UNIVERSITY OF

TORONTO UNIVERSITY OF

TORONTO

Rock Fracture Dynamics and Induced Seismicity Paul Young, Professor of Seismology and Rock Mechanics, University of Toronto, Canada.

Acknowledgements and THANKS to M.H.B (Farzine) Nasseri, Laszlo Lombos, Sebastian Goodfellow, Alex Schubnel, Juan Reynes-Montes, Xueping Zhao, my research group and colleagues (past and present), research partners and sponsors including UofT, Ergotech, MTS, ASC, Itasca, CFI, OIT, MRI, NSERC, MMT.

ARMA Plenary Lecture, Chicago, USA ARMA Plenary Lecture, Chicago, June, 2012

June 25th, 2012

1

UNIVERSITY OF

TORONTO

Overview • Induced Seismicity (Field)  Imaging the subsurface  State of the art, challenges and opportunities

• Rock Fracture Dynamics and Induced Seismicity (Lab)  Some previous research, standing on the shoulders of giants  True-Triaxial lab experiments, induced seismicity and

geophysical imaging  Seismicity, velocity and permeability

• Conclusions and Future Potential for Induced Seismicity ARMA Plenary Lecture, Chicago, June, 2012

2

UNIVERSITY OF

TORONTO

Seismicity: Natural and human sources • When a material ( e.g. rock) undergoes brittle failure, elastic energy is radiated from the point of failure (or slip) into the surrounding medium.

ARMA Plenary Lecture, Chicago, June, 2012

3

UNIVERSITY OF

TORONTO

Imaging the Earth with Seismicity

ARMA Plenary Lecture, Chicago, June, 2012

4

UNIVERSITY OF

TORONTO

Seismic Signals • The signal recorded at any one sensor is the convolution of the source magnitude and other properties, M(t), the transmission media, G(t), and the sensitivity of the instrument, S(t). • Understanding the effect of each is key to understanding seismicity.

ARMA Plenary Lecture, Chicago, June, 2012

S(t) G(t)

M(t)

The “instrument” includes the seismic transducer and all electronics that combine to record the waveform. 5

UNIVERSITY OF

TORONTO

Induced or Triggered Seismicity • There are many examples where human activity causes perturbations of the Earth’s crust that lead directly or indirectly to seismicity. • Studies of such ‘stimulated’ seismicity provide important insights into the factors controlling crustal seismicity, both natural and ‘artificial’. • Monitoring of stimulated seismicity can provide critical feedback on the engineering performance of a particular site or infrastructure. • Induced

Where the causative activity can account for most of the stress change or energy required to produce the seismicity

• Triggered

Where the causative activity accounts for only a fraction of the stress change or energy associated with the seismicity (i.e. tectonic loading plays a primary role)

(after McGarr & Simpson, 1997

ARMA Plenary Lecture, Chicago, June, 2012

6

UNIVERSITY OF

TORONTO

Induced MicroSeismicity (MS) and Scaling Km & 100s Hz

0.1m & 100s kHz

100m & kHz

10m & 10s kHz Thompson, BD Young, RP and Lockner, DA, (2009). JGR, Vol. 114.

ARMA Plenary Lecture, Chicago, June, 2012

7

UNIVERSITY OF

TORONTO

•

MS Locations and Hydraulic Fracturing

The analysis of the spatial distribution of induced MS events provide critical information of the fracture network stimulated through hydraulic treatments: – –

Fracturing extent and geometry Quantification of the stimulated rock volume

UP S W

ARMA Plenary Lecture, Chicago, June, 2012

Single W ing

No

Breakthrough

Yes

Fracture N etw ork W ing Length

173 m

Fracture N etw ork Half Length

133 m (Ewing)

Fracture N etw ork Height

26 m

Fracture N etw ork W idth

18 m

Fracture N etw ork Top

3,976 m (TVD SS)

Fracture N etw ork Bottom

4,002 m (TVD SS)

Fracture N etw ork Azim uth

97 degrees E of N

Fracture N etw ork P lunge

90 degrees

8

UNIVERSITY OF

TORONTO

Challenges and Opportunities • • • • • •

S-wave Location Methods Relative Location Source Mechanisms Analysis of the Continuous Data Streams Enhanced Velocity Models Synthetic Seismicity Modelling

ARMA Plenary Lecture, Chicago, June, 2012

9

UNIVERSITY OF

TORONTO

Increasing MS locations •

The success of MS monitoring relies on the accurate location of the highest number of events. Low signal-to-noise ratios are often experienced in many engineering environments. This has the effect of causing difficulties in the identification and picking of waveform phases, required to obtain source vectors for a successful location from arrays with limited geometry. Different approaches can be used to enhance the number and accuracy of located events:  

