Documente Academic
Documente Profesional
Documente Cultură
⎛ m1 0 0 0 0 ⎞ ⎛1⎞
⎜ ⎜
⎜0 m2 0 0
⎟ 0
⎜1⎟
Mass matrix: M := ⎜ 0 0 m3 0 ⎟ ⋅m 0 Influence Coefficients, r: r := ⎜ 1 ⎟
⎜0 0 0 m4 0 ⎟ ⎜1⎟
⎜ ⎜
⎝0 0 0 0 m5 ⎠ ⎝1⎠
Floor stiffness: k1 := 1 k2 := 1 k3 := 1 k4 := 1 k5 := 1
⎛ k1 + k2 −k2 0 0 0 ⎞ ⎛1⎞
⎜ ⎜
⎜ −k2 k2 + k3 −k3 0 0
⎟ ⎜2⎟
Stiffness matrix: K := ⎜ 0 −k3 k3 + k4 −k4 0 ⎟ ⋅k Floor No.: h := ⎜ 3 ⎟ ⋅ H
⎜ 0 0 −k4 k4 + k5 −k5 ⎟ ⎜4⎟
⎜ ⎜
⎝ 0 0 0 −k5 k5 ⎠ ⎝5⎠
〈 j〉
〈 j〉
mods := stack( z , ψ )
X
Normalized Mode Shapes : z := 0 i := 1 .. N + 1 ψ := ⎯⎯⎯⎯⎯ →
1, j
max X (〈 j〉
)
⎛ 1.252 ⎞
⎜
(ψT⋅M⋅r) j ⎜ −0.394 ⎟ ⎯⎯⎯
1
→
Participation Factor Γ j := ==> Γ = ⎜ −0.208 ⎟ Frequencies and Eigenvalues: f := ⋅ λ
( ψT ⋅ M ⋅ ψ ) j , j ⎜ −0.116 ⎟ 2⋅ π
⎜
⎝ 0.053 ⎠
⎛ 4.398 ⎞
⎜
T( )
ψ ⋅ M ⋅ r j⋅ Γ j ⎜ 0.436 ⎟
Effective Modal Mass and Modal Height
(as a function of "m" and "H," respectively)
MM :=
j m
==> MM = ⎜ 0.121 ⎟
⎜ 0.038 ⎟
==>
∑ MM = 5.00
⎜
⎝ 0.008 ⎠
⎛ 3.513 ⎞
⎜
( ψT ⋅ M ⋅ h ) j 1 ⎜ −1.204 ⎟
MH := ⋅ ==> MH = ⎜ 0.764 ⎟
j
(ψT⋅M⋅r) j H ⎜ −0.594 ⎟
⎜
⎝ 0.521 ⎠
⎛ 2.001 ⎞ ⎛ 0.500 ⎞
⎜ ⎜
→
Modal Periods (sec) and
1 ⎜ 0.685 ⎟ ⎜ 1.459 ⎟
modal frequencies (Hz): T := ==> T = ⎜ 0.435 ⎟ and f = ⎜ 2.300 ⎟
f
⎜ 0.338 ⎟ ⎜ 2.955 ⎟
⎜ ⎜
⎝ 0.297 ⎠ ⎝ 3.370 ⎠
4 4 4 4 4
3 3 3 3 3
2 2 2 2 2
1 1 1 1 1
1 0 1 1 0 1 1 0 1 1 0 1 1 0 1
∑ MM⋅m = 1295.337
⎛ m1 0 0 0 0 ⎞ ⎛1⎞
⎜ ⎜
⎜0 m2 0 0
⎟ 0
⎜1⎟
Mass matrix: M := ⎜ 0 0 m3 0 ⎟ ⋅m 0 Influence Coefficients, r: r := ⎜ 1 ⎟
⎜0 0 0 m4 0 ⎟ ⎜1⎟
⎜ ⎜
⎝0 0 0 0 m5 ⎠ ⎝1⎠
⎛ k1 + k2 −k2 0 0 0 ⎞ ⎛1⎞
⎜ ⎜
⎜ −k2 k2 + k3 −k3 0 0
⎟ ⎜2⎟
Stiffness matrix: K := ⎜ 0 −k3 k3 + k4 −k4 0 ⎟ ⋅k Floor No.: h := ⎜ 3 ⎟ ⋅ H
⎜ 0 0 −k4 k4 + k5 −k5 ⎟ ⎜4⎟
⎜ ⎜
⎝ 0 0 0 −k5 k5 ⎠ ⎝5⎠
〈 j〉
〈 j〉
mods := stack( z , ψ )
X
Normalized Mode Shapes : z := 0 i := 1 .. N + 1 ψ := ⎯⎯⎯⎯⎯ →
1, j
max X (〈 j〉
)
⎛ 1.037 ⎞
⎜
(ψT⋅M⋅r) j ⎜ −0.046 ⎟ ⎯⎯⎯
1
→
Participation Factor Γ j := ==> Γ = ⎜ −0.012 ⎟ Frequencies and Eigenvalues: f := ⋅ λ
( ψT ⋅ M ⋅ ψ ) j , j ⎜ −0.004 ⎟ 2⋅ π
⎜
⎝ 0.002 ⎠
⎛ 4.994 ⎞
⎜
T( )
