Workshop Implementation of the EC 8-3:2005 Assessment and interventions on buildings in earthquake prone areas
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Andreas J Kappos, Professor of Civil Engineering,
City University LondonCity University London
Athens, 12 April 2013
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Eurocode 8 – Part 3 Overview
Relatively recent document, intended to be performance- and displacement-based
‘Flexibility’ to accommodate the large variety of situations arising in y g y gpractice and in different countries
Arguably major advantage, also major weakness!..
Logically structured, but (on drafters’ own admission, see Pinto 2011) missing the support from extended use
improvements to be expected from future (and present?) experiencep
Normative part covering only material-independent concepts and rules; verification formulae are in non-mandatory Informative Annexes
Very limited application, mainly in academic/background studies
the GCSI has enjoyed much more extensive application
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Performance requirements in EC8 – 3
A pilot application of EC8-3. Reflections and comparisons with the GCSI
This presentation
Application of EC8-3 to a realistic building, representative of 1970’s European practice
the ‘SPEAR building’ (designed by Fardis et al.)g ( g y )
Application of all commonly used analysis methods for assessment
Lateral force (elastic) analysis
Multi-modal response spectrum analysis
q-factor approach
Application also of GCSI provisions, wherever different (m-method)
Identification of difficulties in implementing EC8-3
Comparisons with the GCSI
Conclusions and reflections on some issues and future trends
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Specimen tested within the frame of the SPEAR project
Key data for the building
3-storey reinforced concrete (R/C) building
Broadly based on Greek codes of 1954 and 1959
deemed representative of 1970’s European practice
Frame structure, no special seismic provisions and detailing
Intentional weak points:
irregular in plan, torsion problem
indirect beam supports
eccentric connections at some joints
Key data for the building (contnd)
BEAM 1
0,25 0,250,25 0,25
BEAM 2
COLUMNS C1-C5 & C7-C9
0,25
4Φ12 10Φ12COLUMN C6
3
C5 C12Φ12 (MONTAGE)
2Φ12
Φ8/0.20
5
Φ8/0.20
2Φ12 (MONTAGE)
2Φ12
C2
Φ12 Φ12 Φ12 Φ12
Eccentric connections
0,25
STIRRUPS Φ8/25 STIRRUPS Φ8/25
0,25
0,75
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Building model in ETABSin ETABS
Diploma thesisCh. Anastasiadou,
AUTh 2010
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Performance requirements and seismic actions
Significant Damage 10% in 50 years seismic action
Knowledge level: KL3 (full knowledge), not common in actual
buildings (as opposed to Lab specimens)
→ confidence factor: CFKL3=1
Seismic hazard zone: Ι (αg=0.16g) Seismic hazard zone: Ι (αg 0.16g)
Ground conditions: C (dense sand, gravel or stiff clay)
Importance class: ΙΙ (γΙ=1.00)
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Lateral force (elastic) analysis
Multi modal response spectrum analysisElastic response
spectrum
Assessment methods used
Multi-modal response spectrum analysis
q-factor approach
Seismic load combinations:
spectrum
Ex+0.3Ey
Ey+0.3Ex
Assumed stiffnesses: EIef=0.50EIg (EC8 §4.3.1 applies) important difference from GCSI (and ASCE 41-07)! ‘hidden’ in §A3.2.4(5) referring specifically to DL state (not
SD), that if deformations are verified, then EIef=MyLv/3θy
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Range of applicability (according to Eurocode 8 – Part 1):
a) period criteria
2 0s
2.4s4TT c
1
Lateral force (elastic) analysis
No restrictions in Eurocode 8 – Part 1
Additional requirement in EC8–3: ρi = Di / Ci < 2.5 for ductile members
b) Regularity in elevation Additional requirement in EC8–3: ρi = Di / Ci < 2.5 for ductile members
2.0s
Multi-modal response spectrum analysis
q ρi i / i
q= 1.5 for reinforced concrete buildings Verifications based on action effects derived by reduced (1/q) spectrum,
rather than on ρi
Elastic static analysis
q-factor approach
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Brittle components/mechanisms
ρi = Di / Ci < 1.0, Di from analysis (VEd)
Flexure and shear verifications
ρi i / i , i y ( Ed)
Ci based on mean values of material strengths (ΜRm)
ρi >1, Di from capacity design
MRd based on mean values of material strengths, modified for CFKL3 and γΜ
)1min( RcMMγM
Ductile components/mechanisms
Verification based on deformations!
