EC8- Modelling and Analysis
Short Description
EC8...
Description
EUROCODE 8 Background and Applications Dissemination of information for training – Lisbon, 10-11 February 2011 Dissemination of information for training – Lisbon 10-11 February 211
Modelling and Analysis (Chapter 4 of EC8-1)
Peter Fajfar University of Ljubljana
M j K Maja Kreslin li University of Ljubljana
Accelerograms Dissemination of information for training – Lisbon 10-11 February 2011
2
Hysteretic behaviour Dissemination of information for training – Lisbon 10-11 February 2011
STEEL
MASONRY
3
REINFORCED CONCRETE
PRESTRESSED CONCRETE
“Philosophy” of seismic design Dissemination of information for training – Lisbon 10-11 February 2011
4
PERFORMANCE STATE
EART THQUA AKE
operational frequent q R=95 years 41% in 50 years design R=475 years 10% in i 50 years max.considered
safe
near collapse
Performance states Dissemination of information for training – Lisbon 10-11 February 2011
5
Dissemination of information for training – Lisbon 10-11 February 2011
6
Everything should be made as simple as possible possible, but not simpler Albert Einstein
Scope Dissemination of information for training – Lisbon 10-11 February 2011
EC8-1, Chapter 4, Overview and comments
4.2 Characteristics of earthquake resistant buildings 4.3 Structural analysis 4.4 Safetyy verifications
Test building
Modelling g Analysis
Code designed g versus old buildings g
7
Basic principles of conceptual design Dissemination of information for training – Lisbon 10-11 February 2011
Structural simplicity p y Uniformity, symmetry and redundancy Bi-directional resistance and stiffness Torsional resistance and stiffness Diaphragmatic behaviour at storey level Adequate foundation
8
L’Aquila 2009 Dissemination of information for training – Lisbon 10-11 February 2011
9
L’Aquila 2009 Dissemination of information for training – Lisbon 10-11 February 2011
10
L’Aquila 2009 Dissemination of information for training – Lisbon 10-11 February 2011
11
Kobe 1995 Dissemination of information for training – Lisbon 10-11 February 2011
Izmit 1999 12
Kobe 1995 Dissemination of information for training – Lisbon 10-11 February 2011
13
Chile 2010 Dissemination of information for training – Lisbon 10-11 February 2011
14
Kobe 2010 Dissemination of information for training – Lisbon 10-11 February 2011
15
L’Aquila 2009 Dissemination of information for training – Lisbon 10-11 February 2011
16
Montenegro 1979 Dissemination of information for training – Lisbon 10-11 February 2011
17
Kobe 1995 Dissemination of information for training – Lisbon 10-11 February 2011
18
Montenegro 1979 Dissemination of information for training – Lisbon 10-11 February 2011
19
Montenegro 1979 Dissemination of information for training – Lisbon 10-11 February 2011
20
Primary seismic members Dissemination of information for training – Lisbon 10-11 February 2011
Members considered as part of the structural system y that resists the seismic action,, modelled in the analysis for the seismic design situation and fully designed and detailed for earthquake resistance in accordance with the rules of EN 1998
21
Secondary seismic members Dissemination of information for training – Lisbon 10-11 February 2011
Members which are not considered as part of the seismic action resisting g system y and whose strength and stiffness against seismic actions is neglected
22
Structural (ir)regularity Dissemination of information for training – Lisbon 10-11 February 2011
23
Regularity Dissemination of information for training – Lisbon 10-11 February 2011
Regularity in plan
Symmetry Compact plan configuration Adequate in-plan stiffness of the floors Small in in-plan plan slenderness Adequate torsional stiffness
Regularity in elevation
No interruption of lateral load resisting systems in ele ation elevation No abrupt changes of stiffness, mass and overstrength Limitations of setbacks
24
Torsional flexibility Dissemination of information for training – Lisbon 10-11 February 2011
25
T Tx and/or T Ty rx ls and/or ry ls
Torsionally stiff