Increasing the number of events with identified phase arrivals. Using a location algorithm independent of the availability of both P and S wave arrivals and source vector information. 100

% of located M S events

90 80 70 60 50 40 30 20 10 0

H ydrofracture project

ARMA Plenary Lecture, Chicago, June, 2012

10

UNIVERSITY OF

TORONTO

S-wave Polarization Methods •

•

•

A significant number of seismic events recorded during a hydrofracture treatment display a high energy S-wave arrival but have a P-wave that is close to or below the ambient noise level. Traditional location methods, relying on P-wave polarization information to determine the source vector, therefore fail to determine a source location for these events. The S-wave polarization can be investigated using similar methods as those used to analyze the P-wave polarization. Depending upon the nature of the S-wave polarization due to transmission effects, either the full source vector or the plane containing the source vector can be estimated. The use of this additional information in the standard location algorithm has been applied in field operations, increasing up to 8-fold the number of located events. ARMA Plenary Lecture, Chicago, June, 2012

a) MS events located using source vector orientations from P-wave polarization. b) MS events located using source vector orientations from S-wave polarization

11

UNIVERSITY OF

TORONTO

•

•

•

•

Relative location methods (master event or double difference) is based on the use of travel time differences between events, being more robust against uncertainties in the velocity model of the larger rock volume between the events and the sensor aray. The method can provide successful location for microseismicity using a single phase and no polarization information on the target events. The technique provides a means to increase location efficiency and thus provide greater information on the fracture network. A stepwise approach is used over a lattice of master events to overcome the requirement of close separation between target and master events. The technique means that a simpler velocity model can be used for the target events.

ARMA Plenary Lecture, Chicago, June, 2012

Relative location

•

Classical location

Relative Location

12

UNIVERSITY OF

TORONTO

Fracture Network Engineering Interpreting fracture diagnostics from microseismic data: •

•

•

Numerical PFC models can simulate the evolution of fracture volume change and network connectivity and perform simulated fluid circulation; Feedback information between observed and simulated data provide Fracture Network Engineering (FNE). Synthetic seismicity, calculated within the modelled Discrete Fracture Network (DFN), can be compared with that observed during monitoring to relate microseismicity to fracture growth.

ARMA Plenary Lecture, Chicago, June, 2012

13

UNIVERSITY OF

TORONTO

MS from Bonded Particle Models and Synthetic Rocks MPa

9/3

m3/s

2×10-3

Magnitude

Time

Similar MS propagating patterns between field and model  Event locations with time  Linear orientations  Truncation or arrest of events in the N-E and S-W directions

1km

Synthetic MS

Fluid flow and induced cracks 500 400

Magnitude

300 200 100 0

Time

-100

MPa

20/3

-200

m3/s

2×10-3

-300 -400

* Zhao et al., 2011, GRC Annual Meeting ARMA Plenary Lecture, Chicago, June, 2012

-500 -500 -400 -300 -200 -100 0

Synthetic MS

100 200 300 400 500m

Field recorded MS 14

UNIVERSITY OF

TORONTO

•

ARMA Plenary Lecture, Chicago, June, 2012

Frequency (Hz)

•

The continuous microseismic record can be used for investigating the fracture network growth and mechanics of hydrofracture data when no MS are induced or identified. It provides a means for diagnosing the quality of a data set and then optimizing the processing of discrete microseismic events. Continuous records could be used to better understand the fundamental hydrofracture propagation mechanics in different geological and treatment scenarios – by directly correlating with scaled laboratory experiments and dynamic numerical models in which seismic energy release is also mapped.

2000 Slurry Flow Rate (bpm)

Treatment

•

Amplitude

Continuous Record Analysis

1500

Observed Net Pressure (psi)

1000 500 0 0

60

120

180

240

300

Time (Minutes)

15

UNIVERSITY OF

TORONTO

Natural Hazards Application to Volcanic Seismicity

Benson PM, Viciguerra S, Meredith PG and Young RP (2008), Laboratory Simulation of Volcano Seismicity, Science, Vol 322.