ψ ⋅ M ⋅ r j⋅ Γ j ⎜ 0.006 ⎟
Effective Modal Mass and Modal Height
(as a function of "m" and "H," respectively)
MM :=
j m
==> MM = ⎜ 0.000 ⎟
⎜ 0.000 ⎟
==>
∑ MM = 5.00
⎜
⎝ 0.000 ⎠
⎛ 3.048 ⎞
⎜
( ψT ⋅ M ⋅ h ) j 1 ⎜ −38.558 ⎟
MH := ⋅ ==> MH = ⎜ 0.512 ⎟
j
(ψT⋅M⋅r) j H ⎜ −38.503 ⎟
⎜
⎝ 0.501 ⎠
⎛ 5.866 ⎞ ⎛ 0.170 ⎞
⎜ ⎜
→
Modal Periods (sec) and
1 ⎜ 0.900 ⎟ ⎜ 1.111 ⎟
modal frequencies (Hz): T := ==> T = ⎜ 0.482 ⎟ and f = ⎜ 2.074 ⎟
f
⎜ 0.351 ⎟ ⎜ 2.845 ⎟
⎜ ⎜
⎝ 0.299 ⎠ ⎝ 3.341 ⎠
4 4 4 4 4
3 3 3 3 3
2 2 2 2 2
1 1 1 1 1
1 0 1 1 0 1 1 0 1 1 0 1 1 0 1
∑ MM⋅m = 1295.337
⎛ m1 0 0 0 0 ⎞ ⎛1⎞
⎜ ⎜
⎜0 m2 0 0
⎟ 0
⎜1⎟
Mass matrix: M := ⎜ 0 0 m3 0 ⎟ ⋅m 0 Influence Coefficients, r: r := ⎜ 1 ⎟
⎜0 0 0 m4 0 ⎟ ⎜1⎟
⎜ ⎜
⎝0 0 0 0 m5 ⎠ ⎝1⎠
Floor stiffness: k1 := 3 k2 := 1 k3 := 1 k4 := 1 k5 := 1
⎛ k1 + k2 −k2 0 0 0 ⎞ ⎛1⎞
⎜ ⎜
⎜ −k2 k2 + k3 −k3 0 0
⎟ ⎜2⎟
Stiffness matrix: K := ⎜ 0 −k3 k3 + k4 −k4 0 ⎟ ⋅k Floor No.: h := ⎜ 3 ⎟ ⋅ H
⎜ 0 0 −k4 k4 + k5 −k5 ⎟ ⎜4⎟
⎜ ⎜
⎝ 0 0 0 −k5 k5 ⎠ ⎝5⎠
〈 j〉
〈 j〉
mods := stack( z , ψ )
X
Normalized Mode Shapes : z := 0 i := 1 .. N + 1 ψ := ⎯⎯⎯⎯⎯ →
1, j
max X (〈 j〉
)
⎛ 1.327 ⎞
⎜
(ψT⋅M⋅r) j ⎜ −0.658 ⎟ ⎯⎯⎯
1
→
Participation Factor Γ j := ==> Γ = ⎜ −0.441 ⎟ Frequencies and Eigenvalues: f := ⋅ λ
( ψT ⋅ M ⋅ ψ ) j , j ⎜ −0.169 ⎟ 2⋅ π
⎜
⎝ 0.046 ⎠
⎛ 3.316 ⎞
⎜
T( )
ψ ⋅ M ⋅ r j⋅ Γ j ⎜ 0.990 ⎟
Effective Modal Mass and Modal Height
(as a function of "m" and "H," respectively)
MM :=
j m
==> MM = ⎜ 0.634 ⎟
⎜ 0.056 ⎟
==>
∑ MM = 5.00
⎜
⎝ 0.004 ⎠
⎛ 3.555 ⎞
⎜
( ψT ⋅ M ⋅ h ) j 1 ⎜ 0.102 ⎟
MH := ⋅ ==> MH = ⎜ 0.941 ⎟
j
(ψT⋅M⋅r) j H ⎜ 0.114 ⎟
⎜
⎝ 1.509 ⎠
⎛ 2.001 ⎞ ⎛ 0.500 ⎞
⎜ ⎜
→
Modal Periods (sec) and
1 ⎜ 0.724 ⎟ ⎜ 1.382 ⎟
modal frequencies (Hz): T := ==> T = ⎜ 0.523 ⎟ and f = ⎜ 1.913 ⎟
f
⎜ 0.407 ⎟ ⎜ 2.457 ⎟
⎜ ⎜
⎝ 0.341 ⎠ ⎝ 2.929 ⎠
4 4 4 4 4
3 3 3 3 3
2 2 2 2 2
1 1 1 1 1
1 0 1 1 0 1 1 0 1 1 0 1 1 0 1
∑ MM⋅m = 1295.337
Mode Shapes (1st row is 1st floor; each column is a mode shape):
⎛⎜ Tei ⎞
Period Factors: r :=
i, j ⎜ Tj
⎝ ⎠
a) Elements of S-matrix:
S11( a , b , ζ , β ) := 1 ( 2) 2 (
