the Athabasca region. The approach consists of three steps: 1. Construction of an initial static model at the field scale, 2. Simulation of thermal production of heavy oil with two coupled fluid-flow and geomechanical models, 3. Computation of synthetic seismic data at different stages of steam injection. Production scenarios are run to obtain pore pressure, temperature, steam and oil saturations on a detailed reservoir grid around a well pair at several stages of production. Direct coupling with a geomechanical model produces volumetric strain and mean effective stress maps as additional properties. These physical parameters are used to compute new seismic velocities and density for each stage of production. A new synthetic seismic image of the reservoir is generated for each stage of production. The impacts of heterogeneities, production conditions and reservoir properties are evaluated for several simulation scenarios from the beginning of steam injection to 3 years of production. Results show that short-term seismic monitoring can help in anticipating early changes in steam injection strategy. In return, long-term periods allow the behaviour of the steam chamber to be monitored laterally and in the upper part of the reservoir. This study demonstrates the benefit of 4D seismic data in the context of steam-assisted heavy oil production. Modeling of the Steam Chamber Growth During SAGD Eric Bathellier (CGG), Olivier Lerat, Axelle Baroni, Gérard Renard (IFPEN) 2014 PS-AAPG Meeting - April 29 - Bakersfield, CA Steam Assisted Gravity Drainage Vertical 2-D Cross section overburden Steam chamber \ / m injection well J production well underburden Conventional 4D Seismic Data between 2001 and 2004 surveys between 2001 and 2005 surveys Zhang et al., 2005 Map views of 4D seismic amplitude difference Steam injection  reduction of viscosity and mobility ratio  But also: rock and fluid expansion, compaction, oil vaporization,… Continuous and Permanent Seismic Monitoring (SeisMovieTM) Injected steam rate Observation wells Project Objectives  Imaging of the steam chamber evolution from 4D seismic data at early times of SAGD steam injection phase  Demonstration and promotion of the SeisMovie technology in heavy oil and bitumen production  Improvement of the understanding of physical laws driving the petro-elastic model during steam injection Presentation Outline  Workflow  Construction of the full-field static model  Coupled modeling  Seismic modeling  Summary Workflow facies Φ Sat, P, T Stresses TWT Velocities K (initial) Initial Geological Model Sat P T Sagd Pilot Fluid -Flow Simulation Geo mechanical mechanical properties, Model constitutive laws Φ K σ ε Petro-Elastic Params t0 t1 t2 t3 Impedances Maps of Attribute Differences Sensitivity Tests on PEM – One-way coupling of reservoir and geomechanical models – Short to long periods of steam injection (weeks -> 6 years) – Sensitivity study 1. Construction of the Full-field Static Model Geological Model and Static Properties Hangingstone Field Data • Athabasca region (Alberta, Canada) • McMurray Formation • Oil viscosity 1 000 000 cp • Oil density 8° API • 32 horizontal wells • 50 vertical wells • 10 cored wells • Production data (90 months) N E doublet Facies and Logs – Core Calibration CORE ELECTRO DESCRIPTION FACIES RESERVOIR UNIT Lithofacies 1 Lithofacies 2 Lithofacies 3 Lithofacies 4 Lithofacies 5 McMurray Fm. Open bay shales Unit 4 Stacked tidal flats, Channels and bars Unit 3 Tidal ravinement Stacked meandering channels With tidal influence Unit 2 Amalgamated fluvial Braided channels Unit 1 Channel belt incision Coastal plain ? Base Cretaceous unconformity 10 Devonian Carbonates vertical exaggeration x10 90 layers X,Y: 10m x 10m Z: 0.5 - 0.6 m 4.95.106 cells 1 2 reservoir 3 4 heterogeneity 5 Geostatistical Simulation of Facies Distribution 11 Seismic Modeling Before Production materials 1D seismic modeling (reservoir zone): Lithofacies (top), reflectivity coefficients convolved by a 80Hz Ricker wavelet (bottom) 2. Coupled Modeling Reservoir Simulation (Pumaflow) Geomechanical Modeling (Abaqus) Definition of the Local SAGD Reservoir Model Mesh: 10x2.5m; 50x1m; 10x2.5m Y: 41x20m (235,000 cells) Scenario for SAGD Modeling Operating conditions in the wells  Warm up phase – Four months @ constant T = 220°C  Steam injection: up to 6 years – Real injection-production history at wells – Steam trap control implemented Properties Exported to the Reservoir Model 0.001mD 0.005mD 1000mD 3000mD Horizontal Permeability 20% Vertical Permeability 1.0 10% 34% Porosity 0.15 Water Saturation % of steam rate injected in the Percent steam rate injected in thewell well % of steam rate in the injector 12 Date Date Date Date Date Date Date Date Date 11 10 9 8 7 5/7/00 10/7/00 20/7/00 30/7/00 30/8/00 1/10/00 1/11/00 1/12/00 1/1/01 (time (time (time (time (time (time (time (time (time 5 days) 10 days) 20 days) 30 days) 61 days) 92 days) 122 days) 153 days) 183 days) Reservoir simulation with Pumaflow 6 5 4 3 2 1 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 Perforation number heel toe 19 Well pair profile Mechanical Behavior of Shale Materials? 5 23 Explicit coupling (update of permeabilities) One-way coupling heel toe 1 5 9 14 18 23 30 36 41 Materials Temperature Pore Pressure Sec° 9 Sec° 30 Effective Stress 23 Update of the petroelastic model 3. Seismic Modeling Impact of Thermal Production on PEM (Petro-Elastic Model) Velocities Sensitivity Seismic velocities V = f(Knd,Gnd,1/ρ) Incompressibility & shear modulus Knd = f(Kd,Kgr,Kfl,φ), Gnd = f(Gd,Ggr,Gfl,φ) Fluid parameters Kfl=f(P,T,S), Gfl=f(P,visco(T),S), ρfl= f(P,T,S) input visco(T) output P,T,S Reservoir modeling Grain parameters ρgr, Kgr, Ggr input P,T Density ρ = ρgr*(1-φ)+ρfl*φ Drained Modulus Kd =f(σeff) Gd=f(σeff) output σeff, φ Geomechanical modeling Geomodeler (Geometry, Parameters...) Synthetic Seismogram in Time P-wave 80Hz Time 2000 March initial April May June 2001 July Aug. Warm up 1 2 Sept. Oct. Nov. Dec. Jan. production 3 4 5 6 Depth Slices P-wave Seismogram Difference @Z =-314.5m 2-1 3-1 4-1 5-1 6-1 Summary  Fully integrated study from static to dynamic modeling – Facies, petrophysics, geomechanics, petroacoustics  Simulations of full production history – Steam rate matched in the injector – Oil and water rate matched in the producer – Proportion of oil and water respected – Lateral steam connection between sections is taken into account  Impact of heterogeneities on steam chamber development – Influence of shale beds is clear on 3D visualizations – Mechanical behavior of shales needs to be further characterized Summary (continued)  Seismic modeling – Petroelastic modeling shows realistic images – Model updates according to dynamic properties evolution  Monitoring – Improved understanding expected through SeisMovie interpretation Thank you! Questions?