Torsionally flexible
Importance classes Dissemination of information for training – Lisbon 10-11 February 2011
Permanent loads “G”
self weight of the structure + 2 kN/m2
0.8 I =Variable – live loads “Q”
I
2 office building (category B) 2 kN/m = 1.0
I = 1.2
Vertical loads (G, Q) were distributed to the elements with regard to their effective area I = 1 1.4 4
26
Importance factor Dissemination of information for training – Lisbon 10-11 February 2011
27
Importance factor
Return period T (years)
0.8
230
1.0
475
1.2
780
1.3
1000
14 1.4
12 0 1250 (based on data for Slovenia)
Combination of loads (EC0) Dissemination of information for training – Lisbon 10-11 February 2011
G j 1
k,j
"+" P "+" A E,d "+"
28
i 1
2,i
Qk ,i
Permanent loads “G” Prestressing loads “P” Seismic loads “A” Variable – live loads “Q” Q (factor in EC1)
Determination of masses Dissemination of information for training – Lisbon 10-11 February 2011
G
29
kj
""
Ei 2i
Ei
Qki
Pseudo 3D model Dissemination of information for training – Lisbon 10-11 February 2011
30
Cracked sections Dissemination of information for training – Lisbon 10-11 February 2011
In concrete buildings, in composite steelconcrete buildings b ildi and d iin masonry buildings b ildi the h stiffness of the load bearing elements should take into account the effect of cracking (Secant stiffness to the initiation of yielding of the reinforcement). ) The elastic flexural and shear stiffness properties p p of concrete and masonry elements may be taken to be equal to one-half of the corresponding stiffness of the uncracked ncracked elements elements.
31
Cracked sections Dissemination of information for training – Lisbon 10-11 February 2011
32
Sae (g)
Sde (cm)
2.5
T T cr= =2 T T √2 R
20 2.0
N
100
1.5 50
1.0 0.5 0.0
0 0
TTN TTR 1 cr
2
3
T (s)
Accidental eccentricity Dissemination of information for training – Lisbon 10-11 February 2011
33
eaii = 0.05 0 05 Li Li is i th the flfloor-dimension di i perpendicular di l tto th the direction of the seismic action
Methods of analysis Dissemination of information for training – Lisbon 10-11 February 2011
LINEAR
b
NONLINEAR
34
STATIC a
DYNAMIC
Lateral force method
Modal response spectrum analysis
Nonlinear static (pushover) analysis
Nonlinear responsehistory analysis
Sa
T a
combined with response spectrum
b
combined bi d with ith behaviour b h i f t factor
Behaviour factor Dissemination of information for training – Lisbon 10-11 February 2011
Factor used for design purposes to reduce the forces obtained from a linear analysis, y , in order to account for the non-linear response of a structure, associated with the material, the structural system and the design procedures
35
Behaviour factor - background Dissemination of information for training – Lisbon 10-11 February 2011
36
F Fe
Rμ =
Rs =
Fy
R=
Fd
Fe D = =μ Fy Dy
Fy Fd Fe = Rμ Rs Fd
R
Dd Dy
D
D
≡
Eurocode 8 αu q , Rs > α1
Ductility classes Dissemination of information for training – Lisbon 10-11 February 2011
37
F Fe
Fyy2
DCM
Fy1
DCH
Dy1
Dy2
D
D
Behaviour factor Dissemination of information for training – Lisbon 10-11 February 2011
38
Overstrength Dissemination of information for training – Lisbon 10-11 February 2011
39
900
uF
Force
1F
0
0.00
Displacement
Overstrength g factor = u / 1
0.15
Montenegro 1979 Dissemination of information for training – Lisbon 10-11 February 2011
40
Kobe 1995 Dissemination of information for training – Lisbon 10-11 February 2011
41
Overstrength factor Dissemination of information for training – Lisbon 10-11 February 2011
Wall- or wall-equivalent dual systems
wall systems with only two uncoupled walls per horizontal direction: u / 1 = 1,0 other uncoupled wall systems: u / 1 =1,1 wall-equivalent dual, or coupled wall systems: u / 1 = 1,2
Irregular in plan: reduced values Pushover analysis: increased values
42
Lateral force method Dissemination of information for training – Lisbon 10-11 February 2011
43
Regular structures with small influence of higher modes
T1 ≤ 4 TC in T1 ≤ 2.0 s
Fb = Sd(T1) m λ
Lateral force method Dissemination of information for training – Lisbon 10-11 February 2011