S12( a , b , ζ , β ) := −2 ⋅ 1 − 2 ⋅ β ⋅ a − 2 ⋅ 1 − 2 ⋅ β ⋅ b
2) 2
S13( a , b , ζ , β ) := 1 S14( a , b , ζ , β ) := −2 ⋅ ( 1 − 2 ⋅ ζ ) − 2 ⋅ ( 1 − 2 ⋅ β ) ⋅ b
2 2 2
S15( a , b , ζ , β ) := 1 S16( a , b , ζ , β ) := −2 ⋅ ( 1 − 2 ⋅ ζ ) − 2 ⋅ ( 1 − 2 ⋅ β ) ⋅ a
2 2 2
( ) ⋅a2⋅b2
2
S21( a , b , ζ , β ) := S12( a , b , ζ , β ) S22( a , b , ζ , β ) := a + b + 4 ⋅ 1 − 2 ⋅ β
4 4 2
S31( a , b , ζ , β ) := S22( a , b , ζ , β ) ( 2) 2 4 ( 2)
S32( a , b , ζ , β ) := −2 ⋅ 1 − 2 ⋅ β ⋅ b ⋅ a − 2 ⋅ 1 − 2 ⋅ β ⋅ a ⋅ b
2 4
b) Elements of b_vector:
b1( a , b , ζ , β ) := 16⋅ β ⋅ ζ ⋅ a⋅ b
2 2
2 2 2 (2 2 2 )
b2( a , b , ζ , β ) := 4 ⋅ β ⋅ a⋅ b + 4 ⋅ ζ ⋅ ⎡⎣a ⋅ b − 4 ⋅ β ⋅ a⋅ b ⋅ a + b − 4 ⋅ β ⋅ a⋅ b ⎤⎦
2 2
b3( a , b , ζ , β ) := a ⋅ b − 4 ⋅ β ⋅ a⋅ b ⋅ ( a + b − 4 ⋅ β ⋅ a⋅ b ) − 4 ⋅ β ⋅ a ⋅ b ⋅ ( a + b − 8 ⋅ β ⋅ a⋅ b )
2 2 2 2 2 2 2 2 2 2 2 2
b4( a , b , ζ , β ) := 4 ⋅ ζ ⋅ a ⋅ b − a ⋅ b ⋅ ( a + b − 8 ⋅ β ⋅ a⋅ b )
2 4 4 2 2 2 2 2
b5( a , b , ζ , β ) := a ⋅ b
4 4
b6( a , b , ζ , β ) := 0
⎛ b1( a , b , ζ , β ) ⎞
⎜
⎜ b2( a , b , ζ , β ) ⎟
⎜ b3( a , b , ζ , β ) ⎟
b_( a , b , ζ , β ) := ⎜ ⎟
⎜ b4( a , b , ζ , β ) ⎟
⎜ b5( a , b , ζ , β ) ⎟
⎜ b6( a , b , ζ , β )
⎝ ⎠
c) Elements of a_vector:
−1
a_( a , b , ζ , β ) := S( a , b , ζ , β ) ⋅ b_( a , b , ζ , β )
a_( a , b , ζ , β ) 3 a_( a , b , ζ , β ) 4
ABp ( a , b , ζ , β ) := a_( a , b , ζ , β ) 1 + a_( a , b , ζ , β ) 2 CDp ( a , b , ζ , β ) := +
4 2
a a
a_( a , b , ζ , β ) 5 a_( a , b , ζ , β ) 6
EFp( a , b , ζ , β ) := +
4 2
b b
d) Factors A, B, C, and D:
( 2 2 4 3 2 2 )(
DD( a , ζ , β ) := 4 ⋅ a ⋅ β + 1 − 2 ⋅ a + a − 4 ⋅ β ⋅ a⋅ ζ − 4 ⋅ β ⋅ a ⋅ ζ + 4 ⋅ a ⋅ ζ ⋅ 4 ⋅ a ⋅ β + 1 − 2 ⋅ a + a + 4 ⋅ β ⋅ a⋅ ζ + 4 ⋅ β ⋅ a ⋅ ζ + 4 ⋅ a
2 2 2 2 4 3 2
8 6 6 2 6 2 2 4 4 2 4 2 4 4 2 2
a − 4 ⋅ a + 8 ⋅ a ⋅ β − 16⋅ a ⋅ β ⋅ ζ + 3 ⋅ a − 8 ⋅ a ⋅ ζ − 8 ⋅ a ⋅ β + 16⋅ a ⋅ β + 8 ⋅ a ⋅ β
A( a , ζ , β ) :=
DD( a , ζ , β )
6 4 8 2 2 2 6 2 2 6 2 4 4 2 2 4 4 2
2 ⋅ a − 2 ⋅ a + 4 ⋅ a ⋅ ζ − 4 ⋅ a ⋅ β + 32⋅ a ⋅ β ⋅ ζ − 64⋅ a ⋅ β ⋅ ζ − 32⋅ a ⋅ β ⋅ ζ + 64⋅ a ⋅ β ⋅ ζ
