Académique Documents
Professionnel Documents
Culture Documents
Foreword
These lecture notes provides an insight into the concepts, methods and procedures of
structural reliability and risk analysis considering the presence of random uncertainties. The
course is addressed to undergraduate students from Faculty of Engineering in Foreign
Languages instructed in English language as well as to postgraduate students in structural
engineering. The objectives of the courses are:
-
to provide the necessary background to carry out reliability based design and riskbased decision making and to apply the concepts and methods of performance-based
engineering;
The content of the lectures in Structural Reliability and Risk Analysis is:
-
Formulation of reliability concepts for structural components; exact solutions, firstorder reliability methods; reliability indices; basis for probabilistic design codes
Table of Contents
1. INTRODUCTION TO RANDOM VARIABLES THEORY................................................ 6
1.1 Data samples .................................................................................................................... 7
1.2 Indicators of the sample (esantion) .................................................................................. 7
1.3 Probability ........................................................................................................................ 9
1.4 Random variables ........................................................................................................... 10
1.5 Indicators of the probability distributions ...................................................................... 12
2. DISTRIBUTIONS OF PROBABILITY .............................................................................. 15
2.1. Normal distribution ....................................................................................................... 15
2.2. Log-normal distribution ................................................................................................ 19
2.3. Extreme value distributions........................................................................................... 21
2.3.1. Gumbel distribution for maxima in 1 year ............................................................. 22
2.3.2. Gumbel distribution for maxima in N years........................................................... 25
2.4. Mean recurrence interval............................................................................................... 27
3. FUNCTION OF RANDOM VARIABLES ......................................................................... 29
3.1 Second order moment models ........................................................................................ 29
4. STRUCTURAL RELIABILITY ANALYSIS ..................................................................... 33
4.1. The basic reliability problem......................................................................................... 33
4.2. Special case: normal random variables ......................................................................... 35
4.3. Special case: log-normal random variables................................................................... 36
4.4. Calibration of partial safety coefficients ....................................................................... 37
5. SEISMIC HAZARD ANALYSIS........................................................................................ 41
5.1. Deterministic seismic hazard analysis (DSHA) ............................................................ 41
5.2. Probabilistic seismic hazard analysis (PSHA) .............................................................. 42
5.3. Earthquake source characterization............................................................................... 43
5.4. Predictive relationships (attenuation relations) ............................................................. 46
5.5. Temporal uncertainty .................................................................................................... 46
5.6. Probability computations............................................................................................... 47
5.7. Probabilistic seismic hazard assessment for Bucharest from Vrancea seismic source . 47
6. ACTIONS ON STRUCTURES. SNOW LOADS ............................................................... 53
6.1 Introduction .................................................................................................................... 53
6.2 Snow load on the ground................................................................................................ 53
6.3 Snow load on the roof .................................................................................................... 56
6.4 Roof shape coefficients .................................................................................................. 58
6.5 Local effects ................................................................................................................... 60
6.5.1 Local verification .................................................................................................... 60
6.5.2 Exceptional snow drift on the roof .......................................................................... 61
7. ACTIONS ON STRUCTURES. WIND ACTION .............................................................. 62
7.1. General .......................................................................................................................... 62
7.2 Reference wind velocity and reference velocity pressure .............................................. 63
7.3 Terrain roughness and Variation of the mean wind with height .................................... 64
7.4 Wind turbulence ............................................................................................................. 67
7.5 Peak values ..................................................................................................................... 67
7.6 Wind actions................................................................................................................... 68
7.6.1 Wind pressure on surfaces....................................................................................... 69
7.6.2 Wind forces ............................................................................................................ 70
8. ACTIONS ON STRUCTURES. SEISMIC ACTION ......................................................... 71
8.1 Introduction .................................................................................................................... 71
8.2 Representation of the seismic action.............................................................................. 71
References:
1. Structural reliability and risk analysis, Lecture notes (prof.dr.ing. Radu Vacareanu)
http://www.utcb-ccba.ro/attachments/116_Structural%20Reliability%20Lecture%20Notes.pdf
2. Siguranta constructiilor (Dan Lungu & Dan Ghiocel)
3. EN 1991 Eurocode 1: Actions on structures. Part 1-3 General actions Snow Loads
4. EN 1991-1-4 - Eurocode 1: Actions on structures - Part 1-4: General actions - Wind
actions, CEN
5. CR 1-1-4/2012Cod de proiectare. Evaluarea aciunii vntului asupra construciilor.
6. CR 1-1-3/2012 Cod de proiectare. Evaluarea aciunii zpezii asupra construciilor.
7. P100-1/2013 Cod de proiectare seismica - Partea I Prevederi de proiectare pentru
cldiri.
Examination subjects:
1. Random variables, indicators of the sample, histograms
2. Probability, probability distributions function, cumulative density function, indicators
3. Normal distribution
4. Log-normal distribution
5. Extreme value distributions, Gumbel distribution for maxima in 1 year and in N years,
Mean recurrence interval
6. Function of random variables and second order moment models
7. Structural reliability analysis. Special cases
8. Calibration of partial safety coefficients
9. Seismic hazard analysis
10. Snow loads
11. Wind action
12. Seismic action
As
N
Bending
moment
As
Steel area
Compressive
force
Steps: Model (1) Imposed Loads (2) Structural Analysis (3) Sectional Analysis (4)
Design for some limits states (5).
Model (1) - Epistemic uncertainty
Loads (2) dead, wind, live loads are evaluated using the codes.
Random uncertainty (in the codes) - one way to take care, is the use of the safety factors (5%
less than or 5% greater than see the class of concrete)
Structural Analysis (3) - Epistemic uncertainty (linear elastic materials strains proportional
to stresses)
Sectional Analysis (4) - Random + Epistemic uncertainty
Design for some limits states (5) ultimate limit states (safety), serviceability limit states.
For power plants LOCA local lose of cooling agent
Uncertainties can be reduced to physical models:
- random variable (RV) dead load, material strength
- random process wind velocity, ground acceleration.
The random variable (RV) is a quantity whose value cannot be predicted with sufficient
accuracy before performing the experiment.
In engineering statistics one is concerned with methods for designing and evaluating
experiments to obtain information about practical problems, for example, the inspection of
quality of materials and products. The reason for the differences in the quality of products is
the variation due to numerous factors (in the material, workmanship) whose influence cannot
be predicted, so that the variation must be regarded as a random variation.
1.1 Data samples
In most cases the inspection of each item of the production is prohibitively expensive and
time-consuming. Hence instead of inspecting all the items just a few of them (a sample) are
inspected and from this inspection conclusions can be drawn about the totality (the
population).
If X = random variable (concrete compressive strength)
x = value for random variables
The mean value of a sample x1, x2, , xn or, briefly, sample mean, is denoted by x (or mx) and
is defined by the formula:
_
1 n
x xj
(1.1)
n j 1
It is the sum of all the sample values divided by the size n of the sample. Obviously, it
_
measures the average size of the sample values, and sometimes the term average is used for x .
The variance (dispersion) of a sample x1, x2, , xn or, briefly, sample variance, is denoted by
sx2 and is defined by the formula:
_
1 n
s x2
(
x
x
)2
(1.2)
j
n 1 j 1
The sample variance is the sum of the squares of the deviations of the sample values from the
_
mean x , divide by n-1. It measures the spread or dispersion of the sample values and is
always positive.
The square root of the sample variance s2 is called the standard deviation of the sample and is
denoted by sx. s x s x2 . The mean, mx and the standard deviation, sx has the same units.
The coefficient of variation of a sample x1, x2, ,xn is denoted by COV and is defined as the
ratio of the standard deviation of the sample to the sample mean
s
(dimensionless)
(1.3)
COV
x
380
340
390
340
350
370
360
360
370
330
370
400
360
350
360
380
350
400
360
420
350
320
360
360
350
340
390
The statistical relevance of the information contained in Table 1.1 can be revealed if one shall
order the data in ascending order in Table 1.2, first column (320, 330 and so on).
Table 1.2 Frequencies of values of random variable listed in Table 1.1
Compressive
strength
320
330
340
350
360
370
380
390
400
410
420
Absolute
frequency
2
1
3
6
7
4
2
2
2
0
1
Relative
frequency
0.067
0.033
0.100
0.200
0.233
0.133
0.067
0.067
0.067
0.000
0.033
Cumulative
frequency
2
3
6
12
19
23
25
27
29
29
30
Cumulative relative
frequency
0.067
0.100
0.200
0.400
0.633
0.767
0.833
0.900
0.967
0.967
1.000
The number of occurring figures from Table 1.1 is listed in the second column of Table 1.2. It
indicates how often the corresponding value x occurs in the sample and is called absolute
frequency of that value x in the sample.
Dividing it by the size n of the sample one obtains the relative frequency listed in the third
column of Table 1.2.
If for a certain value x one sums all the absolute frequencies corresponding to the sample
values which are smaller than or equal to that x, one obtains the cumulative frequency
corresponding to that x. This yields the values listed in column 4 of Table 1.2. Division by the
size n of the sample yields the cumulative relative frequency in column 5 of Table 1.2.