44
•
Approximate formulas for the period T1
•
Distribution of horizontal forces s i mi Fi Fb sj mj
•
z i mi Fi Fb z j mj
or
Accidental eccentricity
δ = 1 + 1,2 (x/Le) xi Le
Approximate formulas for T1 Dissemination of information for training – Lisbon 10-11 February 2011
45
R l i h Rayleigh
T1 2
u u
2 j
mj
j
pj
j
j
Empirical formula T1 Ct H
3
4
Modal response spectrum analysis Dissemination of information for training – Lisbon 10-11 February 2011
46
Ti 2 Φi i S Ai 2 4
Displacements:
U i max Φi i S Di
Forces:
Fi M Φi i S ai Li i Mi Li ΦTi M s
M i ΦTi M Φ i
Number of modes Dissemination of information for training – Lisbon 10-11 February 2011
The response of all modes of vibration contributing g significantly g y to the g global response p shall be taken into account
the sum of the effective modal masses amounts to at least 90% of the total mass of the structure
allll modes d with ith effective ff ti modal d l masses greater t th than 5% of the total mass are taken into account
47
Effective masses Dissemination of information for training – Lisbon 10-11 February 2011
48
2 L mi i Mi
M i ΦTi M Φi
Li Φ M s T i
n
m
i 1
j 1
m i mj M
Combination of modal responses Dissemination of information for training – Lisbon 10-11 February 2011
EE = √ Σ EEi2
49
(SRSS)
if Tj ≤ 0.9 Ti Otherwise more accurate procedure procedure, such as
CQC
Accidental eccentricity Dissemination of information for training – Lisbon 10-11 February 2011
Accidental torsional effects M X , i FX , i 0.05 LY , i , M Y , i FY , i 0.05 LX , i
50
Pushover analysis Dissemination of information for training – Lisbon 10-11 February 2011
N2 method (basic)
Target displacement: Annex B (informative)
Extended N2
Higher mode effects in plan and elevation Complies with the EC8-3 requirement “4.4.4.5 P Procedure d ffor estimation ti ti off ttorsional i l and d hi higher h mode d effects”
51
Combination of effects of components Dissemination of information for training – Lisbon 10-11 February 2011
52
EEdx "+" " "0 0,30E 30EEdy 0,30EEdx "+" EEdy
SRSS
E1x E1x 1
Fx
Fx
E2x E1y E1x E2x Fy
Fy
Vertical seismic action Dissemination of information for training – Lisbon 10-11 February 2011
If avg is greater than 0,25 g (2,5 m/s2) the vertical component of the seismic action should be taken into account for horizontal or nearly horizontal structural members b spanning i 20 m or more for horizontal or nearly horizontal cantilever components longer than 5 m for horizontal or nearly horizontal pre-stressed components for beams supporting columns in base-isolated structures
53
Displacement calculation Dissemination of information for training – Lisbon 10-11 February 2011
54
d s = qd d e ds di displacement l t induced i d db by th the d design i seismic i i action ti qd behaviour factor for displacements (qd = q, unless otherwise specified) de displacement determined by a linear analysis based on the design response spectrum Upper limit: value from the elastic displacement spectrum Torsional effect are taken into account
Actual displacements Dissemination of information for training – Lisbon 10-11 February 2011
55
F Fe
q = Fe/Fd D = q Dd
Fy Fd
Dd Dy
D
D
Non-structural element Dissemination of information for training – Lisbon 10-11 February 2011
Architectural Architectural, mechanical or electrical element element, system and component which, whether due to lack of strength or to the way it is connected to the structure, is not considered in the seismic design as load carrying y g element
56
Non-structural elements Dissemination of information for training – Lisbon 10-11 February 2011
For non-structural elements of great importance or of a p particularly y dangerous g nature
Floor-response spectra
For other non-structural elements
Simplified procedure
57
Non-structural elements Dissemination of information for training – Lisbon 10-11 February 2011
Simplified analysis
Fa S a Wa a / q a Sa = S[3(1 + z/H) / (1 + (1 – Ta/T1)2)-0,5] Wa weight of the element γa importance factor for the element qa behaviour factor for the element
