Building Near Faults: Soil-Fault-Structure Interaction Nicolas K. - - PowerPoint PPT Presentation

Building Near Faults: Soil-Fault-Structure Interaction Nicolas K. Oettle, Ph.D., P.E.

Jonathan D. Bray, Ph.D., P.E. University of California, Berkeley Funding: National Science Foundation Grant No. 926473

Acknowledgements

Overview

Four Topics:

• Effects of Past Earthquakes
• 2011 Tohoku Earthquake
• Dynamic Modeling
• Mitigation Strategies
• Conclusions

:: 1906 San Francisco Earthquake (Lawson, 1908)

• Faults rupture up to several meters at the ground surface
• Which disturbs structures and the built environment
• This study addresses engineering in fault zones

:: Chi-Chi (Taiwan) Earthquake, 3 to 4.5 m of reverse fault slip (GEER)

Fault-Soil-Structure Interaction

Structural damage from earthquake fault rupture

Deformed Ground Surface

:: (Anastasopoulos et al., 2008)

Fault

Fault-Soil-Structure Problem

Bedrock fault ruptures through soil with overlying structure

Soil Ductility (Bray et al., 1994)

• Showed that soil failure strain is large controlling factor in

soil response

• Critical in estimating damage to earthen dams

Geotechnical Centrifuge Tests (Bransby et al., 2008)

Light Load: q = 37 kPa Heavy Load: q = 91 kPa

• Fault-soil-structure interaction in geotechnical centrifuge
• Showed importance of structure

Numerical Model of SFSI (Anastasopoulos et al., 2008)

• Numerical SFSI model for evaluating fault rupture in soil
• Advanced the state-of-the-art in simulation capabilities

Building in fault zones is still controversial

Project Motivation

• Study the fundamentals of boundary deformation problems
• Analyze dynamic effects of near-field fault slip rate
• Evaluate mitigation strategies for structures in fault zones

:: Darfield Earthquake (Quigley et al., 2011)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 30 35 40 45 50 55 60 Vertical Displacement at Surface (m) Original Horizontal Position (m) Ruptured Soil Unruptured Soil

Deformed Ground Surface Pulse Time History Structure Pulse Time History (Rigid Boundary)

• r

Applied Stress from 1-D Site Response (Deformable) Soil Bedrock Fixed Boundary Not to Scale

2011 Tohoku Earthquake Effects of Past Earthquakes Mitigation Strategies Dynamic Modeling

Collaborators: Jonathan Bray Paper: Oettle, N.K. and Bray, J.D. 2013, JGGE, 139:10, 1637–1647. Collaborators: Jonathan Bray, Keith Kelson, Kazuo Konagai Paper: Oettle, N.K. et al. 2013. 2013 Geo-Congress. Collaborators: Jonathan Bray Paper: Oettle, N.K. and Bray, J.D. 2013. JGGE, 139:11, 1864–1874. Collaborators: Jonathan Bray, Douglas Dreger Paper: Oettle, N.K., Bray, J.D., and Dreger, D.S. Submitted: Soil Dynamics and Earthquake Engineering.

Evolution of Surface Expression

• Observation: response localizes with increasing displacement
• Idea: Faults with prior seismicity could begin with localized

displacement

0.0 0.1 0.2 0.3 35 40 45 50 55 Incremental Displacement (m) Original Horizontal Position (m) 0-0.3 m 0.6-0.9 m 2.1-2.4 m

Soil Fault

Numerical Model

Shear Strain Mesh A numerical model was developed to study this effect

Elastic-perfectly plastic Friction interface

Structural Model

7-10 kPa/floor Model structures included three- and six-story steel moment frames attached to a reinforced-concrete mat foundation 10 m wide bays

Constitutive Model

Yield surface: Flow rule:

:: (Beaty, 2009)

Hardening law (hyperbolic):