B( a , ζ , β ) :=
DD( a , ζ , β )
4 4 2 2 2 2 2 2 2 6 2 4 2 4 4
3 ⋅ a − 8 ⋅ a ⋅ β − 4 ⋅ a + 8 ⋅ a ⋅ ζ − 16⋅ a ⋅ ζ ⋅ β + 1 + 8 ⋅ a ⋅ ζ − 8 ⋅ a ⋅ ζ + 16⋅ a ⋅ ζ
C( a , ζ , β ) :=
DD( a , ζ , β )
−B( a , ζ , β )
D( a , ζ , β ) :=
2
a
a) Get SA values at natural periods of the system (by cubic spline interpolation):
( )
0.4
〈1〉 〈2〉 〈1〉 〈2〉
SAs := interp⎛ vs , RSg , RSg , T ⎞ ⋅ SAF⋅ ⎛⎜
RSd ⎞
vs := lspline RSg , RSg
j ⎝ j⎠
⎝ βj ⎠
b) Get SA values at equipment periods (by cubic spline interpolation):
0.4
〈1〉 〈2〉
SAe := interp⎛ vs , RSg , RSg , Te ⎞ ⋅ SAF⋅ ⎛⎜
RSd ⎞
i ⎝ i⎠
⎝ ζ ⎠
5) Compute the desired Floor Response Spectrum (RSf)
6
SA (g)
0
0 0.5 1 1.5 2 2.5 3
15.1 INTRODUCTION
The basic mode superposition method, which is restricted to linearly elastic analysis,
produces the complete time history response of joint displacements and member
forces. In the past there have been two major disadvantages in the use of this
approach. First, the method produces a large amount of output information that can
require a significant amount of computational effort to conduct all possible design
checks as a function of time. Second, the analysis must be repeated for several
different earthquake motions in order to assure that all frequencies are excited, since
a response spectrum for one earthquake in a specified direction is not a smooth
function.
The purpose of this chapter is to summarize the fundamental equations used in the
response spectrum method and to point out the many approximations and limitations
of the method. For, example it cannot be used to approximate the nonlinear
response of a complex three-dimensional structural system.
The recent increase in the speed of computers has made it practical to run many time
history analyses in a short period of time. In addition, it is now possible to run
design checks as a function of time, which produces superior results, since each
member is not designed for maximum peak values as required by the response
spectrum method.