The graphical representation of the sample values is given by histograms of relative
frequencies and/or of cumulative relative frequencies (Figure 1.1 and Figure 1.2).
If a certain numerical value does not occur in the sample, its frequency is 0. If all the n values
of the sample are numerically equal, then this number has the frequency n and the relative
frequency is 1. Since these are the two extreme possible cases, one has:
- The relative frequency is at least equal to 0 and at most equal to 1;
- The sum of all relative frequencies in a sample equals 1.
If a sample consists of too many numerically different sample values, the process of grouping
may simplify the tabular and graphical representations, as follows (Kreyszig, 1979).
n1 n2
nj
nm
xmin
x x
xmax
A sample being given, one chooses an interval I that contains all the sample values. One
subdivides I into subintervals, which are called class intervals. The midpoints of these
subintervals are called class midpoints. The sample values in each such subinterval are said to
form a class.
The number of sample values in each such subinterval is called the corresponding class
frequency (absolute frequency - nj).
Division by the sample size n gives the relative class frequency (relative frequency m
n
f
f j j and f j = 1). The normalized relative frequency is f jN j . The normalization is
n
x
j 1
with respect to the x.
The relative frequency is called the frequency function of the grouped sample, and the
corresponding cumulative relative class frequency is called the distribution function of the
j
If one chooses few classes, the distribution of the grouped sample values becomes simpler but
a lot of information is lost, because the original sample values no longer appear explicitly.
When grouping the sample values the following rules should be obeyed (Kreyszig, 1979):
All the class intervals should have the same length;
The class intervals should be chosen so that the class midpoints correspond to simple number;
If a sample value xj coincides with the common point of two class intervals, one takes it
into the class interval that extends from xj to the right.
1.3 Probability
Probability is an numerical measure of the chance or likelihood of occurrence of an event
relative to other events.
Letting n to move to infinite (n ), frequencies moves to probabilities and consequently
fj probability. If x 0 then histogram of normalized relative frequencies become the
probability density function (PDF) and the histogram of cumulative relative frequencies
become the cumulative distribution function (CDF).
A random experiment or random observation is a process that has the following properties,
(Kreyszig, 1979):
it is performed according to a set of rules that determines the performance completely;
(1.4)
P(a X b) f (u )du
(1.5)
Hence this probability equals the area under the curve of the density f(x) between x=a and
x=b, as shown in Figure 1.3.
10
f(x)
P(a<X<b)
-4
-3
-2
0 a
-1
1b
- f ( u )du 1
(1.6)
If X is any random variable, then for any real number x there exists the probability P(X x)
corresponding to X x (X assumes any value smaller than x or equal to x) that is a function of
x, which is called the cumulative distribution function of X, CDF and is denoted by F(x).
Thus F(x) = P(X x).
Since for any a and b > a one has P(a X <b) = P(X b) - P(X a)
One shall now define and consider continuous random variables. A random variable X and the
corresponding distribution are said to be of continuous type or, briefly, continuous if the
corresponding distribution function F(x) = P(X x) can be represented by an integral in the
form
x
F ( x ) f ( u )du
(1.7)
where the integrand is continuous and is nonnegative. The integrand f is called the probability
density or, briefly, the density of the distribution. Differentiating one notices that
F(x) = f(x)
In this sense the density is the derivative of the distribution function.
Fx(x)
Shape of
CDF
Fx(a)
0
a
Properties of CDF:
11
- Fx(x) >0
- As x - Fx(x) 0
- As x Fx(x) 1
- if x2 x1 Fx(x2) Fx(x1) it is an increasing function
- Since for any a and b > a Fx(a) = P(X <a); Fx(b) = P(X <b) one has:
b
P( a X b ) f ( u )du F ( b ) F ( a )
a
(1.8a)
xf ( x )dx
(continuous distribution)
(1.8b)
where f(xj) is the probability function of discrete random variable X and f(x) is the density of
continuos random variable X.
The mean is also known as the mathematical expectation of X and is sometimes denoted by
E(X).
The variance of a distribution is denoted by x2 and is defined by the formula:
(1.9)
The positive square root of the variance is called the standard deviation and is denoted by x.
Roughly speaking, the variance is a measure of the spread or dispersion of the values, which
the corresponding random variable X can assume.
The coefficient of variation of a distribution is denoted by Vx and is defined by the formula
Vx
x
x
(1.10)
(1.12)
Skewness is a measure of the asymmetry of the probability distribution of a random variable.
A distribution is said to be symmetric with respect to a number x = c if for every real x,
f(c+x) = f(c-x).
(1.13).
12
Figure 1.4. Asymmetric distributions with positive asymmetry (left) and negative asymmetry
(right) (www.mathwave.com)
The mode of the distribution X is the value of the random variable that corresponds to the
peak of the distribution (the most likely value).
Distributions with positive asymmetry have the peak of the distribution shifted to the left
(mode smaller than mean); distributions with negative asymmetry have the peak of the
distribution shifted to the right (mode larger than mean).
~
The median of the distribution, X , xm is the value of the random variable that have 50%
chances of smaller values and, respectively 50% chances of larger values.
P(X > xm)= P(X xm) = 0.5 = Fx(xm)
For a symmetric distribution the mean, the mode and the median are coincident and the
skewness coefficient is equal to zero.
The fractile xp is defined as the value of the random variable X with p non-exceedance
probability (P(X xp) = p).
13
xp
F X (x )
xp
F X (x )
Cumulative distribution
function
Cumulative distribution
function
Probability distribution
function
Probability distribution
function
f x (x )
FX ( x p )
xp
xp
x
Inferior fractile
Superior fractile
14
2. DISTRIBUTIONS OF PROBABILITY
2.1. Normal distribution
The continuous distribution having the probability density function, PDF
1 x X
X
1
2
f ( x)
e
2 X
(2.1)
0.012
f(x)
Normal distribution
0.01
V=0.10
V=0.20
0.008
V=0.30
0.006
0.004
0.002
0
100
200
300
400
500
600
15
From (2.1) one notices that the normal distribution has the cumulative distribution function, CDF
F (x)
1
1
2 X
1 v X
2 X
dv
(2.2)
Figure 2.2 shows F(x) for same and various values of (and various values of coefficient of
variation V).
1
F(x)
0.9
Normal distribution
0.8
V=0.10
0.7
V=0.20
0.6
V=0.30
0.5
0.4
0.3
0.2
0.1
0
100
200
300
400
500
600
P(a X b) F (b) F (a )
dv
(2.3)
x
dv x
x x
integral (2.2) from - to z =
.
x
X x x x
Fx ( x) P ( X x) P (
( x ) /
u2
2 du
right equals:
16
1
( z )
2
u2
2
du
(2.4)
which is the distribution function of the standardized variable with mean 0 and variance 1 and
x x
has been tabulated and where z =
,but can be represented in terms of the integral.
x
x X
F ( x)
X
(2.5).
The density function and the distribution function of the normal distribution with mean 0 and
variance 1 are presented in Figure 2.3.
0.40
1
2
f(z)
Standard normal
distribution
0.35
0.30
0.25
0.20
0.15
0.10
0.05
z
0.00
-4
1.0
-3
-2
-1
(z)
Standard normal
distribution
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
-4
-3
-2
-1
Figure 2.3. PDF and CDF of the normal distribution with mean 0 and variance 1
1
f(z)
1
z2
e 2 -
17
a X
(2.6).
(2.7).
(2.8).
The meaning of kp becomes clear if one refers to the reduced standard variable z = (x - .
Thus, x = +z and kp represents the value of the reduced standard variable for which:
(z) = p.
Ex: x0,05
0.01
-2.326
0.02
-2.054
0.05
-1.645
0.95
1.645
0.98
2.054
0.99
2.326
18
2 ln X
ln x
1 ln v ln X
2 ln X
d (ln v )
2 ln X
1 ln v ln X
2 ln X
1
dv
v
(2.9).
Since:
x
F (ln x ) f ( v )dv
(2.10)
the probability density function PDF results from (2.9) and (2.10):
1
f(x)
2 ln X
1 ln x ln X
ln X
1 2
e
x
(2.11).
The lognormal distribution is asymmetric with positive asymmetry, i.e. the peak of the
distribution is shifted to the left. The skewness coefficient for lognormal distribution is:
1 3V X V X3
(2.12)
where VX is the coefficient of variation of random variable X. Higher the variability, higher
the shift of the lognormal distribution.
The mean and the standard deviation of the random variable lnX are related to the mean and
the standard deviation of the random variable X as follows:
ln X ln
X
1 V X2
ln X ln( 1 V X2 )
(2.13)
(2.14).
In the case of the lognormal distribution the following relation between the mean and the
median holds true:
ln X ln x0.5
(2.15).
Combining (2.13) and (2.15) it follows that the median of the lognormal distributions is
linked to the mean value by the following relation:
x 0 .5
X
1 V X2
(2.16)
19
ln X ln X
(2.17)
ln X V X
(2.18)
The PDF and the CDF of the random variable X are presented in Figure 2.4 for different
coefficients of variation.
