Fractured Reservoir Simulation Milind Deo and Craig Forster

Fractured Reservoir Simulation Milind Deo and Craig Forster University of Utah Fracture Dominated % of Total Permeability 100% Fracture I II IV 100% Matrix Dominated After Nelson (2001) % of Total Porosity 100% Fracture

Account for Irregular Geometries Hypothetical ‘Real System’ Common Model Properties Regularized Equivalent ECLIPSE Equivalent • Impermeable matrix with f = 0 (Type I, basement reservoir system) • Domain = 1, 000 ft by 200 feet deep • Total feature length = 30, 000 feet • Reference Case: Feature k = 1, 000 md, f = 14 %, width = 0. 5 feet • OOIP = 53, 580 STB • Injection Pressure = 4, 300 psi • Injection Well Production Well

Model Comparison at 900 days 0. 80 0. 65 So 0. 50 0. 35 0. 20 Oil Production (STB/day) Hypothetical ‘Real System’ CVFE 120 100 80 60 40 20 00 Regularized Equivalent CVFE ECLIPSE Equivalent ‘Real System’ CVFE Regularized Equiv. CVFE ECLIPSE Primary Production 300 Time (days) 600 900

Parameter Sensitivity (900 days) 0. 80 0. 65 So 0. 50 0. 35 0. 20 Oil Production (STB/day) 1000 md Base Case 1000 md 1 Connection 120 100 80 60 40 20 00 Geom. Mean 100 md 33: 33 Random k 1000 md Base Case Primary Production 1 Connection GM 100 md 50: 50 GM 100 md 33: 33 300 Time (days) 600 900

Faulted-Fractured Reservoir System Sandstone Increasing Fractures t l au F Increasing Fractures Shale 1000 ft 200 ft k (md) f (%) ft 0 ft ss ne tt hi ck Fa ul Upper Reservoir = 0. 3 100 Lower Reservoir 10000 5 500 11 100 10 10 8 kh (md-ft) 100 Production Wells Secondary Injection Wells (production wells during primary prod’n)

Faulted-Fractured Reservoir 0. 80 0. 65 So Primary Water Flood Production Start after 600 days 600 3000 20 days 6000 days 0. 50 0. 35 0. 20 Production Well Injection Well Viewed From Bottom of Model Domain

Impact of Fault k = 10, 000 md Primary Production Secondary Production after 600 days 0. 80 0. 65 So 0. 50 0. 35 0. 20 800 Primary Production From Fault Lower Res. 400 0 6000 3000 600 20 days 0 6000 3000 Time (days) Water Cut (vol/vol) Injection Well Oil Production Rate (STB/day) Production Well Fault k = 100 md Primary Production 0. 70 0. 35 0 0 3000 Time (days) 6000

Outcrop-to-Simulation 1 m Teasdale Fault, Utah 10 Kilometers Cocks Comb Study N Joint Zone U D Utah High k Features Cottonwood Wash Study Field Area feet 1800 10 1 kh (md-ft) ive 10 Line Drive 2 Dr 25 ne 100 Li k (md) f (%) 16000 2500 Production • Primary 600 days • Production BHP 2200 psi • Injection BHP 3200 psi 200 feet 1800 feet

Oil Production Rate (STB/day) Well Placement Strategies 3500 Line Drive 1 (N to S) 2500 1000 Line Drive 2 Line Drive 1 500 0 3000 Time (days) Primary Production 6000 So 0. 80 0. 65 2100 0. 50 0. 35 Gas Oil Ratio Line Drive 2 (W to E) 1800 1400 1100 0 4000 2000 1000 600 days 0. 20 Line Drive 2 Line Drive 1 3000 Time (days) 6000 View From Bottom of Model Domain

Status and Challenges • Preserve geologic integrity while constructing simulation models • Relationship between discrete-fracture models and dual-porosity models – DFN as upscaling/calibration tool for DP models? – Field-wide DFN models in the near future? • Fundamental work – – – – – Additional physics? Upscaling Integrating geomechanics Dynamic data updaing Different discrete-fracture implementations Discretization schemes Gridding Efficiency of solvers High-performance (parallel) computing

Acknowledgements q U. S. DOE Contract DE-FC 26 -04 NT 15531 through the National Energy Technology Laboratory. q Schlumberger Inc. – Eclipse academic license q Sandia National Laboratories – CUBIT license q Argonne National Laboratory – PETSc q Our eam – – – – – Jim Evans, Professor, Utah State University, Logan, Utah Tom Doe, Golder and Associates Yi-kun Yang, Post-doc Sriram Balasubramaniam, Graduate student Ganesh Balasubramaniam, Graduate student Yao Fu, Graduete student Kan Huang, Graduate student Zhiqiang Gu, Graduate student Huabing Wang, Graduete student