),1min(
RbiRb,Rddi, M
MγM
A pilot application of EC8-3. Reflections and comparisons with the GCSI
• biaxial flexure is reduced to uniaxial one [Eurocode 2 Handbook]
Simplified column check for biaxial flexure (My, Mz, N)
z axis y axis z-axis y-axis
5bM
hM
xi
yi OR 0,2bM
hM
xi
yi Consider both
(i) 0 My
(ii) Mz 0
In all other cases:
if 1b'M
h'M
z
y
b'
Mβh'M z
z 0
if 1b'Mh'M
z
y
0 b'Mβh'
M yz
• the above My, Mz, are compared with corresponding (uniaxial) flexural strengths
z
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Flexural verification based on deformations (§A.3.2.1) Existing software packages do not provide chord rotations (θ) as output
how can θ be calculated from given joint rotations?
π+Ry2
Ry2
α12 Ry3
x1,z1
1 23
x1+ux1,z1+uz1
x2,z2 x3,z3
x2+ux2,z2+uz2
x3+ux3,z3+uz3
Ry2
Ry1
α23
chord rotation at end i: where:ijyii R
ziizjjziizjj uzuza
uzuza
arctantan
general case for calculating θ from
joint rotations Ry on plane x-z
xiizjjij
xiizjjij uxxx
auxxx
a
arctantan
Verification: checking of θ against fractions (dependent on PR) of
dc
ywsx f
f
vc
vum h
Lf
100
35,0225.0
el
25.125;01.0max
';01.0max3.0016.0
1
A pilot application of EC8-3. Reflections and comparisons with the GCSI
a) For internal beam-column joints:
Critical joint verification: Diagonal compression
) j
b) For external beam-column joints:
jcjd
cdcyds2s1Rdjhd hbηv1ηfVfAAγV
v
Mean values of material strengths, modified for CFKL3 and γΜ
jcjd
cdcyds1Rdjhd hbη
v1ηf80.0VfAγV
Ex+0.3Ey Ey+0.3Ex3
ey
3
ResultsResults: Distribution of ρi = Di/Ci (flexure)
2
store
ελαστική στατικήδυναμική φασματική
2
storey
ελαστική στατικήδυναμική φασματική
Lateral force
Dynamic spectral
Lateral force
Dynamic spectral
10 0.5 1 1.5 2
10.5 1 1.5 2
Lateral force Dynamic response spectrum
minρi = 0.015 minρi = 0.003maxρi = 18.37 maxρi = 11.18
Lateral force Dynamic response spectrum
minρi = 0.015 minρi = 0.001maxρi = 7.17 maxρi = 7.49
ρmax/ρmin>>2.5 (ρmin=1.002 for ρ>1) elastic methods not allowed!
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Summary of yrequired verifications
Lateral force (m-method) Dynamic response spectrum (m-method)
% of beam failures in flexure (Ex+0.3Ey)
60
80
100
60
80
100
q-approach
0
20
40
1st 2nd 3rd building0
20
40
1st 2nd 3rd building
80
100 ‘θ-method’ (θreq<3/4θu, based on EI 0 50EI )
0
20
40
60
80
1st 2nd 3rd building
on EIef=0.50EIg)
No failures!
60
80
100
60
80
100
% of beam failures in flexure (Ey+0.3Ex)
Lateral force (m-method) Dynamic response spectrum (m-method)
q-approach
80
100
0
20
40
1st 2nd 3rd building0
20
40
1st 2nd 3rd building
‘θ-method’ (θreq<3/4θu)
0
20
40
60
80
1st 2nd 3rd building
No failures!
% of column failures under M, N (Ex+0.3Ey)
Lateral force (m-method) Dynamic response spectrum (m-method)
60
80
100
60
80
100
q-approach
0
20
40
1st 2nd 3rd building0
20
40
1st 2nd 3rd building
80
100
80
100
‘θ-method’
0
20
40
60
80
1st 2nd 3rd building0
20
40
60
1st 2nd 3rd building
Lateral force Dynamic response spectrum
60
80
100
60
80
100
% of column failures under M, N (Ey+0.3Ex)
q-approach
0
20
40
1st 2nd 3rd building0
20
40
1st 2nd 3rd building
80
100
80
100‘θ-method’
0
20
40
60
1st 2nd 3rd building0
20
40
60
1st 2nd 3rd building
% of beam shear failures (Ex+0.3Ey) Lateral force Dynamic response spectrum
60
80
100
60
80
100
q-approach
0
20
40
1st 2nd 3rd building0
20
40
1st 2nd 3rd building
80
100
0
20
40
60
80
1st 2nd 3rd building
% of ‘failures’Beams in flexure
Comparative results for combination Ex+0.3Ey
% of ‘failures’Columns in flexure
40
60
80
100
lateral force analysis
response spectrum analysis
q-factor approach40
60
80
100
lateral force analysis
response spectrum analysis
q-factor approach
0
20
1st 2nd 3rd building0
20
1st 2nd 3rd building
(lateral force and dynamic analysis based on m-method)
% of ‘failures’Beams in shear
% of ‘failures’Columns in shear
Comparative results for combination Ex+0.3Ey
40
60
80
100
lateral force analysis
response spectrum analysis
q-factor approach 40
60
80
100
lateral force analysis
response spectrum analysis
q-factor approach
0
20
1st 2nd 3rd building0
20
1st 2nd 3rd building