0.012
f(x)
Log-normal distribution
0.01
V=0.10
V=0.20
0.008
V=0.30
0.006
0.004
0.002
0
100
200
300
400
500
600
F(x)
0.9
Log-normal distribution
0.8
0.7
V=0.10
V=0.20
0.6
V=0.30
0.5
0.4
0.3
0.2
0.1
0
100
200
300
400
500
600
Figure 2.4. Probability density function, f(x) and cumulative distribution function, F(x)
of the log-normal distribution for various values of V
The mode median mean of the distribution.
UTCB, Technical University of Civil Engineering, Bucharest
20
Change of variable:
If one uses the reduced variable
ln v ln v
du
1 1
u , then
or (dv = du v lnv,
ln v
dv ln v v
1
1
(z)
2 ln v
(ln v ln v ) / ln v
ln v ln v
. From (2.9) one obtains:
ln v
u2
2
1
1
ln v v du
v
2
u2
2
du
(2.19)
(2.20)
ln X k p ln X
(2.21)
where kp represents the value of the reduced standard variable for which (z) = p.
xm x0 ,5 e ln X ; ln x0.5 ln X relation between the median and the mean
b X
X
a X
ln b ln x
ln x
ln a ln x
ln x
21
0.009
f(x)
0.008
minima
distribution
maxima
distribution
0.007
0.006
all values
distribution
0.005
0.004
0.003
0.002
0.001
0
0
100
200
300
400
500
600
700
800
900
x
1000
Figure 2.5. Distribution of all values, of minima and of maxima of random variable X
2.3.1. Gumbel distribution for maxima in 1 year
The Gumbel distribution for maxima is defined by its cumulative distribution function, CDF:
F ( x ) e e
( x u )
(2.22)
n 1
n k 1 k
6 x
The final form is: u = x 0.45x mode of the distribution (Figure 2.8) and
= 1.282 / x dispersion coefficient (shape factor).
The skewness coefficient of Gumbel distribution is positive constant ( 1 1.139 ), i.e. the
distribution is shifted to the left. In Figure 2.6 it is represented the CDF of Gumbel
distribution for maxima for the random variable X with the same mean x and different
coefficients of variation Vx.
The probability distribution function, PDF is obtained straightforward from (2.22):
( x u )
dF ( x )
(2.23)
f(x)
e ( x u ) e e
dx
The PDF of Gumbel distribution for maxima for the random variable X with the same mean
x and different coefficients of variation Vx is represented in Figure 2.7.
22
Gumbel distribution
for maxima in 1 year
F(x)
0.9
V=0.10
0.8
V=0.20
V=0.30
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
100
200
300
400
500
600
700
Figure 2.6. CDF of Gumbel distribution for maxima for the random variable X
with the same mean x and different coefficients of variation Vx
0.014
f(x)
Gumbel distribution
for maxima in 1 year
0.012
V=0.10
V=0.20
0.01
V=0.30
0.008
0.006
0.004
0.002
x
0
100
200
300
400
500
600
700
Figure 2.7. PDF of Gumbel distribution for maxima for the random variable X
with the same mean x and different coefficients of variation Vx
One can notice in Figure 2.7 that higher the variability of the random variable, higher the shift
to the left of the PDF.
23
0.005
f(x)
0.0045
0.004
0.0035
0.003
0.0025
0.002
0.45
0.0015
0.001
0,164 x
0.0005
0
100
xm x
400
u
300
200
500
600
700
F ( x p ) P( X x p ) p e e
( x p u )
(2.24).
( x p u )
( x p u ) ln( ln p )
x
1
ln( ln p ) x 0.45 x
ln( ln p ) x k Gp x
1.282
(2.25)
where:
(2.26).
The values of kpG for different non-exceedance probabilities are given in Table 2.2.
Table 2.2. Values of kpG for different non-exceedance probabilities p
p
kpG
0.50
-0.164
0.90
1.305
0.95
1.866
0.98
2.593
24
(2.27)
where:
F(x)N years CDF of random variable X in N years
F(x)1 year CDF of random variable X in 1 year.
The Gumbel distribution for maxima has a very important property the reproducibility of
Gumbel distribution - i.e., if the annual maxima (in 1 year) follow a Gumbel distribution for
maxima then the maxima in N years will also follow a Gumbel distribution for maxima:
1( x u1 )
F ( x )N F ( x )1 N e e
ee
e Ne
1( x u1 )ln N
1( x u1 )
ee
1( x ( u1
ln N
))
1
ee
( x u N )
(2.28)
where:
u1 mode of the distribution in 1 year
dispersion coefficient in 1 year
uN = u1 + lnN / mode of the distribution in N years
dispersion coefficient in N years
The PDF of Gumbel distribution for maxima in N years is translated to the right with the amount
lnN / with respect to the PDF of Gumbel distribution for maxima in 1 year, Figure 2.9.
0.0063
f(x)
Gumbel distribution
for maxima
N yr.
1 yr.
0.0042
lnN /
0.0021
lnN /
0
100
200
u1
300
m1
400
500
uN
mN
600
700
800
900
x
1000
Figure 2.9. PDF of Gumbel distribution for maxima in 1 year and in N years
25
V x ,1
x
; V x , N x ; Vx,N < Vx,1 (COV is decreasing with increasing the years)
x ,1
x,N
Also, the CDF of Gumbel distribution for maxima in N years is translated to the right with the
amount lnN / with respect to the CDF of Gumbel distribution for maxima in 1 year, Figure 2.10.
1
F(x)
Gumbel distribution
for maxima
0.9
0.8
0.7
N yr.
1 yr.
0.6
0.5
0.4
0.3
0.2
0.1
0
100
200
300
400
500
600
700
800
900
x
1000
Figure 2.10. CDF of Gumbel distribution for maxima in 1 year and in N years
Important notice: the superior fractile xp (p >> 0.5) calculated with Gumbel distribution for
maxima in 1 year becomes a frequent value (sometimes even an inferior fractile if N is large,
N 50) if Gumbel distribution for maxima in N years is employed, Figure 2.11.
0.0063
f(x)
Gumbel distribution
for maxima
N yr.
1 yr.
0.0042
probability of
exceedance of x p
in N years
0.0021
probability of
exceedance of x p
in 1 year
0
100
200
300
400
xp
500
600
700
800
900
x
1000
26
years
The mean recurrence interval, MRI of a value larger than x of the random variable X may be
defined as follows:
MRI ( X x )
1
1
1
1
P1 year X x 1 P1 year X x 1 Fx x 1 p
(2.29)
where:
p is the annual probability of the event (Xx) and FX(x) is the cumulative distribution function of X.
Thus the mean recurrence interval of a value x is equal to the reciprocal of the annual
probability of exceedance of the value x. The mean recurrence interval or return period has an
inverse relationship with the probability that the event will be exceeded in any one year.
For example, a 10-year flood has a 0.1 or 10% chance of being exceeded in any one year and
a 50-year flood has a 0.02 (2%) chance of being exceeded in any one year. It is commonly
assumed that a 10-year earthquake will occur, on average, once every 10 years and that a 100year earthquake is so large that we expect it only to occur every 100 years. While this may be
statistically true over thousands of years, it is incorrect to think of the return period in this
way. The term return period is actually misleading. It does not necessarily mean that the
design earthquake of a 10 year return period will return every 10 years. It could, in fact, never
occur, or occur twice. This is why the term return period is gradually replaced by the term
recurrence interval. Researchers proposed to use the term return period in relation with the
effects and to use the term recurrence interval in relation with the causes.
The mean recurrence interval is often related with the exceedance probability in N years. The
relation among MRI, N and the exceedance probability in N years, Pexc,N is:
Pexceedance, N (>x) = 1 Pnon-exceedance, N (x) = 1 - [Pnon-exceedance, 1 (x)]N =1-pN
27
1
T
e
Pexceedance, N (>x) = 1 - 1
T ( x )
1
1
MRI ( X x p )
; MRI ( X x p ) = T ( X x p ) ;
1 P1year X x 1 p
(2.30)
1
50years
1 0,98
Usually the number of years, N is considered equal to the lifetime of ordinary buildings, i.e.
50 years. Table 2.3 shows the results of relation (2.30) for some particular cases considering
N=50 years.
Table 2.3 Correspondence amongst MRI, Pexc,1 year and Pexc,50 years
Mean recurrence interval,
years MRI, T ( x )
10
30
50
100
225
475
975
2475
Probability of exceedance
in 1 year, Pexc,1 year
Probability of exceedance
in 50 years, Pexc,50 years
0.10
0.03
0.02
0.01
0.004
0.002
0.001
0.0004
0.99
0.81
0.63
0.39
0.20
0.10
0.05
0.02
The modern earthquake resistant design codes consider the definition of the seismic hazard
level based on the probability of exceedance in 50 years. The seismic hazard due to ground
shaking is defined as horizontal peak ground acceleration, elastic acceleration response
spectra or acceleration time-histories. The level of seismic hazard is expressed by the mean
recurrence interval (mean return period) of the design horizontal peak ground acceleration or,
alternatively by the probability of exceedance of the design horizontal peak ground
acceleration in 50 years. Four levels of seismic hazard are considered in FEMA 356
Prestandard and Commentary for the Seismic Rehabilitation of Buildings, as given in Table
2.4. The correspondence between the mean recurrence interval and the probability of
exceedance in 50 years, based on Poisson assumption, is also given in Table 2.4.