For three dimensional seismic motion, the typical modal Equation (13.6) is rewritten
as
&& n + 2 ζ n ω n y(t)
& n + ω n y(t)n =
2
y(t)
(15.1)
&& gx + pny u(t)
pnx u(t) && gy + pnz u(t)
&& gz
&y(t)
& n + 2 ζ n ω n y(t)
& n + ω n y(t)n = pni u(t)
2
&& g (15.2)
RESPONSE SPECTRUM ANALYSIS 3
Equation (15.2) can be solved for &y&(t ) and substituted into Equation (15.3) which
yields
Therefore, for the special case of zero damping, the total acceleration of the system
is equal to ω 2 y (t ) . For this reason, the displacement response spectrum curve is
normally not plotted as modal displacement y (ω ) MAX vs ω . It is standard to
present the curve in terms of S(ω ) vs. a period T in seconds. where
2π
S (ω ) a = ω 2 y (ω ) MAX and T= (15.5a) and (15.5b)
ω
The pseudo-acceleration spectrum, S(ω ) a , curve has the units of acceleration vs.
period which has some physical significance for zero damping only. It is apparent
that all response spectrum curves represent the properties of the earthquake at a
specific site and are not a function of the properties of the structural system. After
4 STATIC AND DYNAMIC ANALYSIS
The maximum modal displacement, for a structural model, can now be calculated
for a typical mode n with period Tn and corresponding spectrum response value
S (ω n ) . The maximum modal response associated with period Tn is given by
S (ω n )
y (Tn ) MAX = (15.6)
ω 2n
The corresponding internal modal forces, f kn , are calculated from standard matrix
structural analysis using the same equations as required in static analysis.
A ten second segment of the Loma Prieta earthquake motions, recorded on a soft site
in the San Francisco Bay Area, is shown in Figure 15.1. The record has been
corrected, by use of an iterative algorithm, for zero displacement, velocity and
acceleration at the beginning and end of the ten second record. For the earthquake
motions given in Figure 15.1a, the response spectrum curves for displacement and
pseudo-acceleration are summarized in Figure 15.2a and 15.2b
The velocity curves have been intentionally omitted since they are not an essential
part of the response spectrum method. Furthermore, it would require considerable
space to clearly define terms such as peak ground velocity, pseudo velocity
spectrum, relative velocity spectrum and absolute velocity spectrum.
RESPONSE SPECTRUM ANALYSIS 5
25
20
15
10
-5
-10
-15
-20
-25
0 1 2 3 4 5 6 7 8 9 10
TIME - seconds
-2
-4
-6
-8
-10
-12
0 1 2 3 4 5 6 7 8 9 10
TIME - seconds
20
18
16
14
12
10
8
1.0 Percent Damping
6 5.0 Percent Damping
4
2
0
0 1 2 3 4 5
PERIOD - Seconds
100
90
80
1.0 Perent Damping
70 5.0 Percent Damping
60
50
40
30
20
10
0
0 1 2 3 4 5
PERIOD - Seconds
The maximum ground acceleration, for the earthquake defined by Figure 15.1a, is
20.01 percent of gravity at 2.92 seconds. It is important to note that the pseudo
acceleration spectrum, shown in Figure 15.2b, has the same value for a very short
period system. This is due to the physical fact that a very rigid structure moves as a
rigid body and the relative displacements within the structure are equal to zero as
indicated by Figure 15.2a. Also, the behavior of a rigid structure is not a function of
the viscous damping value.
The maximum ground displacement shown in Figure 15.1b is -11.62 inches at 1.97
seconds. For long period systems, the mass of the one-degree-of-freedom structure
does not move significantly and has approximately zero absolute displacement.
Therefore, the relative displacement spectrum curves, shown in Figure 15.2a, will
converge to 11.62 inches for long periods and all values of damping. This type of
real physical behavior is fundamental to the design of base isolated structures.
The relative displacement spectrum, Figure 15.2a, and the absolute acceleration
spectrum, Figure 15.2b, have physical significance. However, the maximum
relative displacement is directly proportional to the maximum forces developed in
the structure. For this earthquake the maximum relative displacement is 18.9 inches
at a period of 1.6 seconds for one percent damping and 16.0 inches at a period of
four seconds for five percent damping. It is important to note the significant
difference between one and five percent damping for this typical soft site record.
The most conservative method that is used to estimate a peak value of displacement
or force within a structure is to use the sum of the absolute of the modal response
values. This approach assumes that the maximum modal values, for all modes,
occur at the same point in time.
Another very common approach is to use the Square Root of the Sum of the
Squares, SRSS, on the maximum modal values in order to estimate the values of
displacement or forces. The SRSS method assumes that all of the maximum modal
values are statistically independent. For three dimensional structures, in which a
large number of frequencies are almost identical, this assumption is not justified.
The peak value of a typical force can now be estimated, from the maximum modal
values, by the CQC method with the application of the following double summation
equation:
F= ∑∑ f ρ
n m
n nm fm (15.9)
where f n is the modal force associated with mode n . The double summation is
conducted over all modes. Similar equations can be applied to node displacements,
relative displacements and base shears and overturning moments.