58
Floor acceleration spectrum (simplified) Dissemination of information for training – Lisbon 10-11 February 2011
59
Sa 6 agS 5 0 . . 0 0 1 = = = z Hz Hz H
5 4 3 2 1 0 0
1
2
3
4
Ta T1
Normalized floor acceleration spectrum ( a = 1, (q 1 a = 1, 1 z height h i ht up to t the th floor fl , H total t t l height, h i ht Ta period of the element, T1 period of the structure)
Additional measures for masonry infilled frames Dissemination of information for training – Lisbon 10-11 February 2011
Provisions apply to frame or frame equivalent dual concrete systems of DCH and to steel or steel-concrete composite moment resisting frames of DCH with interacting nonengineered masonry infills
Recommendation: adopt also for DCM or DCL concrete, concrete steel or composite structures with masonry infills
Irregularities I l iti iin elevation l ti Irregularities g in p plan Damage limitation of infills
60
Friuli 1976 Dissemination of information for training – Lisbon 10-11 February 2011
61
Montenegro 1979 Dissemination of information for training – Lisbon 10-11 February 2011
Izmit 1999 62
Safety verifications (1) - Ultimate limit state Dissemination of information for training – Lisbon 10-11 February 2011
Resistance condition Ed ≤ R d Ed demand, d d Rd capacity it
P- effects need not be taken into account if P d θ = tot r 0,10 V tot h
63
Safety verifications (2) Dissemination of information for training – Lisbon 10-11 February 2011
Global and local ductility condition
Specific material related requirements shall be satisfied satisfied, including, when indicated, capacity design provisions Prevention of storey mechanisms
M Rc 1,3 M Rb
64
Capacity design Dissemination of information for training – Lisbon 10-11 February 2011
YES !
65
NO !
Capacity design Dissemination of information for training – Lisbon 10-11 February 2011
66
Kobe 1995 Dissemination of information for training – Lisbon 10-11 February 2011
67
Kobe 1995 Dissemination of information for training – Lisbon 10-11 February 2011
68
Kobe 1995 Dissemination of information for training – Lisbon 10-11 February 2011
69
Safety verifications (3) Dissemination of information for training – Lisbon 10-11 February 2011
Equilibrium condition R i t Resistance off h horizontal i t l di diaphragms h Resistance of foundations Seismic joint condition
70
Damage limitation state Dissemination of information for training – Lisbon 10-11 February 2011
71
Limitation of interstorey y drift •
non-structural elements of brittle materials dr ≤ 0.005 0 005 h , dr ≤ 0.01h 0 01h
•
ductile non-structural elements dr ≤ 0.0075 h
•
non-structural elements do not to interfere with structural deformations, or without non-structural elements dr ≤ 0.010 h dr ≤ 0.02h
= 0.4 (importance classes III and IV) = 0.5 (importance classes I in II)
Return period versus (importance) factor Dissemination of information for training – Lisbon 10-11 February 2011
72
Return period T (years)
Return period T (years)
50
0.48
100
0.60
200
0.76
475
1.00
1000
1 30 1.30
10000
2.57 Valid for Slovenia
Test example Dissemination of information for training – Lisbon 10-11 February 2011
RC building g 6 stories + 2 basements
73
Description of building Dissemination of information for training – Lisbon 10-11 February 2011
74
SCHEMATIC SECTION
Description of building Dissemination of information for training – Lisbon 10-11 February 2011
75
TYPICAL PLAN
Description of building Dissemination of information for training – Lisbon 10-11 February 2011
76
BASEMENT
Seismic actions Dissemination of information for training – Lisbon 10-11 February 2011
ELASTIC RESPONSE SPECTRUM ag = I ·agR = 0.25g
importance class II (I = 1.0) 1 0)
Soil B, Type 1 S = 1.2, TB = 0.15 s,TC = 0.5 s, TD = 2.0 s
Damping 5%
77
Vertical actions Dissemination of information for training – Lisbon 10-11 February 2011
Permanent loads “G”
self weight of the structure + 2 kN/m2
Variable – live loads “Q”
office building (category B) 2 kN/m2
Vertical loads (G, Q) were distributed to the elements
78