Table 2.4. Correspondence between mean recurrence interval and probability of exceedance
in 50 years of design horizontal peak ground acceleration as in FEMA 356
Seismic Hazard
Mean recurrence interval
Probability of
Level
(years)
exceedance
SHL1
72
50% in 50 years
SHL2
225
20 % in 50 years
SHL3
475
10 % in 50 years
SHL4
2475
2 % in 50 years
28
and l, W are described deterministically, the design condition (ultimate limit state condition)
is:
M max M cap ;
ql 2
y W
8
and considering
S sectional effect of load, and
R sectional resistance
it follows that:
ql 2
; R y W .
8
If Y (X1, X2, Xi,.. Xn) is the function of random variables how to find fy, my , y
To answer the question, two cases are considered in the following:
1.
2.
29
For the random variable X one knows: the probability density function, PDF, the cumulative
distribution function, CDF, the mean, m and the standard deviation, . The unknowns are the
PDF, CDF, m and for the random variable Y.
If one applies the equal probability formula, Figure 3.2 (a linear transformation keeps the
distribution):
(3.1)
1
1
1 y b
y b
f X ( x) f X
; fY ( y ) f X ( x )
dy
a
a
a a
(3.2)
dx
y
y=ax+b
fY ( y ) dy
y dy
fY ( y ) ?
f X ( x)dx
x
x dx
f X (x)
mY
yf
( y )dy
( x)dx a
2
Y
( a x b) f
(y m
xf
( x)dx b
( x)dx a mX b
) fY ( y )dy (a x b a mX b) 2 f X dx
2
a 2 ( x m X ) 2 f X ( x )dx a 2 X2
Y a X
mY a mX b
VY
(3.3)
Y
a X
mY a mX b
30
1.
2.
f (a)
f (a)
( x a)
( x a) 2 .... Taylor series
1!
2!
Y
( X i m X i ) ... linear
i 1
i m
n
relation
mY Y ( mx1 , mx2 mxi .. mxn ) - First order approximation of mean
y2
(3.4)
Y 2
X i - First order approximation of the variance
i 1
i m
n
X
2
n y
y
y
y
2
X2
X2
Y2
Xn
1 x 2
2
i 1 xi
x1 m
xn m
m
2
Xi
m
(3.5)
Relations 3.4 and 3.5 are the basis for the so-called First Order Second Moment Models,
FOSMM.
Few examples of FOSMM are provided in the following:
Y X1 X 2
mY Y ( mx1 , mx2 ) = m X m X
1
y
y
X2
X2 12 X2 12 X2 X2 X2 ;
Y2
1
2
1
2
1
2
x1 m
x2 m
Y X2 1 X2 2
31
Y a X b
mY a m X b
Y2 a 2 X2 ; Y a X
Y X1 X 2
mY Y ( mx1 , mx2 ) = m X m X
1
y
y
X2
X2 x1 2m X2 x2 2m X2 m X2 X2 m X2 X2
1
2
1
2
2
1
1
2
x1 m
x2 m
2
Y
m X2 2 X2 m X2 1 X2
X2
X2
Y
1
2
1
VY
2 2 VX21 VX22
mY
mX mX
m X2
mX
1
32
(4.1a)
=P(R-S0)
(4.1b)
=P(R/S1)
(4.1c)
(4.1d)
or, in general
=P(G(R ,S)0)
(4.1e)
where G( ) is termed the limit state function and the probability of failure is identical with the
probability of limit state violation.
Re= P(R>S) - probability of reliability
Pf + Re=1
(R>S safe; R=S limit state; R<S failure)
Quite general density functions fR and fS for R and S respectively are shown in Figure 4.1
together with the joint (bivariate) density function fRS(r,s). For any infinitesimal element
(rs) the latter represents the probability that R takes on a value between r and r+r and S a
value between s and s+s as r and s each approach zero.
Figure 4.1. Joint density function fRS(r,s), marginal density functions fR(r) and fS(s)
and failure domain D, (Melchers, 1999)
UTCB, Technical University of Civil Engineering, Bucharest
33
In Figure 4.1, the Equations (4.1) are represented by the hatched failure domain D, so that the
probability of failure becomes:
Pf PR S 0 f RS r , s dr ds
(4.2).
D
sr
f R r f S s dr ds
F x f x dx
R
(4.3)
Relation (4.3) is also known as a convolution integral with meaning easily explained by
reference to Figure 4.2. FR(x) is the probability that Rx or the probability that the actual
resistance R of the member is less than some value x. Let this represent failure. The term fs(x)
represents the probability that the load effect S acting in the member has a value between x
and x+x as x 0. By considering all possible values of x, i.e. by taking the integral over all
x, the total probability of failure is obtained. This is also seen in Figure 4.3 where the density
functions fR(r) and fS(s) have been drawn along the same axis.
1 F R (x), f S (x)
0.9
0.8
0.7
R
S
0.6
0.5
0.4
0.3
0.2
P(R<x)
0.1
0
-4 -3 -2 -1
R=x
2 3
x
9 10 11
f R (x), f S (x)
amount of overlap
of f R( ) and fS()
rough indicator of pf
0.0042
R
S
0.0021
0
100
200
S
300
400
S500
k Rk
600
R
700
800
900
x
1000
34
Pf
1 F
( x) f R ( x)dx
(4.4)
which is simply the sum of the failure probabilities over all cases of resistance for which the
load effect exceeds the resistance.
4.2. Special case: normal random variables
For a few distributions of R and S it is possible to integrate the convolution integral (4.3)
analytically.
One notable example is when both are normal random variables with means R and S and
variances R2 and S2 respectively.
The safety margin Z=R-S (Cornell approach) follow an normal distribution with a mean and
variance given by :
Z = R - S
(4.5a)
Z2 = R2 + S2
Z R2 S2
(4.5b)
0 Z
Pf PR S 0 PZ 0 FZ ( z 0 )
Z
(4.6)
where ( ) is the normal standard cumulative distribution function (for the standard normal
variate with zero mean and unit variance).
0.035
f Z (z)
0.03
Z<0
0.025
Z>0
Failure
Safety
Failure domain
(z<0)
0.02
Safe domain
(z>0)
0.015
0.01
0.005
Pf
0 0
-60
z=r-s=0
Limit state
function
-50
-40
-30
-20
-10
10
20
z
35
The random variable Z = R - S is shown in Figure 4.4, in which the failure region Z 0 is
shown shaded. Using (4.5) and (4.6) it follows that (Cornell, 1969):
0
R
S
1
Pf
2
2
R S
where =
(4.7)
z
is defined as reliability (safety) index.
z
If either of the standard deviations R and S or both are increased, the term in square brackets
in (4.7) will become smaller and hence Pf will increase. Similarly, if the difference between the
mean of the load effect and the mean of the resistance is reduced, Pf increases. These
observations may be deduced also from Figure 4.3, taking the amount of overlap of fR( ) and fS()
as a rough indicator of Pf.
Pf
10-3
10-4
10-5
10-6
3.09
3.72
4.27
4.75
S
S
Z
ln
ln R ln
R
S
V R V R2 VS2
(4.8b)
R
1 VR2
ln R ;
ln R ln(1 VR2 ) VR ;
Relation (4.1d) then becomes
ln S ln
S
1 VS2
ln S
ln S ln(1 VS2 ) VS
0 Z
R
Pf P ln 0 PZ 0
S
(4.9)
where ( ) is the standard normal distribution function (zero mean and unit variance). The
R
random variable Z= ln is shown in Figure 4.5, in which the failure region Z 0 is shown
S
shaded.
Using (4.8) and (4.9) it follows that
36
0 ln
S
Pf
1
V 2 V 2
S
R
where =
(4.10)
z
is defined as reliability (safety) index,
z
R
S
ln
V R2 VS2
0.035
(4.11)
f Z (z)
0.03
Z<0
0.025
Z>0
Failure
Safety
0.02
0.015
0.01
0.005
Pf
0 0
-60
-50
-40
-30
-20
-10
10
20
z
R
S
ln
R
S
V R2 VS2
1 VR
37
ln R
S
VR VS
(4.12).
ln R
S
R e VR VS R e VR e VS R e VR S e VS (4.13)
VR VS
S
S
where e VR and e VS are called safety coefficients, SC with respect to the mean.
Rq and Sp are characteristics values (fractiles) of the random variables R respectively S.
Rq e
ln R k q ln R
ln R
Sp e
ln S k p ln S
ln S
k V
R e q R e
S p S e k p VS
Rq
( k )V
k q VR
R e
k p VS
S e
(VR VS )
k q VR
k p VS
k q VR
k p VS
( k )V
p
S
e
( kq )V
R
e
( k )V
q
p
R
S
Rq e
S p e
p
q
q e
( k q )VR
p e
( k p )VS
But one needs the SC with respect to the characteristic values of the loads and resistances, the
so-called partial safety coefficients, PSC.
kq= -1,645 (p=0.05)
kp= 1,645 (p=0.95)
kq= -2,054 (p=0.02)
kq= 2,054 (p=0.98)
To this aim, one defines the limit state function used in the design process:
Rdesign
S design
(4.14)
where 0.98 and 0.05 are called partial safety coefficients, PSC.