ARMA Plenary Lecture, Chicago, June, 2012

16

UNIVERSITY OF

TORONTO

Rock Fracture Dynamics Facility (RFDF)

ARMA Plenary Lecture, Chicago, June, 2012

17

UNIVERSITY OF

TORONTO

T rue-T riaxial G eophysical Im aging C ell and P olyaxial T esting M achine

•Polyaxial servo-controlled loading system; 6800 kN axial, 3400 kN lateral •Polyaxial (true triaxial) and triaxial geophysical imaging cells •Temp. up to 200 oC •Full waveform continuous Acoustic Emission (18 sensor 3D array sampled at 10MHz – up to 8hrs) •3D velocity measurement system ( including 6P and 12S axial sensors) •Pore pressure control and 3D permeability along independently controlled axes ARMA Plenary Lecture, Chicago, June, 2012

18

18

UNIVERSITY OF

TORONTO

True-Triaxial Geophysical Imaging Cell One of the few True Triaxial

Rock Deformation Facilities with Integral Geophysical Imaging for laboratory

experiments and modelling of rock fracture: • • • •

3D geophysical measurements provide data to validate models 3D directional permeability measurements Coupled hydraulic, stress, and thermal conditions Laboratory simulation of the engineered subsurface environment of the Earth

ARMA Plenary Lecture, Chicago, June, 2012

True Triaxial Geophysical Imaging cell

19

UNIVERSITY OF

TORONTO

S om e C itations for T rue-T riaxial E xperim ents Mogi, K. (1966). Some precise measurements of fracture strength of rocks under uniform compressive strength, Rock Mech. Engin. Geology 4, 51-55 Mogi, K. (1971). Fracture and flow of rocks under high triaxial compression, J. Geophys. Res. 76, 1255-1269. Takahashi, M and Koide, H. (1989). Effect of intermediate principal stress on strength and deformation behaviour of sedimentary rocks at the depth shallower than 2000 m. In Rock at Great Depth (eds. V. Maury and D, fourmaintraux) (Balkema, Rotterdam), PP.19-26 Haimson, B. and Chang, C. (2000). A new true triaxial cell for testing mechanical properties of rocks, and its use to determine rock strength and deformability of Westerly granite, Int. J. Mech. Min. Sci.37, 285-296. King, M.S. (2002). Elastic wave propagation in and permeability for rocks with multiple parallel fractures. IJRMMS 39: 1033-1043. Haimson, B. (2006). True triaxial stresses and the brittle fracture of rock. Pure and Applied Geophysics, 163: 1101-1130. MQ, Y. (2009). True-triaxial strength criteria for rock. IJRMMS, 46, PP. 115-127. King, M.S., Pettitt, W.S., Haycox, J.R. & Young, R.P. (2011). Acoustic emission associated with the formation of fracture sets in sandstone under polyaxial stress conditions. Geophysical Prospecting. doi:10.111/j.1365-2478.2011.00959.x Young, R.P., Nasseri, M.H.B. and Lombos, L. (2012). Imaging the Effect of the Intermediate Principal Stress on Strength, Deformation and Transport Properties of Rocks Using Seismic Methods. In True Triaxial Testing of Rocks. Ed. Marek Kwasniewski , Xiaochun Li, Manabu Takahashi. CRC Press. ARMA Plenary Lecture, Chicago, June, 2012

20

UNIVERSITY OF

TORONTO

Research Objectives of Current Study 1.

Effect of intermediate stress on seismic and transport properties of rocks (here some results for Fontainebleau Sandstone)

2.

Evolution of velocity and acoustic emission for imaging fracture growth within a true triaxial system

3.

Permeability measurements along three independent orthogonal axes and application of effective medium theory and numerical methods

ARMA Plenary Lecture, Chicago, June, 2012

21

UNIVERSITY OF

TORONTO

Results: Effect of TT Stress on 3D deformation and Vp

σ3=10, σ2=20 MPa

ARMA Plenary Lecture, Chicago, June, 2012

σ3=10, σ2=50 MPa

22

UNIVERSITY OF

TORONTO

Imaging Velocity Structure σ3=2.5, σ2=50, σ1=550 MPa

σ3=10, σ2=10, σ1=10 MPa

σ3=10, σ2=50, σ1=500 MPa

3 2 σ3=10, σ2=50, σ1=100 σ3=10, σ2=50, σ1=50 MPa MPa

4400Plenary m/s Lecture, Chicago,4625 ARMA June, 2012

σ3=10, σ2=50, σ1=400 MPa σ3=10, σ2=50, σ1=300 MPa

4850

5073

530023

UNIVERSITY OF

TORONTO

Acoustic Emission Evolution (Induced Seismicity in the Lab)

