CPT Applications -Liquefaction 1
Peter K. Robertson
CPT in Geotechnical Practice
Santiago, Chile
July, 2014
What is Soil Liquefaction?
• Loss of strength and/or stiffness due toundrained loading
• Major factor in earthquakes
– Sand boils
– Ground cracks
– Slumping of embankments
– Lateral spreading
– Ground oscillations
• Flow slides
– Statically or dynamically triggered
Definitions of Liquefaction
• Cyclic (seismic)Liquefaction– Zero effective stress
(during cyclic loading)
• Flow (static)Liquefaction– Strain softening
response
– Part 2
Cyclic (seismic) Liquefaction
• Zero effective stress dueto undrained cyclicloading
• Shear stress reversal
– Level or gently slopingground
• Controlled by size andduration of cyclicloading
• Large deformationspossible
Moss Landing
Kobe
Cyclic Liquefaction – Lab Evidence
Zero stiffness
Liquefaction is not ‘yes’ or ‘no’ – it’s a gradual processwith increasing pore pressures and decreasing effective
stress with reducing soil stiffness.
Liquefaction evaluationAfter Seed et al 2003
Liquefaction is a multi-step process with liquefactioninduced deformations often the primary goal.
Mitigation is warranted when consequences threaten theproject not when triggering is likely
Simplified Procedure
• Following the 1964 earthquakes in Alaska andNiigata, Japan, the “Simplified Procedure” wasdeveloped by Seed & Idriss (1971) for evaluatingseismic demand and liquefaction resistance ofsands
• CSR7.5,1 = 0.65 (amax/g) (sv/s’v) rd / (MSF*Ksigma)
• CRR7.5,1 = fn of penetration resistance (SPT & CPT)
• NCEER Workshop (1996/7) to developconsensus update paper (Youd et al, 2001)
Cyclic liquefaction – design process
Sites defined as:
Level ground, gently sloping (< 5 degrees) or levelwith nearby steep slope or free-face
Sequence to evaluate (cyclic) liquefaction:
1. Evaluate susceptibility to cyclic liquefaction
2. Evaluate triggering of cyclic liquefaction
3. Evaluate post-earthquake deformations
Sand-like and Clay-like soils
Sand-like soils
• Fine-grained soils that are essentially non-plastic and behavevery similar to sands
• Cyclic resistance within framework based on in-situ tests
Clay-like soils
• Clays and plastic silts that are more easily sampled, are lessaffected by sampling disturbance and exhibit stress-historynormalized strength properties.
• Cyclic resistance estimated based on in-situ testing, laboratorytesting and empirical correlations based on undrained shearstrength
Susceptibility to cyclic liquefaction
Seed et al, 2003
Bray & Sancio, 2006
CPT SBT
Sand-like
Clay-like
Ic = 2.6(+/-)
Behavior Characteristics
Physical Characteristics
Transition from sand to clay-like behavior
Case history field observations
• Holocene age, clean sand deposits (Ko ~ 0.5)
• Level or gently sloping ground
• Corrected to magnitude M = 7.5 earthquake
• Depth range 1 to 12m
– 95% < 10m
• Surface observations
of liquefaction
• Average CPT values
Case history range of values• CPT tip resistance: 10 < qc < 120 atm (av = 52atm, stdev = 38)
• Friction ratio: 0.1% < Fr < 3% (av = 0.86%, stdev = 0.5%)
• SBT Index: 1.4 < Ic < 2.6 (av = 2.0, stdev = 0.28)
• Fines content: 0 < FC < 85% (most < 20% & mostly NP-low PI)
• EQ magnitude: 6.0 < Mw < 9.0 (av = 7.0)
• PGA: 0.09 < a(max) < 0.84g (av = 0.21g, stdev = 0.14g)
• Depth of liquefaction: 1 < zliq < 12m (av = 4.4m, stdev = 1.2m)
• Vert. eff. stress: 0.19 < s’vo < 1.5 atm (av=0.5 atm, stdev=0.24)
• Holocene-age, uncemented, silica-based soil (~NC)
• 27% ‘Non-Liq’ case out of 253 total
Origin of CPT-based methods
All methods have similar origins:
Case histories (each summarized to 1 data point)
– CSR7.5,s’=1 = 0.65 (amax/g) (sv/s’v) rd / MSF * Ks
– Normalization (qc1N) and ‘fines’ correction to getnormalized clean sand equivalent (qc1N,cs or Qtn,cs)
Each method made different assumptions for: rd, MSF, Ks,normalization of qc & ‘fines correction’
RW98 Moss06 IB08&14
Cyc
lic
Str
ess
Rat
io,(
CS
R)
Challenges with samplingProblem with corrections based on‘fines content’:• soil behaviour not controlled onlyby fines content• random sampling every 1.5 m (~5ft) – will get a limited view of actualsoil conditions
Better to select sampling depth basedon adjacent CPT
Updated database on SBTn chart
Data after Boulanger & Idriss, 2014
Ic = 2.6
All cases have CPT SBTnIc < 2.60
Robertson, 1990
Clean sand equivalent, Qtn,cs
Soils with same‘clean sandequivalent’Qtn,cs have
similarbehavior
Increased resistanceto loading
Same ‘clean sand equivalent’penetration resistance
- same in-situ state
Consequences of Liquefaction
• Post-earthquake settlement caused byreconsolidation of liquefied soils, plus possibleloss of ground (ejected) and localized shearinduced movements from adjacent footings, etc.