Assuming that S and R are log-normally distributed, one has:
ln R 1.645 ln R
ln S 2.054 ln S
R0.05 e
S 0.98 e
ln R
e 1.645VR R e 1.645VR
(4.15)
ln S
e 2.054VS S e 2.054VS
(4.16)
VR
R0.05 R e 1.645VR e
S 0.98 S e 2.054VS
e S e 1.645VR
e 2.054VS
(4.17)
38
( 1.645 )V
R
R0.05 e
S0.98 eS
0.98
0.05
0.05 e
( 1.645 )V R
0.98 e
( 2.054 )V S
(4.18)
(4.19)
The partial safety coefficients 0.05 and 0.98 as defined by Equations 4.19 and 4.20 depend on
the reliability index and the coefficients of variation for resistances and loads, respectively.
If the reliability index is increased, the partial safety coefficient for loads 0.98 increases
while the partial safety coefficient for resistance 0.05 decreases. The theory of partial safety
coefficients based on lognormal model is incorporated in the Romanian Code CR0-2013
named Cod de proiectare. Bazele proiectarii structurilor in constructii (Design Code. Basis
of Structural Design).
g
0.98
2.0
0.9
F0.05
1.8
0.8
1.6
1.4
0.7
1.2
VS
1.0
0
0.2
0.4
VR
0.6
0.6
0.05
0.1
0.15
0.2
Figure 4.6. Variation of partial safety coefficients with the coefficient of variation of load
effect (left) and of resistance (right)
Eurocode 1 (EN1-1990) Basis of Structural design
= 4,7 ultimate Limit states (ULS)
= 2,9 Serviceability Limit states (SLS)
Rd Ed (design condition)
Actions (F) : Permanent actions (G), Variable actions (Q), Accidental action (A)
Fk characteristic value (upper fractile)
Fd design value
f - partial safety coefficient applied to the action
Fd = f Fk
Ed = Sd E(Fd)
- design value of section effort of the load
E(Fd) sectional effect of the design action (load)
Sd - safety coefficient considering the errors involved in the model for calculation of the
sectional effort of the load
39
Ed = E( Sd f Fk) = E( F Fk)
Xk characteristic value of the strength of the material (inferior fractile)
Xd design value of the strength of the material
m - partial safety coefficient applied to the strength of the material
Xd =
1
m
Xk
1
1
1
1
R( X d ) R(
X k ) R(
Xk )
Rd
Rd m
M
Other representative values of variable actions used in the design, Figure 4.7 are:
the combination value, represented as a product of 0Qk used for the verification of
ultimate limit states and irreversible serviceability limit states;
the frequent value, represented as a product 1Qk, used for the verification of ultimate
limit states involving accidental actions and for verifications of reversible
serviceability limit states; this value is closed to the central value of the statistical
distribution;
15
Valoare caracteristica, Qk
10
20
20
15
Valoare
instantanee Q
10
25
0
0
0.01
0.02
Timp
0.03
0.04
fQ
10
20
30
40
50
60
70
j 1
i2
j 1
i 1
Ed = Gk , j P AEd 2 ,i Qk ,i
- seismic combination
40
1.E-01
PSHA - Bucharest
1.E-02
1.E-03
1.E-04
100
150
200
250
300
PGA , cm/s2
350
400
1
; y mean annual rate of ground exceedance of mean parameter (PGA-peak
y
ground acceleration, SA-spectral acceleration, peak ground displacement, etc)
The two methods of seismic hazard analysis are deterministic and probabilistic ones.
MRI ( y )
Site
R hypocentral
distance
M- Source (focus)
2. Selection of source-to-site distance parameters (either epicentral or hypocentral
distance) for each source zone;
3. Selection of controlling earthquake (expected to produce the strongest shaking at the
site); use predictive (attenuation) relations for computing ground motion produced at
the site by earthquakes of magnitudes given in step 1 occurring at each source zone;
41
4. Define the seismic hazard at the site in terms of peak ground acceleration PGA,
spectral acceleration SA, peak ground velocity PGV, etc.
All the steps are summarized in Figure 5.1.
Source 3
M3
Source 1
M1
R1
R3
Site
R2
Source 2
M2
Step 1
Step 2
M
Ground
3
Motion
Parameter, M
2
M
Y
Y1
Y
2
Y Y3
YN
R2
R3
Step 3
R1
Distance
Step 4
42
4. Uncertainties in earthquake location, size and ground motion prediction are combined
and the outcome is the probability that ground motion parameter will be exceeded
during a particular time period.
All the steps are summarized in Figure 5.2.
Source 3
M3
Source 1
M1
Average
rate,
log
3
R
Site
f
R
Source 2
M2
Magnitude
Step
1
Step
2
P(Y>y*
)
Ground
Motion
Parameter,
Y
Distance, R
Step
3
Parameter value, y*
Step
4
Figure 5.2. Steps in PSHA
43
Gutenberg & Richter (1944) organized the seismic data in California according to the number
of earthquakes that exceeded different magnitudes during a time period. The key parameter in
Gutenberg & Richters work was the mean annual rate of exceedance, M of an earthquake of
magnitude M which is equal to the number of exceedances of magnitude M divided by the
length of the period of time. The Gutenberg & Richter law is (Figure 5.3):
lg M = a - b M
(5.1)
NM
T
10a
1
b small
b
b large
(5.3)
M = e M
(5.4)
44
The original Gutenberg & Richter law (5.1) is unbounded in magnitude terms. This leads to
unreliable results especially at the higher end of the magnitude scale. In order to avoid this
inconsistency, the bounded recurrence law is used.
The bounded law is obtained and defined hereinafter.
The form (5.4) of Gutenberg & Richter law shows that the magnitudes follow an exponential
distribution. If the earthquakes smaller than a lower threshold magnitude Mmin are eliminated,
one gets (McGuire and Arabasz, 1990):
FM(M) = P[Mag. M M Mmin] = 1 - P[Mag. > M M Mmin] =
= 1
fM(M) =
M min
=1-
e M
M min
= 1 - e ( M M min )
(5.5)
dFM ( M )
= e ( M M min) .
dM
(5.6)
Mmin is the mean annual rate of earthquakes of magnitude M larger or equal than Mmin.
If both a lower threshold magnitude Mmin and a higher threshold magnitude Mmax are taken
into account, the probabilistic distribution of magnitudes can be obtained as follows (McGuire
and Arabasz, 1990); (Bounded Guttenberg-Richter law).
The cumulative distribution function must have the unity value for M = Mmax. This yields:
FM ( M )
1 e ( M M min )
=
(5.7)
FM (M ) = P[Mag. M Mmin M Mmax] =
FM ( M max ) 1 e ( M max M min )
dFM ( M )
e ( M M min )
=
.
f M (M ) =
dM
1 e ( M max M min )
(5.8)
M M min [1 FM ( M )] = M min
(5.9)
where M min = e M min is the mean annual rate of earthquakes of magnitude M larger or
equal than Mmin.
Finally one gets (McGuire and Arabasz, 1990):
M min
M = e
M = e
(5.10)
1 e ( M max M )
1 e ( M max M min )
45
PY y* | m, r 1 PY y* | m, r 1 FY y *
where F is the CDF of ground motion parameter, usually assumed lognormal.
(5.11)
Y
P(Y>y*|m,r)
y*
fY(y|m,r)
r
46
R hypocentral
distance
M- Source (focus)
The attenuation relation for subcrustal earthquakes (Mollas & Yamazaki, 1995):
ln PGA = c0 + c1 Mw + c2 lnR +c3R +c4 h +
where: PGA is peak ground acceleration at the site,
Mw- moment magnitude,
R - hypocentral distance to the site,
h - focal depth,
c0, c1, c2, c3, c4 - data dependent coefficients
- random variable with zero mean and standard deviation = ln PGA .
5.7. Probabilistic seismic hazard assessment for Bucharest from Vrancea seismic source
The Vrancea region, located when the Carpathians Mountains Arch bents, is the source of
subcrustal (60-170km) seismic activity, which affects more than 2/3 of the territory of
Romania and an important part of the territories of Republic of Moldova, Bulgaria and
Ukraine. According to the 20th century seismicity, the epicentral Vrancea area is confined to a
rectangle of 40x80km2 having the long axis oriented N45E and being centered at about 45.6o
Lat.N 26.6o and Long. E.
47
Two catalogues of earthquakes occurred on the territory of Romania were compiled, more or
less independently, by Radu (1974, 1980, 1995) and by Constantinescu and Marza (1980,
1995) Table 5.1. The Radus catalogues are more complete, even the majority of significant
events are also included in the Constantinescu and Marza catalogue. The magnitude in Radu
catalogue is the Gutenberg-Richter (1936) magnitude, MGR. The magnitude in Constantinescu
& Marza catalogue was the surface magnitude, MS. Tacitly, that magnitude was later
assimilated as Gutenberg-Richter magnitude (Marza, 1995).
Table 5.1. Catalogue of subcrustal Vrancea earthquakes (Mw 6.3 ) occurred during the 20th century
Date
1903 13 Sept
1904
6 Feb
1908 6 Oct
1912 25 May
1934 29 March
1939
5 Sept
Time
(GMT)
h:m:s
Lat. N
Long.