Seismic masses (1) Dissemination of information for training – Lisbon 10-11 February 2011
79
Masses from permanent loads “G” factor 1.0 Masses from live loads “Q” Q factor Ei Ei 2 i
factor = 1.0 (roof storey), = 0.5 (other) factor 2i = 0 0.3 3 (category B) 15% (30%) mass from Q is taken into account
Seismic masses (2) Dissemination of information for training – Lisbon 10-11 February 2011
80
Level
Storey mass m (ton)
Moment of inertia MMI (ton*m2)
ROOF
372
33951
5
396
36128
4
396
36128
3
396
36128
2
396
36128
1
408
37244
=
2362 ton
* Only masses above level 0 are taken into account
l 2 b2 MMI m l m 12 2 s
Structural model – general (1) Dissemination of information for training – Lisbon 10-11 February 2011
3D (spatial) model All element are modelled as line elements • peripheral walls are modelled with line elements and a rigid beam at the top of the each element
Effective widths of beams (EC2) Rigid g offsets are not taken into account • Infinitely stiff elements are used only in relation to walls W1 and W2
Rigid diaphragms at each floor • slabs are not modelled
81
Structural model – general (2) Dissemination of information for training – Lisbon 10-11 February 2011
Masses and mass moments of inertia are lumped at centres of masses • Only masses above the top of the peripheral walls are taken into account
Cracked elements are considered • 0.5*As, 0.5*I, 0.1*It
All elements are fully fixed in foundation Infills are not considered
82
Structural model – general (3) Dissemination of information for training – Lisbon 10-11 February 2011
83
Structural model – effective width EC2 Dissemination of information for training – Lisbon 10-11 February 2011
84
Structural model – peripheral walls Dissemination of information for training – Lisbon 10-11 February 2011
85
Structural regularity Dissemination of information for training – Lisbon 10-11 February 2011
86
Criteria for regularity in elevation Criteria for regularity in plan 1) slenderness < 4
Lmax 4 Lmin
2) eccentricity < 30% 30%* torsional radius 3)) torsional radius < radius of gy gyration
Direction X: e0 X 0.30 r X Direction Y: e0Y 0 0.30 30 rY
Direction X:
rX ls
Direction Y:
rY l s
Structural regularity in plan Dissemination of information for training – Lisbon 10-11 February 2011
87
Structural eccentricity e0 and centre of stiffness
3 static load cases in each storey (FXi = 1, FYi = 1, Mi = 1) Loads are applied in centres of mass (CM) Determine rotation RZi due to FXi, FYi and Mi Determine e0i and centres of stiffness (XCRi, YCRi) e0 X ,i e0Y ,i
RZ ,i FX ,i 1 RZ ,i (Mi 1)
RZ ,i FY ,i 1 RZ ,i (Mi 1)
XCRi e0 X ,i XCMi
YCRi e0Y ,i YCMi
Structural regularity in plan Dissemination of information for training – Lisbon 10-11 February 2011
88
Torsional T i l radius di (r ( X, r Y)
3 static load cases in each storey (FTXi = 1, FTYi = 1, MTi = 1) L d are applied Loads li d iin centres t off stifness tif (CR) Determine rotations RZi (MTi), displacement UXi (FXi) and UYi (FYi) Determine torsional (KM,i) and lateral stiffnesses (KFX,i, KFY,i) Determine rXi and rYi K M ,i
1 1 1 , K FX ,i , K FY ,i RZ ,i MT ,i 1 U X ,i FTX ,i 1 UY ,i FTY ,i 1
rX ,i
K M ,i K FY ,i
and
rY ,i
K M ,i K FX ,i
Structural regularity - criteria Dissemination of information for training – Lisbon 10-11 February 2011
89
Criteria for regularity in elevation Criteria for regularity in plan 1) 2) 3))
Lmax 4 Lmin
Direction X: e0 X 0.30 r X Direction Y: e0Y 0.30 rY Direction X:
rX ls
Direction Y:
rY l s
Structure is regular in p plan and in elevation
Irregular in elevation if basement is also considered !?
Structural type of the building Dissemination of information for training – Lisbon 10-11 February 2011
UNCOUPLED WALL SYSTEM
The structural system is defined as a wall system system, when 65% (or more) of the shear resistance is contributed by walls Application of shear resistance is difficult
EC8 allows that shear resistance may be substituted by shear h fforces
Base (above basement) shear force taken by walls amounts to 72% (direction X) and 92% (direction Y) of the total shear force
Dual wall equivalent system?