Validation

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 20 30 40 50 60 70 Vertical Displacement at Surface (m) Original Horizontal Position (m) Dots - Centrifuge Test Solid Lines - Numerical Model 1.87 m vert. base offset 0.98 m vert. base offset 0.7 m vert. base offset

Numerical model was validated with centrifuge data

Boundary Deformation Induced Localization

Zone of High Stress Ratio Principal Stresses Fault Movement Surface Deformation

• Shear band formation propagates upward with

increasing displacement

• Initial K0 stress state altered to failure stress state

Shear Band Boundary Deformation

The Effect from Previous Ruptures

• Fault rupture may already be localized

– Weakened shear zone – Existing stress state

:: 1906 San Francisco Earthquake (Lawson, 1908)

Effect of Historical Seismicity

• Assumed continuation of prior earthquake
• More localized deformation field

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 30 35 40 45 50 55 60 Vertical Displacement at Surface (m) Original Horizontal Position (m) Ruptured Soil Unruptured Soil

Effect of Prior Ruptures on SSI

Without prior rupture:

• Broad deformation
• Fault splitting

With prior rupture:

• Localized displacement
• Foundation separation

Boundary Displacement Required for Localization

Compression

• Based on several numerical models
• Depends on soil height, failure strain, fault type

20 40 60 80 100 120 140 0% 5% 10% 15% Soil Thickness / Required Vertical Base Offset PS Compression (Loading) Failure Strain Solid - Reverse Fault Dashed - Normal Fault

Normal and Reverse Fault Stress Fields

Zone of High Stress Ratio Shear Band Principal Stresses Fault Movement Graben Plane Strain Compression Unloading Zone of High Stress Ratio Surface Deformation Shear Band Zone of High Stress Ratio Shear Band Principal Stresses Fault Movement Surface Deformation Plane Strain Extension Loading

• 150
• 100
• 50

50 100 50 100 150 200

t = (σ1-σ3)/2 (kPa) s = (σ1+σ3)/2 (kPa)

Reverse Fault Stress Path Normal Fault Stress Path Initial Stress Peak Stress Ratio Critical State Stress Ratio Stress Path Directions

Fundamentally different stress fields for normal and reverse faults

Stress Paths: Reverse: Normal:

Required Boundary Deformation

Controlled by field stress path failure strain

20 40 60 80 100 120 140 0% 5% 10% 15% Soil Thickness / Required Vertical Base Offset PS Compression (Loading) Failure Strain Solid - Reverse Fault Dashed - Normal Fault 20 40 60 80 100 120 140 0% 5% 10% 15% 20% Soil Thickness / Required Vertical Base Offset Stress Path-Dependent Failure Strain Solid - Reverse Fault Dashed - Normal Fault

• Developed the potential importance of prior fault ruptures
• Elucidated the correct mechanics of fault rupture

mechanisms in soil

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 30 35 40 45 50 55 60 Vertical Displacement at Surface (m) Original Horizontal Position (m) Ruptured Soil Unruptured Soil

Deformed Ground Surface Pulse Time History Structure Pulse Time History (Rigid Boundary)

• r

Applied Stress from 1-D Site Response (Deformable) Soil Bedrock Fixed Boundary Not to Scale

2011 Tohoku Earthquake Effects of Past Earthquakes Mitigation Strategies Dynamic Modeling

Collaborators: Jonathan Bray Paper: Oettle, N.K. and Bray, J.D. 2013, JGGE, 139:10, 1637–1647. Collaborators: Jonathan Bray, Keith Kelson, Kazuo Konagai Paper: Oettle, N.K. et al. 2013. 2013 Geo-Congress. Collaborators: Jonathan Bray Paper: Oettle, N.K. and Bray, J.D. 2013. JGGE, 139:11, 1864–1874. Collaborators: Jonathan Bray, Douglas Dreger Paper: Oettle, N.K., Bray, J.D., and Dreger, D.S. Submitted: Soil Dynamics and Earthquake Engineering.