E
26.6
RADU Catalogue,
1994
h, km
>60
I0 MGR Mw
7 6.3 -
MARZA www.infp.ro
Catalogue, Catalogue,
1998
1980
I0
Ms
Mw
6.5
5.7
6.3
08:02:7
45.7
02:49:00
45.7
26.6
75
5.7
6.3
6.6
21:39:8
18:01:7
20:06:51
45.7
45.7
45.8
26.5
27.2
26.5
150
80
90
8
7
7
6.8
6.0
6.3
8
7
8
6.8
6.4
6.3
7.1
6.7
6.6
06:02:00
45.9
26.7
120
5.3
6.1
6.2
1940 22 Oct
06:37:00 45.8
26.4
122 7 / 8 6.5 - 7
6.2
6.5
1940 10 Nov
01:39:07 45.8
26.7
1501)
9 7.4 - 9
7.4
7.7
1945 7 Sept
15:48:26 45.9
26.5
75 7 / 8 6.5 - 7.5
6.5
6.8
1945 9 Dec
06:08:45 45.7
26.8
80
7 6.0 - 7
6.2
6.5
1948 29 May
04:48:55 45.8
26.5
130 6 / 7 5.8 - 6.5
6.0
6.3
1977 4 March 2)
19:22:15 45.34 26.30
109 8 / 9 7.2 7.5 9
7.2
7.4
1986 30 Aug
21:28:37 45.53 26.47
133
8 7.0 7.2 7.1
1990 30 May
10:40:06 45.82 26.90
91
8 6.7 7.0 6.9
1990 31 May
00:17:49 45.83 26.89
79
7 6.1 6.4 6.4
1)
Demetrescus original (1941) estimation: 150Km; Radus initial estimation (1974) was 133 km
2)
Main shock
Nov.10 , 1940 destruction in the epicentral area and in Moldavia, around 1000 deaths, the
largest RC building in Bucharest (Carlton) collapsed;
March 4, 1977 more than 1500 deaths, more than 11000 injured, more than 2 billion
$ losses, 31 buildings with more than 4 stories collapsed.
As a systematization requirement for seismic hazard assessment, usually it is recommended
the use of the moment magnitude, Mw. For Vrancea subcrustal events the relation between
two magnitudes can be simply obtained from recent events data given in Table 5.1:
Mw MGR+ 0.3
6.0 < MGR < 7.7
(5.15)
Even the available catalogues of Vrancea events were prepared using the Gutenberg-Richter
magnitude MGR, the recurrence-magnitude relationship was herein newly determined using the
moment magnitude Mw. The relationship is determined from Radus 20th century catalogue of
subcrustal magnitudes with threshold lower magnitude Mw=6.3.
The average number per year of Vrancea subcrustal earthquakes with magnitude equal to and
greater than Mw, as resulting also from Figure 5.5, is:
log n(Mw) = 3.76 - 0.73 Mw
(5.16)
48
0.1
0.01
1 e 1 .687(8 .1 M w )
n M w e8 .654 1 .687M w
1 e1 .687( 8. 1 6 .3 )
M w, max = 7.8
8.1
0.001
6.3
6.0
6.7
6.4
7.1
6.8
7.5
7.2
7.9
7.6
8.3
8.0
Moment magnitude, M w
Figure 5.5. Magnitude recurrence relation for the subcrustal Vrancea source (M w6.3)
The values of surface rupture area (SRA) and surface rupture length (SRL) from Wells and
Coppersmith (1994) equations for "strike slip" rupture were used to estimate maximum
credible Vrancea magnitude. According to Romanian geologists Sandulescu & Dinu, in
Vrancea subduction zone: SRL 150200 km, SRA<8000 km2. Based on this estimation, from
Table 5.2 one gets:
Mw,max= 8.1.
(5.17)
Table 5.2. Application of Wells and Coppersmith equations to the Vrancea source (mean values)
M Mw
Event
Experienced
SRA, km2
2)
If the source magnitude is limited by an upper bound magnitude Mw,max, the recurrence
relationship can be modified in order to satisfy the property of a probability distribution,
Equation 5.18:
n M w e
M w
1 e
1 e
M w ,max M w
M w ,max M w0
(5.18)
1 e 1.687 (8.1 M w )
1 e 1.687 (8.1 6.3)
(5.19)
49
In Eq.(5.18), the threshold lower magnitude is Mw0=6.3, the maximum credible magnitude of
the source is Mw,max=8.1, and = 3.76 ln10 = 8.654, = 0.73 ln10 =1.687.
The maximum credible magnitude of the source governs the prediction of source
magnitudes in the range of large recurrence intervals, where classical relationship (5.16) does
not apply, Table 5.3.
Table 5.3. Mean recurrence interval (MRI) of Vrancea magnitudes, (Mw)=1/n(Mw)
Date
GutenbergRichter
magnitude, MGR
10 Nov. 1940
7.4
4 March 1977
7.2
30 Aug. 1986
30 May 1990
7.0
6.7
Moment
MRI from Eq.
magnitude,
(5.18),
Mw
years
8.1
8.0
778
475
7.9
356
7.8
217
7.7
148
100
7.6
108
7.5
82
7.4
63
50
7.3
50
7.2
40
7.0
26
The depth of the Vrancea foci has a great influence on the experienced seismic intensity. The
damage intensity of the Vrancea strong earthquakes is the combined result of both magnitude
and location of the focus inside the earth.
The relationship between the magnitude of a destructive Vrancea earthquake (Mw6.3) and the
corresponding focal depth shows that higher the magnitude, deeper the focus:
ln h = - 0.866 + 2.846 lnMw - 0.18 P
(5.20)
where P is a binary variable: P=0 for the mean relationship and P=1.0 for mean minus one
standard deviation relationship.
The following model was selected for the analysis of attenuation (Mollas & Yamazaki, 1995):
ln PGA = c0 + c1 Mw + c2 lnR +c3R +c4 h +
(5.21)
where: PGA is peak ground acceleration at the site, Mw- moment magnitude, R - hypocentral
distance to the site, h - focal depth, c0, c1, c2, c3, c4 - data dependent coefficients and - random
variable with zero mean and standard deviation = ln PGA, Table 5.4. Details are given
elsewhere (Lungu et.al., 2000, Lungu et. al. 2001). To obtain the values of table 5.4 we used the
records from 1977, 1986 and 1990s earthquakes.
Table 5.4. Regression coefficients inferred for horizontal components of peak ground
acceleration during Vrancea subcrustal earthquakes, Equation (4.21)
c0
3.098
c1
1.053
c2
-1.000
c3
-0.0005
c4
-0.006
lnPGA
0.502
The application of the attenuation relation 5.21 for the Vrancea subcrustal earthquakes of
March 4, 1977, August 30, 1986 and May 30, 1990 is represented in Figures 5.6, 5.7 and 5.8.
50
500
Attenuation relation - March 4, 1977; Mw=7.5, h=109km
450
400
median
PGA, cm/s^2 ..
350
median+stdev
300
250
200
150
100
50
0
0
50
100
150
200
250
D, km
Figure 5.6. Attenuation relation applied for March 4, 1977 Vrancea subcrustal source
250
Attenuation relation - August 30 1986; Mw=7.2, h=133km
PGA, cm/s^2..
200
150
100
median
median+stdev
50
0
0
50
100
150
200
250
D, km
Figure 5.7. Attenuation relation applied for August 30, 1986 Vrancea subcrustal source
400
Attenuation relation - May 30, 1990; Mw=7.0, h=91km
350
PGA, cm/s^2..
300
250
200
150
median
median+stdev
100
50
0
0
50
100
150
200
250
D, km
Figure 5.8. Attenuation relation applied for May 30, 1990 Vrancea subcrustal source
51
For a given earthquake recurrence, the mean annual rate of exceedance of a particular value of
peak ground acceleration, PGA*, is calculated using the total probability theorem (Cornell,
1968, Kramer, 1996):
( PGA PGA*) M min P ( PGA PGA* | m, r ) f M (m) f R (r )dmdr
(5.22)
where:
(PGA>PGA*) mean annual rate of exceedance of PGA* ;
- Mmin is the mean annual rate of earthquakes of magnitude M larger or equal than Mmin;
- P(PGA>PGA*|m,r) probability of exceedance of PGA* given the occurrence of an
earthquake of magnitude m at source to site distance r. This probability is obtained from
attenuation relationship (4.21) assuming log-normal distribution for PGA;
- fM(m) probability density function for magnitude;
- fR(r) probability density function for source to site distance.
The probability density function for magnitude is obtained from Eq. (4.8) (Kramer, 1996).
The probability density function for source to site distance is considered, for the sake of
simplicity, uniform over the rectangle of 40x80km2 having the long axis oriented N45E and
being centered at about 45.6o Lat.N and 26.6o Long. E.
The mean annual rate of exceedance of PGA the hazard curve - for Bucharest site and
Vrancea seismic source is represented in Figure 5.9.
The hazard curve can be approximated by the form H ko a g k , where ag is peak ground
acceleration, and ko and k are constants depending on the site (in this case ko=1.176E-05,
k=3.0865).