90
Behaviour factor q Dissemination of information for training – Lisbon 10-11 February 2011
Structural St t l type: t uncoupled l d wall ll system t Ductility class: DCM
q0 3.0 Structural (ir)regularity: regular g in elevation - no reduction q0 Factor associated with prevailing failure mode: kw = 1 q kw q0 3.0
91
Periods, effective masses and modal shapes (1) Dissemination of information for training – Lisbon 10-11 February 2011
Meff,UX Meff,UY (%) (%)
92
Mode
T (sec)
1
0.92
80.2
0.0
0.2
2
0 68 0.68
00 0.0
76 3 76.3
00 0.0
3
0.51
0.2
0.0
75.2
4
0.22
15.0
0.0
0.2
5
0.15
0.0
18.5
0.0
6
0.12
0.2
0.0
17.6
Meff =
95 7 95.7
94 7 94.7
93 1 93.1
ETABS program
Meff,MZ (%)
Periods, effective masses and modal shapes (2) Dissemination of information for training – Lisbon 10-11 February 2011
1 MODE – predominantly translational in X direction 1.
93
Periods, effective masses and modal shapes (3) Dissemination of information for training – Lisbon 10-11 February 2011
2 MODE – translational in Y direction 2.
94
Periods, effective masses and modal shapes (4) Dissemination of information for training – Lisbon 10-11 February 2011
3 MODE – predominantly torsional 3.
95
Modal response spectrum analysis RSA Dissemination of information for training – Lisbon 10-11 February 2011
Modal response spectrum analysis was performed independently p y for the g ground excitation in two horizontal direction Combination of diferent modes – CQC Combination of results in two directions – SRSS Design spectrum was used Accidental eccentricity was taken into account Seismic S i i d design i situation it ti
96
Accidental torsional effects Dissemination of information for training – Lisbon 10-11 February 2011
Results of analysis without accidental torsion (SSRS p of accidental of two horizontal directions)) + envelope torsional effects
SRSS ((EX, EY) + ENVE(±M ( X, ± MY) Results of analysis without accidental torsion + accidental id t l ttorsional i l effects, ff t ffor each hh horizontal i t l direction. SRSS combination of two horizontal directions
SRSS (EX ± MX, EY ± MY)
97
RSA – Accidental torsional effects Dissemination of information for training – Lisbon 10-11 February 2011
98
RSA – shear forces Dissemination of information for training – Lisbon 10-11 February 2011
12% % of the total weight
99
15% of the total weight
RSA - displacements Dissemination of information for training – Lisbon 10-11 February 2011
100
RSA – Damage limitations Dissemination of information for training – Lisbon 10-11 February 2011
101
RSA – second order effects Dissemination of information for training – Lisbon 10-11 February 2011
102
Force distribution Dissemination of information for training – Lisbon 10-11 February 2011
103
Direction X
Lateral force method
Force distribution Dissemination of information for training – Lisbon 10-11 February 2011
104
Direction Y
Lateral force method
Shear forces Dissemination of information for training – Lisbon 10-11 February 2011
105
Direction X
Lateral force method
Shear forces Dissemination of information for training – Lisbon 10-11 February 2011
106
Direction Y
Lateral force method
Code designed versus old buildings Dissemination of information for training – Lisbon 10-11 February 2011
SPEAR BUILDING
107
Pushover curves Dissemination of information for training – Lisbon 10-11 February 2011
108
30
Test EC8 H
27
= 6.5
24 F b / W [[%]
21 18 15
12
= 3.2
9
Test EC8 H 1st yield of beam Test 1st yield of beam column NCyield of column 1st Design force NC
6 3 0 X direction
0.0
0.3
0.6
0.9
1.2 1.5 1.8 d n / H [%]
2.1
2.4
2.7
Determination of seismic capacity (NC) Dissemination of information for training – Lisbon 10-11 February 2011
Test EC8 H
109
Probability of “failure” Dissemination of information for training – Lisbon 10-11 February 2011
110
PGA475 = 0.25 0 25 g x 1.15 1 15 = 0 0.29 29 g PGAC = 0.25 g
(test building),
(seismic hazard map, map soil type C) PGAC = 0.77 g
PNC = 0.78 x 10-2
or 32% in 50 years
PNC = 2.67 x 10-4
or 1.3% in 50 years
(EC8 building)
(test building) (EC8 building)
Discussion of results Dissemination of information for training – Lisbon 10-11 February 2011
111
PGAC = 0.77 g “The The code is too conservative!? conservative!?” PNC,50 = 1.3 13% “The probability is too high!?” How high is the tolerable probability? How safe is safe enough?
View more...
Comments