• Lateral spreading due to ground geometry
• Loss of shear strength, leading to instability ofslopes and embankments – strain softeningresponse – flow liquefaction
Factors that can affectliquefaction-induced deformations• Soil Characteristics
– relative density (state), fines content, plasticity, age, degree ofsaturation, cementation, prior stress & strain history, in-situ stress state,depositional environment
• Earthquake Characteristics– level and duration of shaking, frequency content of motions
• Site Characteristics– topography, geometry, stratigraphy, lateral variability, static
piezometric profile, hydraulic conductivity, pore water & void re-distribution
• Other factors– 3-D effects, ground cracking effects, low permeable surface layer, near
by foundations
Post Earthquake Deformations
• Estimating deformations in sandy soils isdifficult (even under static loading)
• For low to moderate risk projects semi-empirical models are common & appropriate
– CPT-based methods provide continuous profiles ofvolumetric & shear strain
• For high risk projects numerical analyses canbe appropriate, if initial screening indicates aneed
CPT-based post earthquakedeformations
(Zhang, Robertson & Brachman, 2002 & 2004)
• Based on extensive laboratory test results
– Ishihara and Yoshimine (1992)
• Links CPT and factor of safety to volumetric& shear strains for clean sands
• Apply CPT ‘clean sand equivalent’ approachto get profiles of volumetric & shear strainsand hence, displacements
• Calibrated with case histories
Evaluation of CPT Settlements• Zhang et al. (2000) showed:
– Good results when applied toMarina District (1998) andTreasure Island (1998) casehistories
– Importance of thin layers andtransition zones
– Importance of other factors (3-D, depth, proximity to footings,ejected soil, etc.)
• Lee et al. (2000) showed:
– Good results when applied toTaiwan (1999) case histories
Challenges estimating verticalsettlements
Liquefiedsoil
Liquefiedsoil
Liquefiedsoil
Liquefiedsoil
(a) (b) (c) (d)
Evaluation of Lateral SpreadApproach
• D. Chu & J. Stewart (2006) showed that Zhanget al (2004) CPT and Youd et al (2002) SPTmethods produced good results (slight over-predictions).
• Layers below base elevation of free face (z >2H) should not be included.
Transition zoneCPT data in‘transition’when
cone is moving from one soil typeto another when there is
significant difference in soilstiffness/strength (e.g. soft clay to
sand)
CPT data within transition zonewill be misinterpreted
In interlayered depositsthis can result in
excessive conservatism
Ahmadi & Robertson, 2005
Transitionzone
detection
Based on rate ofchange of Ic near
boundary of Ic = 2.60
Very important forliquefaction analysis
“CLiq” softwarewww.geologismiki.gr
Depth of liquefaction
Ishihara, 1985
Ishihara (1985) showed thatsurface damage from
liquefaction is influenced bythickness of liquefied layer and
thickness of non-liquefiedsurface layer.
Cetin et al (2009) proposedsimple weighting of vol. strainwith depth to produce similar
results
Summary
• CPT can provide estimates of:– Potential for cyclic liquefaction & softening (via CRR)– Post earthquake settlement and lateral displacement profiles
• Updates on:– Stress normalization– Extended to include clay-like soils
• No need for an Ic cut-off– Identification of transition zones
• Seismic CPT provides two independentmethods to evaluate soil response in same soil
Summary• Recommend removing transition zones
– CLiq provides auto feature to remove
• Recommend ‘weighting’ strains with depth
– CLiq provides ‘weighting’ feature
• Recommend sensitivity analysis to evaluatesensitivity of output (deformation) to mainvariables (e.g. EQ load, etc.)
• Often no single answer – requires somejudgment
Top Related