2011 Tohoku Aftershock

Site Down Fault Ridge Up Three years ago, a fault ruptured through this ridge

Site Geology

The tertiary ridge overlies cretaceous bedrock

Field Deformation Measurements

:: (Karabacak et al., 2011)

Terrestrial LiDAR fault deformation on ridge by Prof. Kazuo Konagai of the University of Tokyo

Field Deformation Measurements

LiDAR measured 3D deformation field as determined from the

• riginal pool elevation

Conceptual Site Geometry

Pool/Gym Tertiary Ridge Sedimentary Rocks Cretaceous Abukuma Bedrock Not to Scale South 1.2 m

A tertiary ridge overlying cretaceous bedrock with the pool and gymnasium on the ridge, above the fault

Numerical Modeling

• At least 20 m of deformable media is necessary to deform

the ground surface this broadly

• Elasto-plastic analysis indicates 2% or higher axial failure

strain without prior ruptures matches the LiDAR data

• Subsequent geophysics confirm this model
• 1.2
• 1.0
• 0.8
• 0.6
• 0.4
• 0.2

0.0 35 45 55 65 Vertical Displacement (m) Original Horizontal Position (m)

2% (or higher) Failure Strain Needed 5 m 50 m 20 m

Improved Building Performance

:: (GEER, 2011)

• 20 m of deformable media changed surface expression of

the boundary deformation problem from localized to broad

• Ridge likely had no previous ruptures at this location
• Other areas with soft sediments on this fault ruptured

discretely and likely had prior ruptures

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 30 35 40 45 50 55 60 Vertical Displacement at Surface (m) Original Horizontal Position (m) Ruptured Soil Unruptured Soil

Deformed Ground Surface Pulse Time History Structure Pulse Time History (Rigid Boundary)

• r

Applied Stress from 1-D Site Response (Deformable) Soil Bedrock Fixed Boundary Not to Scale

2011 Tohoku Earthquake Effects of Past Earthquakes Mitigation Strategies Dynamic Modeling

Collaborators: Jonathan Bray Paper: Oettle, N.K. and Bray, J.D. 2013, JGGE, 139:10, 1637–1647. Collaborators: Jonathan Bray, Keith Kelson, Kazuo Konagai Paper: Oettle, N.K. et al. 2013. 2013 Geo-Congress. Collaborators: Jonathan Bray Paper: Oettle, N.K. and Bray, J.D. 2013. JGGE, 139:11, 1864–1874. Collaborators: Jonathan Bray, Douglas Dreger Paper: Oettle, N.K., Bray, J.D., and Dreger, D.S. Submitted: Soil Dynamics and Earthquake Engineering.

Dynamic Modeling

• Very near fault ground motions
• Based on numerical work by

Dreger et al. (2011)

• Slip rate 0.5 to 1.0 m/s

:: Dreger et al. (2011)

Deformed Ground Surface Pulse Time History Structure Pulse Time History (Rigid Boundary)

• r

Applied Stress from 1-D Site Response (Deformable) Soil Bedrock Fixed Boundary Not to Scale

• 300
• 200
• 100

100 200 300 Acceleration (cm/s2) This Study Dreger et al. (2011) 20 40 60 80 Velocity (cm/s) 10 20 30 40 17 18 19 20 21 Displacement (cm) Time (s)

Change in Soil Stresses

• Very unusual site response due to fling-type motions
• Could cause tension in soil if fast enough slip velocity

0.0 0.4 0.8 1.2 1.6 50 100 150 200 250 300 350 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Boundary Velocity (m/s) Stress (kPa) Time (s) Reverse Fault, 0.8 m/s σyy σxx Vboundary

Increase in Fault Diversion

• Free-field solution not changed considerably
• A dynamic analysis influences the amount of building

movement

0.0 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1 1.2 Movement of Structure as Fraction of Pseudo-Static Movement Slip Velocity (m/s) Pseudo-Static Displacement Dynamic Displacement Free-Field Boundary Condition

Pseudostatic: 0.8 m/s:

Structural Mass

Increasing the size of the structure decreases building movement

0.0 0.2 0.4 0.6 0.8 1.0 20 40 60 80 100 120 Movement of Structure as Fraction of Pseudo-Static Movement Mat Pressure (kPa) Pseudo-Static Displacement Dynamic Displacement

• Fault rupture dynamics can have a moderate affect on the

predicted soil-structure interaction

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 30 35 40 45 50 55 60 Vertical Displacement at Surface (m) Original Horizontal Position (m) Ruptured Soil Unruptured Soil

Deformed Ground Surface Pulse Time History Structure Pulse Time History (Rigid Boundary)

• r

Applied Stress from 1-D Site Response (Deformable) Soil Bedrock Fixed Boundary Not to Scale

2011 Tohoku Earthquake Effects of Past Earthquakes Mitigation Strategies Dynamic Modeling

Collaborators: Jonathan Bray Paper: Oettle, N.K. and Bray, J.D. 2013, JGGE, 139:10, 1637–1647. Collaborators: Jonathan Bray, Keith Kelson, Kazuo Konagai Paper: Oettle, N.K. et al. 2013. 2013 Geo-Congress. Collaborators: Jonathan Bray Paper: Oettle, N.K. and Bray, J.D. 2013. JGGE, 139:11, 1864–1874. Collaborators: Jonathan Bray, Douglas Dreger Paper: Oettle, N.K., Bray, J.D., and Dreger, D.S. Submitted: Soil Dynamics and Earthquake Engineering.

Design Strategies for Mitigation

1. Spread fault deformation 2. Design structure to move 3. Divert the fault

Design Strategies: Engineered Fill

Limited Shear Band Propagation Fault Movement Distributed Ground Deformation Spreading of Fault Movement Ductile Engineered Fill 500 1,000 1,500 2,000 2,500 0.3 0.6 0.9 Moment in 2nd Floor (kN·m) Vertical Fault Displacement (m) Stiff Previously Ruptured Native Soil More Ductile Engineered Fill Less Ductile Engineered Fill Yielding

• Engineered fill spreads fault deformation
• Note: first plastic behavior in structure typically bending in

beams at beam-column joints

Design Strategies: Mat Foundations

Thick mats support structure and limit distress to superstructure

Thinner Mat Columns Floor Beams Thicker Mat

Design Strategies: Mat Foundations

500 1,000 1,500 2,000 2,500 1 2 3 Moment in 2nd Floor (kN·m) Mat Thickness (m)

0 m 0.05 m 0.1 m 0.2 m 0.3 m 0.6 m

Reverse Fault Vertical Fault Displacement: Yield Moment

• Increasing mat thickness decreases structural loads
• Can decrease structural distress to elastic levels or

to life-safety levels

Design Strategies: Mat Foundations

Thick mats are resilient to complex fault deformation

Thick Mat Foundation Fault 2 Fault 1

Design Strategies: Fault Diversion

:: 1999 Kocaeli Earthquake (Lettis et al., 2000)

Heavy structures diverted fault rupture in Kocaeli

Design Strategies: Fault Diversion

Ground improvement diverts well located faults

Reverse Fault Soil Ground฀ Improvement Six-story Structure

Normal Fault Soil Diaphragm Wall Three-story Structure Tiebacks

Design Strategies: Fault Diversion

Structural walls can divert well located faults

Design Strategies: Fault Diversion

Reverse Fault Soil Anchors Three-story Structure

Ground anchors can hold a building to one side of a fault

Design Strategies: Fault Diversion

Fault Seismic Gap Structure Excavation

Gaps or soft ground can accommodate fault displacement Mitigation strategies can work and come in a variety of styles

Conclusions

• Historic seismicity affects soil-fault-structure interaction
• Can explain localized fault deformation observed
• Japan showed how soil can spread fault rupture
• Slip rate also affects the boundary deformation solution
• Safe structural design in fault zones is achievable
• Spread fault deformation
• Design structure to move
• Divert the fault