(lateral force and dynamic analysis based on m-method)
Comparative results for combination Ey+0.3Ex
% of ‘failures’Beams in flexure
% of ‘failures’Columns in flexure
100
40
60
80
100
lateral force analysis
response spectrum analysis
q-factor approach 40
60
80
100
lateral force analysis
response spectrum analysis
q-factor approach
0
20
1st 2nd 3rd building0
20
1st 2nd 3rd building
(lateral force and dynamic analysis based on m-method)
% of ‘failures’Beams in shear
% of ‘failures’Columns in shear
Comparative results for combination Ey+0.3Ex
100 100
40
60
80
lateral force analysis
response spectrum analysis
q-factor approach 40
60
80
lateral force analysis
response spectrum analysis
q-factor approach
0
20
1st 2nd 3rd building0
20
1st 2nd 3rd building
% of joints failing in diagonal compression Ex+0.3Ey
Beam-column joint verifications
% of joints failing in diagonal compression Ey+0.3Ex
40
60
80
100
lateral force analysis
response spectrum analysis
q-factor approach40
60
80
100
lateral force analysis
response spectrum analysis
q-factor approach
0
20
1st 2nd 3rd building
q-factor approach
0
20
1st 2nd 3rd building
q factor approach
average D/C ratios for columns (Ex+0.3Ey)
Comparisons with the GCSI (θ vs. m method, lateral force)
average D/C ratios for beams (Ex+0.3Ey)
0 5
1.0
1.5
2.0
2.5
3.0
GCSI
EC8-3
0 5
1.0
1.5
2.0
2.5
3.0
GCSI
EC8-3
CGSI (and ASCE 41-06) m-method results in substantially more unfavourable results than EC8-3!
0.0
0.5
1st 2nd 3rd building0.0
0.5
1st 2nd 3rd building
average D/C ratios for columns (Ex+0.3Ey)
Comparisons with the GCSI (θ vs. m method, response spectrum)
average D/C ratios for beams (Ex+0.3Ey)
0 5
1.0
1.5
2.0
GCSI
EC8-30.4
0.6
0.8
1.0
GCSI
EC8-3
again, CGSI (and ASCE 41-06) m-method results in substantially more unfavourable results than EC8-3!
0.0
0.5
1st 2nd 3rd building0.0
0.2
1st 2nd 3rd building
Comparisons with the GCSI (θ vs. m method)
What are the reasons for the discrepancy?
e ifications sho ld be ca ied o t fo stiffnesses compatible ith verifications should be carried out for stiffnesses compatible with θu, θy important, since for old buildings EIef<<0.5EIg (here 0.5EIg was used in all cases, as allowed by §4.3 of EC8-3)
different fractions of allowable deformation are specified by each code (e.g. for SD, θd=3/4θu by EC8-3, =0.5(θu+θy)/γRd by GCSI)
in calculating m approximate LVL/2 is assumed for beams and columns whereas θ corresponds to the exact Lcolumns, whereas θ corresponds to the exact LV
m-factor used to reduce seismic action effects only, θ includes contribution from gravity loads as well
for same EIef and θd, the two procedures give the same outcome only in a cantilever, undeformed under gravity loads
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Analysis methods work as expected (same as in the GCSI) Dynamic response spectrum method leads to most favourablefavourable results q-approach leads to most unfavourableunfavourable results
Conclusions and reflections
q pp Realistic structures do not satisfy ρi<2.5 requirement for elastic analysis
(static+dynamic) need for inelastic analysis! (this also true for the GCSI) from GCSI based studies, pushover analysis more favourable results!
Preferable to use procedures prescribed in more detailed and comprehensive codes (like the GCSI)
Existing software packages do not provide chord rotations while θ Existing software packages do not provide chord rotations, while θ method (for flexure) is typically more favourable than m method!
Shear verification (capacity-based) not influenced by analysis method Assessment becomes complicated (and perhaps error-prone) due to
using different material strengths (with/without γM-factors) things equally (actually, a bit more) complex in the GCSI!
A pilot application of EC8-3. Reflections and comparisons with the GCSI
Thank you, Thank you, y ,y ,hope I have stirred hope I have stirred some discussion!some discussion!
websites: http://www.city.ac.uk/engineering-maths/staff/professor-andreas-kappos
http://ajkap.weebly.com/english.html
e-mail: [email protected]