Performance assessment under multiple hazards D. Vamvatsikos , Dept - - PowerPoint PPT Presentation

Performance assessment under multiple hazards

• D. Vamvatsikos, Dept of Civil and Environmental Engineering, University of Cyprus, Cyprus
• E. Nigro, Dept of Structural Engineering, University of Naples “Federico II”, Naples, Italy

L.A. Kouris, G. Panagopoulos, A.J. Kappos, Dept of Civil Engineering, Aristotle University of Thessaloniki, Greece

• T. Rossetto & T.O. Lloyd, Dept of Civil, Environmental and Geomatic Engineering, University College

London, UK

• T. Stathopoulos, Dept of Building, Civil and Environmental Engineering, Concordia University, Canada

Introduction

• Vulnerability can be defined in multiple ways

– it can be evaluated using widely different formats that are typically inconsistent with each other, especially when considering different hazards!

• Emergence of multi-hazard assessment concepts, hence

– important to collectively discuss such methods – understand their merits – attempt to cast them in a format that is suitable for integration within a single practical assessment framework

• Here: review of vulnerability assessment methods for

– earthquake hazard – landslides and flowslides – tsunami – strong wind and hurricanes

• 1. Vulnerability of

structures to

earthquakes

Classification of the building stock

Code Structural system Storey number

MSt1-2 Stone masonry 1 2 MSt3+ 3+ MBr1-2 Brick masonry 1 2 MBr3+ 3+

Code Structural system Infills

Building height

• No. of

storeys

RC1L

Frame

No infills

Low-rise 1 to 3 RC1M Medium- rise 4 to 7 RC1H High-rise 8+ RC3.1L

Regularly infilled

Low-rise 1 to 3 RC3.1M Medium- rise 4 to 7 RC3.1H High-rise 8+ RC3.2L

Soft storey (pilotis)

Low-rise 1 to 3 RC3.2M Medium- rise 4 to 7 RC3.2H High-rise 8+ RC4.1L

Dual (walls+frames)

No infills

Low-rise 1 to 3 RC4.1M Medium- rise 4 to 7 RC4.1H High-rise 8+ RC4.2L

Regularly infilled

Low-rise 1 to 3 RC4.2M Medium- rise 4 to 7 RC4.2H High-rise 8+ RC4.3L

Soft storey (pilotis)

Low-rise 1 to 3 RC4.3M Medium- rise 4 to 7 RC4.3H High-rise 8+

RISK-UE system, as adapted by AUTh Team

• several systems available…
• need to converge!

Digital map (ArcMap) Building inventory (Εxcel)

UID

GIS

Example of compilation of inventory in Grevena loss assessment project (AUTh)

Damage definition

• six (5+1) damage states (DS0 to DS5)
• damage threshold different for R/C, URM …

Damage State Damage state label Range of damage factor Central damage factor (%) DS0 None DS1 Slight 0-1 0.5 DS2 Moderate 1-10 5 DS3 Substantial to heavy 10-30 20 DS4 Very heavy 30-60 45 DS5 Collapse 60-100 80

Ground motion characterisation

• Choice of a ground motion parameter that represents the

seismic demand is crucial. Possible choices:

macroseismic intensity based approaches (e.g. ATC-13)

• can be misleading!(rather subjective quantity, associated with

great uncertainty, dependent on building stock performance)

• but: (limited) available damage data is usually associated (only)

with intensity levels!

direct ground motion quantities, such as PGA or PGV

• r even

spectral quantities, like Sd (HAZUS) or Sa (Kiremidjian 1996)

• pertinent empirical data very scarce!

Determination of vulnerability functions

• Empirical approach [e.g. Spence et al. 2008; Rota et al. 2008]
• most common problem in purely empirical approach: lack of

(sufficient and reliable) statistical data for several intensities

• Judgement-based and rating methods
• expert opinion [e.g. ATC-13]

subjectivity ?...

• ‘scoring’ method (questionnaires) [e.g. GNDT]

scores=?...

• Analytical approach
• from (equiv.)SDOF to response-history of full structures!
• might seriously diverge from reality, usually overestimating loss!
• Hybrid approach
• combines empirical data with inelastic analysis (static/dynamic)
• critical point: definition of damage in each component!

The ‘primary’ vulnerability curve (evolution of damage with intensity of seismic action)

• Damage-state medians determined from analytical L – PGA

relationship, scaled based on statistical data available

e.g. DS4 (L=30%)

• lognormal distribution assumed for fragility analysis
• for given distribution type, only LDSi and β needed
• fragility curves for R/C and URM buildings developed by Kappos et al. using the

hybrid approach

PGA (g) P[ds>ds

PGA (g) P[ds>ds

typical fragility curves for R/C buildings typical fragility curves for URM buildings

The probabilistic vulnerability curve ( fragility curve)

Epistemic uncertainty

• 200 realizations of static

pushover capacity curves for a two story masonry building, caused by epistemic uncertainty [Vamvatsikos & Pantazopoulou 2010]

• Uncertainty is important (lack of knowledge of properties, coarse models)
• Recent methodologies remain computationally intensive and often difficult

to apply for practical purposes…

• 2. Vulnerability of structures subjected to

landslides and flowslides

Eruption of St. Helens volcano (1980) resulting in a catastrophic landslide Landslide triggered by the eruption

• f Stromboli in 2002

Landslides triggered by the 1949 Khait earthquake, Tajikistan: major bend in Khait landslide path

Building damage due to debris flow

breaking of brick or tuff external walls in R.C. framed structure height of debris flow

Building damage due to debris flow: Reinforced concrete buildings

Failure of corner column Breaking of brick or tuff external walls

Building damage due to debris flow: Reinforced concrete buildings (contnd.)