1.E-01
PSHA - Bucharest
1.E-02
1.E-03
1.E-04
100
150
200
250
300
PGA , cm/s2
350
400
Figure 5.9. Hazard curve for Bucharest from Vrancea seismic source
52
53
meteorological data are statistically independent. The statistical analysis use annual maximum
values, which in this case is the maximum associated with winter snow. Although in some
geographic regions over a long period of time can identify certain trends in climate evolution,
the actual practice setting snow load on the ground they do not take into account.
Romanian law CR 1-1-3/2012, is following the SR EN 1991-1-3, and establish that
characteristic value of snow load on the ground is defined as the snow load on the ground
based on an annual probability of exceedence of 0,02, i.e. IMR=50 years, which is an upper
fractile value of a random variable whose values are measured as annual maximum, using the
Gumbel probability distribution for maxima.
The Gumbel probability distribution for maxima was the EN 1991-1-3 recommendation
because after analyzing the data available at European level when drawing up standard (2003),
this probability distribution proved to be most suitable modeling snow in most European
countries (Switerland, Italy, Grece, Norway, Sueden, Finland, Island, Germany, France, Great
britain) weather stations.
The characteristic values of snow load on the ground in Romania, sk, are indicated in the
zoning map in Figure 6.1. The values shown are valid for designing the action for altitudes of
the sites A 1000 m. The values in Figure 6.1 are minimum values used in the design action
of snow load.
Figure 6.1 Characteristic value of snow load on the ground zoning map, sk, kN/m2, for
altitudide of the sites A 1000 m
The altitude of the site is height above mean sea level, of the site where the structure is to be
located, or is already located for an existing structure.
For the altitude of the site 1000m < A 1500m, the characteristic value of snow load on the
ground is:
sk(1000m < A 1500m) = 2,0 + 0,00691 (A-1000)
(6.1)
54
(6.2)
Ground level snow loads for any mean recurrence interval different to that for the
characteristic snow load, sk, (which by definition is based on annual probability of exceedence
of 0,02) may be adjusted to correspond to characteristic values by application of relation 6.3.
ln( ln p )
1 0,45
V1
1,282
sp
sk
1 2,593 V1
(6.3)
where:
sk is the characteristic snow load on the ground (kN/m2), with a return period of 50 yearsannual probability of exceedence of 0,02 (annual probability of nonexceedence p=0.98) - in
accordance with EN 1990:2002
sp is the snow load on the ground having the p annual probability of nonexceedence
V1 is the coefficient of variation of annual maximum snow load (in Romania V1 is 0,351,0).
Figure 6.2 are exemplified ratios of the snow load on the ground having MRI = 75years,
respectively MRI = 100years and with characteristic snow load on the ground (MRI=50
years) for different values of the coefficient of variation V1.
Raportul sIMR/sIMR=50ani
1.2
sIMR=100ani/sIMR=50ani
1.15
1.1
sIMR=75ani/sIMR=50ani
1.05
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Coeficientul de variatie V1
Figure 6.2 Ratios of snow load on the ground having MRI = 75years and MRI = 100 years
and with characteristic snow load on the ground (MRI=50 years)
The exceptional snow load on the ground is the load of the snow layer on the ground
resulting from a snow fall which has an exceptionally infrequent likelihood of occurring due
to the temperate climate is not applicable for Romania. If is the case, the EN 1991-1-3
recommended to double value of the sk.
The bulk weight density of snow varies. In general it increases with the duration of the snow
cover and depends on the site location, climate and altitude. Indicative values for the mean
bulk weight density of snow on the ground given in Table 6.1 may be used.
55
(6.2)
56
where:
Isis the importance/exposure coefficient for the snow load;
i is the snow load shape coefficient;
sk is the characteristic value of snow load on the ground
Ce is the exposure coefficient
C t is the thermal coefficient.
The snow loads on roofs for the accidental design situations where exceptional snow drift is
the accidental action shall be determined as follows:
s = Is i sk
(6.3)
where:
Isis the importance/exposure coefficient for the snow load;
i is the snow load shape coefficient;
sk is the characteristic value of snow load on the ground
The exposure coefficient Ce should be used for determining the snow load on the roof. The
choice for Ce should consider the future development around the site. Ce should be taken as
1,0 unless otherwise specified for different topographies. He characterizes the overall effect of
the wind on snow deposit on the building by the surrounding topography and natural
environment and / or built in the vicinity of the building.
Table 6.2 Values of Ce for different topographies
Topographies
Windswept
Ce
0,8
Normal
1,0
Sheltered
1,2
The thermal coefficient C t, is the coefficient defining the reduction of snow load on roofs as
a function of the heat flux through the roof, causing snow melting. The thermal coefficient C t
57
should be used to account for the reduction of snow loads on roofs with high thermal
transmittance (> 1 W/m2K) for some glass covered roofs, because of melting caused by heat
loss. Based on the thermal insulating properties of the material and the shape of the
construction work the thermal coefficient value is determined by special studies, see ISO
4355 - 1998. Ct should be taken as 1,0.
The roughness of the roof surface influences the snow sliding and is difficult to be evaluated.
For example, in some areas of the roof, the small objects might prevent natural snow slide
(other than parapets). However, since it is considered that the snow slides off the roof entirely
(when there are no obstacles or parapets) if the roof angle above 60, those shape coefficients
are zero for these parts of roofs.
6.4 Roof shape coefficients
The roof snow load shape coefficient is ratio of the snow load on the roof to the undrifted
snow load on the ground, without the influence of exposure and thermal effects.
European values of coefficients of EN 1991-1-3 were calibrated based on the analysis results
of experimental studies both on site (in-situ) and in the wind tunnel and the comparative
analysis of coefficients prescriptions from different countries, see Formichi, P., 2008. EN
1991 Eurocode 1: Actions on structures. Part 1-3 General actions Snow Loads,
presentation at Workshop Eurocodes. Background and Applications, 18-20 Feb., Brussels,
60p.
The results of in-situ measurements in the US, Canada, Norway and England were
completed with the results of measurements conducted special campaigns in Europe for
studying deposits of snow on roofs (winter 1998/1999). The measurements were very
detailed, both in terms of meteorological parameters (wind speed, wind direction, air
temperature, air humidity, solar radiation, rainfall regime, etc.) and in terms of roof types
(shape dimensions, inclinations, the surface heat transfer from inside the building, roof
insulation, etc.), altitude, exposure (wind, sun), drifts of snow on the roof at various points,
etc. In England measurements were performed on 25 types of roofs in 18 different sites at
altitudes from 5m to 656 m. In the Italian Alps measurements were performed on 13 rooftops
in 7 different sites at altitudes from 88m to 1340m and in the Dolomites on sites. In Germany
measurements were performed on three roofs in 2 different sites at altitudes of 141m and
880m, and in Switzerland on 35 rooftops at 8 different sites at altitudes from 570m to 1628m.
In total, 81 measurements were carried out on the roof, see Sanpaolesi L., 1999. Scientific
Support Activity In The Field Of Structural Stability Of Civil Engineering Works - Snow
Loads, Final Report, Commission Of The European Communities, DGIII - D3, Contract n
500990/1997, 172p. These in-situ measurement information were completed with results of
laboratory tests at "Jules Verne climatic tunnel" Centre Scientifique et Technique du's
Btiment (CSTB), Nantes.
The snow load shape coefficients that should be used for roofs are given in Table 6.3 and in
Figure 6.3. Where snow fences or other obstructions exist or where the lower edge of the roof
is terminated with a parapet, then the snow load shape coefficient should not be reduced
below 0,8.
58
Table 6.3 Snow load shape coefficients used for monopitch roofs, pitched roofs and for multispan roofs
Angle of pitch of roof, 0
00 300
600
0,8
0,0
1,6
Figure 6.3 Snow load shape coefficients used for monopitch roofs, pitched roofs and for
multi-span roofs
Case (i)
1(2)
1(1)
1(2)
1(1)
0.51(2)
Figure 6.4 Snow load shape coefficients and the load arrangements used for monopitch roofs,
pitched roofs and for multi-span roofs
In the case of monopitch roofs, the load arrangement of Figure 6.4 should be used for both the
undrifted and drifted load arrangements.
In the case of pitched roofs the load arrangement of Figure 6.4 are as follows:
- The undrifted load arrangement which should be used is case (i);
- The drifted load arrangements which should be used are cases (ii) and (iii), unless
specified for local conditions.
For multi-span roofs the load arrangement of Figure 6.4 are as follows:
- The undrifted load arrangement which should be used is shown in case (i);
- The drifted load arrangement which should be used is shown in case (ii), unless
specified for local conditions.
The snow load shape coefficients that should be used for cylindrical roofs, in absence of snow
fences, are given in the Figure 6.5.
0,8
Cazul (i)
Cazul (ii)
0,53
ls/4
ls/4
ls/4
ls/4
ls
60
The snow load shape coefficients and drift lengths for quasi-horizontal roofs should be taken
as follows (see Figure 6.1), unless specified for local conditions:
1 = 0,8
2 = h / sk with restrictions 0,8 2 2,0
- weight density of snow (2 kN/m3)
ls = 2 h with restrictions 5 m ls 15 m.