The cross-modal coefficients, ρ nm , for the CQC method with constant damping are
8ζ 2 (1 + r ) r 3/ 2
ρnm = (15.10)
(1 − r 2 ) 2 + 4ζ 2r (1 + r ) 2
RESPONSE SPECTRUM ANALYSIS 9
The problems associated with the use of the absolute sum and the SRSS of modal
combination can be illustrated by their application to the four story building shown
in Figure 15.3. The building is symmetrical; however, the center of mass, of all
floors, is located 25 inches from the geometric center of the building.
The direction of the applied earthquake motion, a table of natural frequencies and
the principal direction of the mode shape are summarized in Figure 15.4. One notes
the closeness of the frequencies which is typical of most three dimensional building
structures that are designed to equally resist earthquakes from both directions.
Because of the small mass eccentricity, which is normal in real structures, the
fundamental mode shape has x, y, as well as torsion components. Therefore, the
model represents a very common three dimensional building system. Also, note that
there is not a mode shape in a particular given direction as implied in many
building codes and some text books on elementary dynamics.
The building was subjected to one component of the Taft, 1952, earthquake. An
exact time history analysis, using all 12 modes, and a response spectrum analysis
were conducted. The maximum modal base shears in the four frames for the first
five modes are shown in Figure 15.5.
Figure 15.6 summarizes the maximum base shears, in each of the four frames, using
different methods. The time history base shears, Figure 15.6a, are exact. The
SRSS method, Figure 15.6b, produces base shears which under-estimate the exact
values in the direction of the loads by approximately 30 percent and over-estimate
the base shears normal to the loads by a factor of ten. The sum of the absolute
RESPONSE SPECTRUM ANALYSIS 11
values, Figure 15.6c, grossly over-estimates all results. The CQC method, Figure
15.6d, produces very realistic values that are close to the exact time history solution.
Figure 15.5. Base Shears in Each Frame for First Five Modes
The modal cross-correlation coefficients for this building are summarized in Table
15.1. It is of importance to note the existence of the relatively large off-diagonal
terms that indicate which modes are coupled.
Mode 1 2 3 4 5 ω n rad/sec
If one notes the signs of the modal base shears, shown in Figure 15.3, it is apparent
how the application of the CQC method allows the sum of the base shears in the
direction of the external motion to be added directly. In addition, the sum of the
base shears, normal to the external motion, tend to cancel. The ability of the CQC
method to recognize the relative sign of the terms in the modal response is the key to
the elimination of errors in the SRSS method.
RESPONSE SPECTRUM ANALYSIS 13
Design spectra are not uneven curves as shown in Figure 15.2 since they are
intended to be the average of many earthquakes. At the present time, many building
codes specify design spectra in the form shown in Figure 15.7.
3
Normallized Pseudo Acceleration
2.5
1.5
0.5
0
0 2 4 6 8 10
PERIOD - Seconds
The Uniform Building Code has defined specific equations for each range of the
spectrum curve for four different soil types. For major structures it is now common
practice to develop a site-dependent design spectrum which includes the effect of
local soil conditions and distance to the nearest faults.
14 STATIC AND DYNAMIC ANALYSIS
It is reasonable to assume that motions that take place during an earthquake have
one principal direction [1]. Or, during a finite period of time, when maximum
ground acceleration occurs, a principal direction exists. For most structures this
direction is not known and, for most geographical locations, cannot be estimated.
Therefore, the only rational earthquake design criterion is that the structure must
resist an earthquake of a given magnitude from any possible direction. In addition to
the motion in the principal direction, a probability exists that motions normal to that
direction will occur simultaneously. In addition, because of the complex nature of
three dimensional wave propagation, it is valid to assume that these normal motions
are statistically independent.
Based on these assumptions, a statement of the design criterion is "a structure must
resist a major earthquake motion of magnitude S1 for all possible angles θ and, at
the same point in time, resist earthquake motions of magnitude S2 at 90o to the angle
θ ". These motions are shown schematically in Figure 15.1.
RESPONSE SPECTRUM ANALYSIS 15
The stated design criterion implies that a large number of different analyses must be
conducted in order to determine the maximum design forces and stresses. It will be
shown, in this section, that maximum values for all members can be exactly
evaluated from one computer run in which two global dynamic motions are applied.
Furthermore, the maximum member forces calculated are invariant with respect to
the selection system.
