Thomas Dewers 1 , Alex Rinehart 1 , Jon Major 2 , Photos placed in - PowerPoint PPT Presentation

G EOMECHANICS OF G EOLOGIC C ARBON Photos placed in horizontal position S TORAGE : H AZARDS , L ONG -T ERM S EALING , with even amount of white space between photos AND S TORAGE S ECURITY and header Thomas Dewers 1 , Alex Rinehart 1 , Jon Major 2 , Photos placed in horizontal position Peter Eichhubl 2 , Pania Newell 1 , Mario with even amount of white space Martinez 1 , Joseph Bishop 1 , Steven Bryant 2 between photos and header 1 Sandia National Laboratories, Albuquerque NM and 2 University of Texas, Austin TX Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE -AC04-94AL85000. SAND NO. 2011-XXXXP SAND No. 2012-xxxxP

CFSES Approach to GCS Geomechanics  Caprock Heterogeneity and Injection Response  Time dependent growth of fractures in cap rock  Heterogeneity and geomechanical response of subsurface reservoirs  Chemo-mechanical coupling during injection  Modeling of consequences of CO 2 injection  Parametric coupled flow and geomechanical analysis of surface uplift 2

Core samples for geomechanical testing from NETL MGSC, SECARB, and SWP Mt Simon Sandstone @ 2 km Tuscaloosa Sands @ 3km US National Energy Technology Laboratory Partnership Demonstration Projects 3

Heterogeneity and geomechanical response of subsurface reservoirs 4

Core, Well Logs and Sampling • Three well-log units (U, M, L) • Three sampled lithofacies (I, II, III) • Similar porosities but markedly different permeabilities • Distribution of facies similar to those on east flank of Illinois Basin (Saeed and Evans, 2012) • Similar to lower portions of Illinois Basin lithofacies (Bowen et al. 2011) incl. main injection horizon but lacking upper “B - cap” muddy facies 5 Dewers et al., IJGHGC, Accepted

Lithofacies Interpretation • I Lithofacies (main injection unit in IB): quartz-rich sand flat B1 facies of Saeed and Evans [2012] or the “sandy tidal” facies of Fischietto [2009] • II Lithofacies: heterolithic T2 “mixed flat” facies and “sand flat to tidal channel” B2 facies of Saeed and Evans [2012] or the “mixed fluvial -eolian tidal” and “braided fluvial” facies of Fischietto [2009] • III Lithofacies: mud flat T1 facies of Saeed and Evans [2012] or the muddy tidal facies of Fischietto [2009] I Lithofacies II Lithofacies III Lithofacies 6

Yield and Failure Envelopes Failure envelope:     a I F ( I ) a a e a I 2 1 f 1 1 3 4 1 Yield Surface f ( I ) f ( I )  f 1 c 1 J   2 ( )   f ( I ) F ( I ) N f 1 f 1        ( I )( I ( I ))    1 1 1 2 f c ( I , ) 1   1   2 2 X (After Brannon et al., 2009; Pelessone, 1989) 7

Kayenta Model Validation Kayenta* Includes: • Non-Associative Plasticity • Stress Invariant Dep. Failure • Elliptical Cap Surface • Kinematic Hardening • Isotropic Hardening • Nonlinear Elasticity • Elastic-Plastic Coupling *Developed by Brannon et al. 2009 8 Dewers et al., IJGHGC, Accepted

Modeling of consequences of CO2 Injection 9

Application: Leakage Pathway Development in Jointed Caprock Overpressure, MPa, Effective Stress, MPa, 5 Y at 5 Y at 3MT/Y 3MT/Y CO2 Saturation, 5 Y at 3MT/Y CO2 Saturation, 5 Y at 5MT/Y Coupled injection, multiphase flow, and geomechanics simulation showing range of over pressure and effective stress (top, at 5 yrs for 3 MT/yr) and CO 2 saturation (bottom at 3 MT/yr (a), and 5 Mt/yr (b). Overpressure induced by higher injection rates results in opening of caprock fractures and leakage. Martinez et al., IJGHGC, 2013 10

Application: Induced Seismicity In Mt Simon Sandstone Site of Youngstown Ohio 2011 earthquakes thought to be triggered by fluid injection into the Mt Simon Locations of earthquake epicenters (circles) and injection wells for Guy, Arkansas 2010-2011 and Lake, Ohio 1983- 1986 earthquake swarms Simulated pore pressure during injection, with failure occuring within gold shadied regions Person et al., Groundwater, 2013 11

Application: Reservoir Drawdown and Wellbore Damage During CO2 Injection and Brine Withdrawal Map view of wellfield with injectors (triangles) and extractors (circles) Drawdown at Mt Simon extraction wells can induce shear failure and wellbore damage Heath et al., ES&T, 2014 Threshold pressure for breakouts/well 12 sanding determined experimentally

Chemo-mechanical coupling during injection 13

Does chemistry of injectate influence geomechanical response? • Pressure response during injection suggests a “ geomechanical event” • Tracer studies show increase in permeability with “event” Field and simulated pressure response at Cranfield injection zone (Kim and 14 Hosseini, 2013)

Lithofacies • Mixed chlorite- and quartz- cemented muddy cross- bedded fine sandstone (Facies B); • Quartz-cemented tabular very fine sandstone (Facies C) • Chlorite-cemented conglomeratic sandstone (Facies A); Geomechanical Testing • T, P and major ion fluid chemistry • UCS, hydrostatic and triaxial stress paths • Pore fluid equilibrated with scCO2 Rinehart and Dewers, 2015 15

Chemical effects may have caused Yield and Failure “ Geomechanical Event” Surfaces • Accelerated creep at low stresses In situ consistent with stress corrosion model (pH-activated?) • Failure of chlorite facies below in situ stresses Facies A • Grain-coating chlorite delamination (pH activated?) Facies B Facies C Degradation of elastic moduli masks the compactional- to-dilational turn-around in volume strain 16

Post-70 MPa Test (Away Post-70 MPa Test Untested From Fracture) (Near Fracture) Untested has chlorite cements, opaque grains, grain casts with remnant cements. Tests show pore collapse, grain shattering with increased intensity near 17 fracture, and concentration of opaque grains and pore-filling clay minerals.

Untested Post-70 MPa Test Well-defined crystalline chlorite rims Degraded chlorite rims without crystal habit • Exposure to CO2-equilibrated brine at 100 ° C overnight (24 hours) influences chlorite cement crystallinity. • Post-test photomicrograph is away from fracture feature, degradation primarily a chemical effect but is pervasive through the sample. • Other reservoir facies show enhanced creep but no substantial degradation. 18

SolidWorks View of in situ short-rod fracture tester WAVE SPRING SAMPLE ACTUATOR LVDT

Conclusions  Sandstone reservoirs exhibit a range of heterogeneity and rock mechanical responses to changing stress paths associated with injection. Weaker facies (e.g. Cenozoic sands in Gulf Coast & Colorado Plateau, facies II and III for Mt Simon) exhibit elastic-plastic coupling.  Chemo-mechanical effects – good (improved injectivity & sweep efficiency) or bad (time-dependent leakage through caprock)?  Validated Kayenta constitutive model captures essential features of sandstone reservoir geomechanical behavior. It can be included in most FEM models.  Experimental/Modeling approach for weak sandstone reservoirs informs models on caprock leakage, induced seismicity, and wellbore damage. Can inform regulatory constrains on injectivity (i.e. frac gradient) and withdrawal (borehole shear failure). 20