The design of those parts of a roof cantilevered out beyond the walls should take account of
snow overhanging the edge of the roof, in addition to the load on that part of the roof. The
loads due to the overhang may be assumed to act at the edge of the roof and may be calculated
as follows:
se
se = k s2 /
se - snow load per metre length due to the overhang
s - the most onerous undrifted load case
- weight density of snow (3 kN/m3)
k - a coefficient to take account of the irregular shape of the
snow
k = 3/d with restrictions k d ,
61
v(z,t)
v(z,t)
vm(z)
vm(z)
62
1
vb2 0.625 vb2 m / s
2
(7.2)
The conversion of velocity pressure averaged on 10 min into velocity pressure averaged on
other time interval can be computed from relation (9.1):
10min
1.1q1h
0.7q1min
0.44qb3 s .
b qb
b
(7.3)
The reference wind velocity pressure qb (in kPa) with mean recurrence interval MRI=50 yr.
averaged on 10 min. at 10 m above ground in open terrain is presented in Figure 7.1.
Figure 7.1. The reference wind velocity pressure, [kPa] with MRI=50 yr.
wind velocity averaged on 10 min. at 10 m above ground in open terrain. The mapped values
are valid for sites situated bellow 1000 m
63
The mean wind velocity vm(z) at a height z above the terrain depends on the terrain roughness
and orography and on the basic wind velocity, vb and should be determined using Expression
(7.4).
vm(z) = co(z) cr (z ) vb
(7.4)
where:
co(z) - is the orography factor, taken as 1,0
cr (z ) -is the roughness factor
7.3 Terrain roughness and Variation of the mean wind with height
The roughness of the ground surface is aerodynamically described by the roughness length, zo,
(in meters), which is a measure of the size of the eddies close to the ground level. Various
terrain categories are classified in Table 7.1 according to their approximate roughness lengths.
Table 7.1. Roughness length zo, in meters, for various terrain categories 1)
Terrain
Terrain description
category
0
Sea or coastal area exposed to the open sea
I
Lakes or flat and horizontal area with negligible vegetation and
without obstacles
II
Area with low vegetation such as grass and isolated obstacles
(trees, buildings) with separations of at least 20 obstacle heights
III
Area with regular cover of vegetation or buildings or with
isolated obstacles with separations of maximum 20 obstacle
heights (such as villages, suburban terrain, permanent forest)
IV
Area in which at least 15 % of the surface is covered with
buildings and their average height exceeds 15 m
1)
2)
1
1
0.05
0.3
1.0
10
2)
For the full development of the roughness category, the terrains of types 0 to III must prevail in the
up wind direction for a distance of at least of 500m to 1000m, respectively. For category IV this
distance is more than 1 km.
z
k r ln
z0
cr z
cr z min
for z z min
Where,
kr is terrain factor depending on the roughness length calculated with:
z
k r 0,189 0
0,05
0 , 07
(7.6).
z0 and zmin values are given in the Table 7.1. The values kr are indicated in Table 7.2.
UTCB, Technical University of Civil Engineering, Bucharest
64
II
III
IV
0,155
0,024
0,169
0,028
0,189
0,036
0,214
0,046
0,233
0,054
(a)
(a)
(b)
(b)
(a)
(a)
(b)
(b)
65
The logarithmic profile of the mean wind velocity is valid for moderate and strong winds
(mean velocity > 10 m/s) in neutral atmosphere (where the vertical thermal convection of the
air may be neglected).
Wind velocity will be the same
(almost constant)
(almost constant)
200
180
Terrain category IV
160
140
Terrain category II
120
Terrain category I
100
Terrain category 0
80
60
40
20
0
0
0.5
1.5
2.5
3.5
z 0
(7.7)
z
2
k r z 0 ln for z min z z max 200 m
2
z 0
cr z
c 2 z z
for z z min
min
r
(7.8)
Where:
66
v
v m z
(7.9)
The recommended rules for the determination of Iv (z) are given in Expression (7.10):
2,5 ln z
z
I v z
0
for z z min
I v z z min
(7.10)
The values of (Table 7.3) depends of terrain roughness factor (z0, m):
(7.11)
0
2,74
I
2,74
II
2,66
III
2,35
IV
2,12
(7.12)
v p z c pv z v m z ;
(7.13)
c pv z
v p z
v m z
1 g I v z 1 3,5 I v z
(7.14)
Where:
g is the peak factor, recommended value g=3,5
cpv(z) is gust factor for mean wind velocity.
The peak velocity pressure qp(z) at height z, which includes mean and short-term velocity
fluctuations, should be determined by:
q p z
1
2
2
vm z 1 g I v z qm z 1 7 I v z
2
67
The gust factor for velocity pressure is the ratio of the peak velocity pressure to the mean
wind velocity pressure.
c pq z
q p z
qm z
1 2 g I v z 1 7 I v z
(7.15)
The peak velocity pressure at the height z above ground is the product of the gust factor, the
roughness factor and the reference velocity pressure.
(7.16)
q p z ce z q b
(7.17)
The exposure factor is defined as the product of the gust and roughness factors:
ce z cpq z cr2 z
(7.18)
If we cannot neglect the orography effect, the exposure factor ce(z) is:
(7.19)
200
z, m
180
160
Teren categoria IV
140
120
100
Teren categoria II
80
Teren categoria I
60
Teren categoria 0
40
20
0
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
68
the wind, the horizontal frictional forces acting tangentially to the surface can have a
significant effect.
The wind action is classified as a fixed variable action; actions are assessed as wind pressure /
suction or forces and are represented by their characteristic values.
Wind actions on structures and structural elements shall be determined taking account of both
external and internal wind pressures.
(7.20)
where:
qp(ze)
ze
cpe
Iw
The wind pressure acting on the internal surfaces of a structure, wi, should be obtained from
wi Iw c pi q p z i
(7.21)
where:
qp(zi) is the peak velocity pressure ;
zi
is the reference height for the internal pressure;
cpi
is the pressure coefficient for the internal pressure.
69
(7.22)
or by vectorial summation over the individual structural elements by using Expression (7.23):
Fw Iw cd
qp ze Aref
(7.23)
elements
Where:
qp(ze)
cd
cf
Aref
(ii)
The wind force, Fw acting on a structure or a structural element may be determined by
vectorial summation of the forces Fw,e and Fw,i, calculated from the external and internal
pressures using Expressions (7.24) and (7.25) and the frictional forces resulting from the
friction of the wind parallel to the external surfaces, calculated using Expression (7.26).
external forces:
Fw,e cd
w z A
e
ref
(7.24)
surfaces
internal forces:
Fw,i
w z A
i
ref
(7.25)
surfaces
friction forces:
F fr Iw c fr q p z e A fr
(7.26)
where:
cd
we(ze)
wi(zi)
Aref
cfr
Afr
70
AE d = I,e AE k
(8.1)
71
S e ( T ) a g T
(8.2)
where
Se (T) is the elastic response spectrum;
T
is the vibration period of a linear single-degree-of-freedom system;
ag
is the design ground acceleration (m/s2), from zonation map;
T is the normalized elastic response spectrum of absolute accelerations.
In the case of 0,05% viscous damping, the normalized elastic response spectrum of absolute
accelerations is defined by the following relations, Figure 8.3:
0 1 T
0 T TB
(T) 1
TB < T TC
TC < T TD
T
(T) 0 C
T
T T
(T) 0 C D
T2
TD < T 5s
where:
0
T
TB
TC
TD
TB
(8.3)
(8.4)
(8.5)
(8.6)
The values of the periods TB, TC , TD describing the shape of the elastic response spectrum
depend upon the ground type. In Romania, based on seismic records, the zonation of control
period TC (s) is given in Figure 8.2.
72
Table 8.1 Values of the periods TB, TC , TD describing the horizontal elastic response spectra
TC (s)
TB (s)
TD (s)
0,7
0,14
3,0
1,0
0,2
3,0
1,6
0,32
2,0
3
= 2.5
=0,05
2.5
(T)
2
1.75/T
1.5
5.25/T 2
1
T C =0.7s
0.5
0
T B =0.14
0
0.5
T D =3
1
1.5
2
2.5
Perioada T , s
3.5
3
=2.5
=0,05
2.5
(T)
2
2.5/T
1.5
7.5/T 2
1
0.5
T B =0.2 T C =1.0s
0
0
0.5
T D =3
1.5
2
2.5
Perioada T , s
3.5
3
=2.5
=0,05
2.5
4/T
(T)
2
1.5
8/T 2
1
0.5
T B =0.32
0
0
0.5
T C =1.6s
1
T D =2
1.5
2
2.5
Perioada T , s
3.5
Figure 8.3 Shapes of normalized elastic response spectrum for TC = 0,7s, 1,0s and 1,6s (5%
damping)
73
The elastic displacement response spectrum, SDe(T) (in meters), shall be obtained by direct
transformation of the elastic acceleration response spectrum, Se(T), using the following
expression:
T
S De ( T ) S e ( T )
2
(8.7)
q
S d ( T ) a g 1
T
0 < T TB
(8.9)
TB
T > TB
Sd (T) a g
(T )
0 ,2 a g
q
(8.10)
74