Plastic collapse mechanism of columns

Building damage due to debris flow: Masonry (URM) buildings

Masonry building impacted by debris flows Remaining parts of masonry buildings

Mechanical models for assessment of buildings

Type-A Mechanism Collapse of the tuff or brick external walls

2 2 2 2 2

2 1 2 cos cos 2 1 1 cos 2 cos

u u f u u

P p L L p C V V g p P g g V L L

• uncertainties in both the hydrodynamic and structural models!

Mechanical models for assessment of buildings – R/C

Type-B Mechanism Three-plastic-hinge collapse mechanism in reinforced concrete columns

2 2 2 2

16 16

u u f u u

M q L q C D V D V g q M g g V D L D

Mechanical models for assessment of buildings – R/C

Type-C Mechanism Two-plastic-hinges collapse mechanism in reinforced concrete columns

2 2 , 2

4 4 4

u u u u u i i i i

M q L q M g g V D L D M g V L D

Mechanical models for assessment of buildings – R/C

Type-D Mechanism Shear collapse mechanism in reinforced concrete columns

1 1

2 2 1 / 2 /

u u u u u

T q L T q L L L L L q g V D

Mechanical models for assessment of buildings - URM

Type-E Mechanism Debris flow impact against the ground floor walls of masonry buildings

2 2

4 1 3 1 2 1 cos cos

u uv uo k u uv uo

M L p L s p b p g g V p p

• fragility curves?...

• 3. Vulnerability to Strong Wind events -

Hurricanes

23

 Vulnerability of structures is assessed through:

• damage assessment
• field examinations of wind-structure interaction
• hurricane risk assessment from the insurance

perspective

24

Homes destroyed by the storm in Plaquemines, Parish, Louisiana

Image from NOAA

25

Highway 90 bridge from Biloxi, Mississippi to Ocean Springs lies in a twisted mass as result

• f catastrophic wind and storm surge from Hurricane Katrina

(photo from FEMA)

Damage Assessment

26

 Several studies assessed damage through on-site

• bservations

 Components that suffered excessive damage were

• roofs
• gable-end walls
• connections and sheathing

 Fully engineered buildings showed superior

performance over pre- or non-engineered buildings

Field Studies

27

 Valuable information has been produced by field (full-

scale) studies

 Monitoring of both

• wind characteristics and
• impact on structures

 Evaluating wind-induced envelope pressures and

structural response

Hurricane Risk Assessment

 Risk assessment involves several points of view (e.g.

engineering, economic etc.)

 Research aims at developing appropriate risk

assessment models influenced by various factors such as:

• weather data
• type of building
• occupancy
• construction method

• 4. Vulnerability to Tsunami

Tsunami vulnerability is still in its infancy Generation of tsunami vulnerability curves has been hampered by the rarity of events

• Leading to lack of knowledge on tsunami

behaviour near and on-shore

• Difficulty of current numerical models to

accurately reproduce velocity profiles onshore

Empirical tsunami vulnerability curves

• Majority of existing tsunami vulnerabilty curves are empirical.
• Some examples:

Peiris (2006):

• For low-rise URM houses in

Sri Lanka (SL) affected by the 2004 Indian Ocean Tsunami (n=8672, from 11 locations)

• SL census data post-

tsunami

• 3 damage states
• X-axis = submerged height

Empirical tsunami vulnerability curves

Examples of empirical tsunami vulnerabilty curves

Koshimura et al. (2009):

• Buildings in Banda Aceh affected

by Indian Ocean Tsunami (mix of low-rise timber and non- engineered RC)

• Based on damage interpreted

from pre- and post-satellite imagery

• 1 damage state (collapse)
• X-axis = inundation depth,

velocity, and hydrodynamic force, calculated numerically and very coarsely (e.g. assuming buildings as a roughness coefficient)

Spatial distribution of structural damage interpreted from post-tsunami satellite image

Tsunami Vulnerability - Gaps

• Lack of vulnerability curves for engineered buildings
• Lack of analytical vulnerability curves
• Few curves with x-axes other than water depth
• this is not representative of the impact forces and pressures that

are better represented by other parameters, e.g. velocity and momentum

• Lack of field measurements of onshore flows and tsunami

characteristics

• Difficulty in simulating tsunami onshore flow parameters

numerically and experimentally

• Accounting for interactions of flows with bathymetry, coastal

barriers etc. and variation of flow between buildings