90
S2
90
S1
Plan View
0
Figure 15.7 indicates that the basic input spectra S1 and S2 are applied at an
arbitrary angle θ . At some typical point within the structure, a force, stress or
displacement F is produced by this input. In order to simplify the analysis, it will
be assumed that the minor input spectrum is some fraction of the major input
spectrum. Or,
S2 = a S1 (15.11)
Recently, Menun and Der Kiureghian [3] presented the CQC3 method for the
combination of the effects of orthogonal spectrum.
16 STATIC AND DYNAMIC ANALYSIS
where
F02 = ∑ ∑ f 0 n ρ nm f 0 m (15.13)
n m
F902 = ∑ ∑ f 90 n ρ nm f 90 m (15.14)
n m
F0− 90 = ∑ ∑ f 0 n ρ nm f 90 m (15.15)
n m
FZ2 = ∑ ∑ f z n ρ nm f z m (15.16)
n m
in which f 0 n and f 90 n are the modal values produced by 100 percent of the lateral
spectrum applied at 0 and 90 degrees respectively and f z n is the modal response
from the vertical spectrum which can be different from the lateral spectrum.
It is important to note that for equal spectra a = 1 , the value F is not a function of
θ and the selection of the analysis reference system is arbitrary. Or,
F02 + F902 + Fz
2
FMAX = (15.17)
This indicates that it is possible to conduct only one analysis, with any reference
system, and the resulting structure will have all members that are designed to equally
resist earthquake motions from all possible directions. This method is acceptable by
most building codes.
RESPONSE SPECTRUM ANALYSIS 17
For a=1 the CQC3 method reduces to the SRSS method. However, this can be over
conservative since real ground motions of equal value in all directions have not been
recorded. Normally, the value of θ in Equation (15.12) is not known; therefore, it
is necessary to calculate the critical angle that produces the maximum response.
Differentiation of Equation (15.12) and setting the results to zero yields
1 −1 2 F0−90
θcr = tan [ 2 ] (15.17)
2 F0 − F902
Two roots exist for Equation (15.17) that must be checked in order that the
following equation is maximum:
At the present time no specific guidelines have been suggested for the value of a .
Reference [3] presented an example with values a between 0.50 and 0.85.
The previously presented theory clearly indicates that the CQC3 combination rule,
with a equal 1.0, is identical to the SRSS method and produces results, for all
structural systems, which are not a function of the reference system used by the
engineer. One example will be presented in order to show the advantages of the
method. A very simple one-story structure, shown in Figure 15.8, was selected to
compare the results of the 100/30 and 100/40 percentage rules with the SRSS rule.
Note that the masses are not at the geometric center of the structure. The structure
has two translations and one rotational degrees-of-freedom located at the center of
mass. The columns, which are subjected to bending about the local 2 and 3 axes,
are pinned at the top where they are connected to an in-plane rigid diaphragm.
18 STATIC AND DYNAMIC ANALYSIS
3 2
Typical Column:
X = Y = 106.065 ft. I 22 = 100 ft 4
4
3 2 I 33 = 200ft 4
Sym. E = 30 k / ft 2
3
X = Y = 70.717 ft. L = 10 ft
M TOP = 0.25k - sec 2 / ft
3 3
1 2 2 2
X
0
X = 100 ft. X = 150 ft.
The periods and normalized base shear forces associated with the mode shapes are
summarized in Table 15.2. Since the structure has a plane of symmetry at 22.5
degrees, the second mode has no torsion and has a normalized base shear at 22.5
degrees with the x-axis. Due to this symmetry, it is apparent that columns 1 and 3
(or columns 2 and 4) should be designed for the same forces.
Period X- Y- X- Y-
Mode Seconds Force Force Moment Moment Torsion
The definition of the mean displacement response spectrum used in the spectra
analysis is given in Table 15.3.
RESPONSE SPECTRUM ANALYSIS 19
Spectral
Mode Period X- Y- Displacement
Seconds MASS MASS Used For
Analysis
1 1.01029 14.43 84.081 1.00
2 0.76918 85.36 14.65 1.00
3 0.43102 0.22 1.27 1.00
The moments about the local 2 and 3 axes at the base of each of the four columns
for the spectrum applied separately at 0.0 and 90 degrees are summarized in Tables
15.4 and 15.5 and are compared to the 100/30 rule.
For this example, the maximum forces do not vary significantly between the two
methods. However, it does illustrate that the 100/30 combination method produces
20 STATIC AND DYNAMIC ANALYSIS
moments which are not symmetric, whereas the SRSS combination method produces
logical and symmetric moments. For example, member 4 would be over-designed by
3.4 percent about the local 2-axis and under-designed by 7.8 percent about the local
3-axis using the 100/30 combination rule.