ARMA Plenary Lecture, Chicago, June, 2012

24

UNIVERSITY OF

C1

TORONTO

1 3 2

3 2 3 ARMA Plenary Lecture, Chicago, June, 2012

25

UNIVERSITY OF

TORONTO

C2

1 3 2

3 2 3 ARMA Plenary Lecture, Chicago, June, 2012

26

UNIVERSITY OF

TORONTO

C3

1 3 2

3 2 3 ARMA Plenary Lecture, Chicago, June, 2012

27

C4

UNIVERSITY OF

TORONTO

1 3 2

3 2 3 ARMA Plenary Lecture, Chicago, June, 2012

28

C5

UNIVERSITY OF

TORONTO

1 3 2

3 2 3 ARMA Plenary Lecture, Chicago, June, 2012

29

UNIVERSITY OF

C1

C2

TORONTO

1

3

3

ARMA Plenary Lecture, Chicago, June, 2012

C1-5

C5

1 3

1

1

3

C4

C3

1 3

1 3 30

UNIVERSITY OF

C1

TORONTO

C4

ARMA Plenary Lecture, Chicago, June, 2012

C2

C5

C3

C1-5

31

UNIVERSITY OF

TORONTO

Fracture Pattern Under True-Triaxial Testing

σ3=10, σ2=50, σ1= 550 MPa

σ3 σ2

Side view parallel to σ2

Top view parallel to σ1 Polarized light

σ3 σ3

σ3

Deformed Pore space ARMA Plenary Lecture, Chicago, June, 2012

σ2

Epi-fluorescent

32

UNIVERSITY OF

TORONTO

AE Location and Failure Planes σ3=10, σ2=50, σ1=550 MPa

1 3 ARMA Plenary Lecture, Chicago, June, 2012

33

UNIVERSITY OF

TORONTO

3D Directional Permeability under Sealed Edges

ARMA Plenary Lecture, Chicago, June, 2012

34

UNIVERSITY OF

TORONTO

Evolution of 3D Permeability Experimental Data σ3=10, σ2=20 MPa

ARMA Plenary Lecture, Chicago, June, 2012

σ3=10, σ2=50 MPa

35

UNIVERSITY OF

TORONTO

3D Compressional and Shear wave velocities σ3=10, σ2=50 MPa

σ3=10, σ2=20 MPa

(a)

(b)

a) depicts variation of VP and VS1along three principal stress axes as a function of σ1 stress while σ2 and σ3 was kept at 20 and 10 MPa ,b) shows similar variation while σ2 and σ3 was kept at 50 and 10 MPa ARMA Plenary Lecture, Chicago, June, 2012

36

UNIVERSITY OF

TORONTO

P erm eability P rediction, S tatistical A pproach σ3=10, σ2=50 MPa 5E-15

kX predicted

4.5E-15

kY Predicted

4E-15

kZ Predicted

k, m 2

3.5E-15

kX Exp. data

3E-15

kY Exp. data

2.5E-15

kZ Exp.data

2E-15 1.5E-15 1E-15 5E-16 0 0

100

200

300

400

500

600

σ 1, MPa Crack density from inversion of seismic wave velocities ARMA Plenary Lecture, Chicago, June, 2012

Permeability After Gueguen and Dienes (89) 37

UNIVERSITY OF

TORONTO

Summary of True-Triaxial Results and Future work

• 3D velocity shows the effect of unequal stresses causing global compaction prior to fracture development when preferred tension cracks were formed parallel to the σ1 and σ2 plane. • Evolution of AE events confirm progressive development of such planes parallel to σ1 and σ2 plane. • 3D permeability is achievable within our TTT cell and satisfies Darcy’s law. • Inherent fabric orientation and initial 3D fracture networking was not the same between the two samples. • Increments of axial stress and compaction in the σ1 direction did influence the transport properties in the other two horizontal directions. 3D fracture networking affects lateral K values when axial stress is increased. • Predicted values using both methods (statistical and numerical) captured the overall trend of the 3D permeability variations as a function of axial stress increments. • Further development of platens underway for simultaneous resistivity, permeability and compliance. Initial tests are with machinable ceramics and non conductive coatings. • Future work on lab hydraulic fracturing under true-triaxial conditions to help validate Bonded Particle and Synthetic Rock numerical models for true-triaxial conditions at elevated temperatures. ARMA Plenary Lecture, Chicago, June, 2012

38

UNIVERSITY OF

TORONTO

Future Potential for Induced Seismicity Monitoring • Reservoir Stimulation via Hydraulic Fracturing

• Shale Gas • Mass Mining • Geothermal • CO2 Sequestration • Deep Geological Disposal of Radioactive Waste • Laboratory Investigations ARMA Plenary Lecture, Chicago, June, 2012

39

UNIVERSITY OF

TORONTO

Thank you

ARMA Plenary Lecture, Chicago, June, 2012

40