The SRSS and 100/40 design moments about the local 2 and 3 axes at the base of
each of the four columns are summarized in Tables 15.6 and 15.7
The results presented in Tables 15.6 and 15.7 also illustrate that the 100/40
combination method produces results which are not reasonable. Because of
symmetry, members 1 and 3 and members 2 and 4 should be designed for the same
moments. Both the 100/30 and 100/40 rules fail this simple test.
RESPONSE SPECTRUM ANALYSIS 21
For three dimensional response spectra analyses, it has been shown that the "design
of elements for 100 percent of the prescribed seismic forces in one direction plus 30
or 40 percent of the prescribed forces applied in the perpendicular direction" is
dependent on the user’s selection of the reference system. These commonly used
"percentage combination rules" are empirical and can underestimate the design
forces in certain members and produce a member design which is relatively weak in
one direction. It has been shown that the alternate building code approved method,
in which an SRSS combination of two 100 percent spectra analyses with respect to
any user defined orthogonal axes, will produce design forces that are not a function
of the reference system. Therefore, the resulting structural design has equal
resistance to seismic motions from all directions.
The use of the CQC3 method should be used if a value of a less than 1.0 can be
justified. It will produce realistic results that are not a function of the user selected
reference system.
It is apparent that use of the response spectrum method has limitations, some of
which can be removed by additional development. However, it will never be
accurate for nonlinear analysis of multi-degree of freedom structures. The author
believes that in the future more time history dynamic response analyses will be
conducted and the many approximations associated with the use of the response
spectrum method will be avoided. Some of these additional limitations will be
discussed in this section.
22 STATIC AND DYNAMIC ANALYSIS
All displacements produced by the response spectrum method are positive numbers.
Therefore, a plot of a dynamic displaced shape has very little meaning since each
displacement is an estimation of the maximum value. Inter-story displacements are
used to estimate damage to nonstructural elements and cannot be calculated directly
from the probable peak values of displacement. A simple method to obtain a
probable peak value of shear strain is to place a very thin panel element, with a
shear modulus of unity, in the area where the deformation is to be calculated. The
peak value of shear stress will be a good estimation of the damage index. The
current code suggests a maximum value of 0.005 horizontal drift ratio, which is the
same as panel shear strain if the vertical displacements are neglected.
The fundamental equation for the calculation of the stresses within the cross section
of a beam is
P My x Mx y
σ= + + (15.19)
A Iy Ix
This equation can be evaluated for a specified x and y point in the cross section and
for the calculated maximum spectral axial force and moments which are all positive
values. It is apparent that the resulting stress may be conservative since all forces
will probably not obtain their peak values at the same time.
For response spectrum analysis, the correct and accurate approach for the
evaluation of equation (15.19) is to evaluate the equation for each mode of vibration.
This will take into consideration the relative signs of axial forces and moments in
each mode. An accurate value of the maximum stress can then be calculated from
the modal stresses using the CQC double sum method. It has been the author’s
experience, with large three dimensional structures, that stresses calculated from
modal stresses can be less than 50 percent of the value calculated using maximum
peak values of moments and axial force.
RESPONSE SPECTRUM ANALYSIS 23
Unfortunately, most design check equations for steel structures are written in terms
of "design strength ratios" which are a nonlinear function of the axial force in the
member; therefore, the ratios cannot be calculated in each mode. A new
approximate method, to replace the current state of the art approach of calculating
strength ratios based on maximum peak values of member forces, is proposed by the
author. This would involve first calculating the maximum axial force. The design
ratios would then be evaluated mode by mode, assuming the maximum axial force
reduction factor remains constant for all modes. The design ratio for the member
would then be estimated by a double-sum modal combination method such as the
CQC3 method. This approach would improve accuracy and still be conservative.
With respect to the interesting problem of calculating the maximum shear force in a
bolt, it is not correct to estimate the maximum shear force from a vector summation
since the x and y shears do not obtain their peak values at the same time. A correct
method of estimating the maximum shear in a bolt is to check the maximum bolt
shear at several different angles about the bolt axis. This would be a tedious
approach using hand calculations; however, if the approach is built into a post
processor computer program, the computational time to calculate the maximum bolt
force is trivial.
The same problem exists if principal stresses are to be calculated from a response
spectrum analysis. One must check at several angles in order to estimate the
maximum and minimum value of the stress at each point in the structure.
24 STATIC AND DYNAMIC ANALYSIS
15.10 SUMMARY
In this chapter it has been illustrated that the response spectrum method of dynamic
analysis must be used carefully. The CQC method should be used to combine modal
maxima in order to minimize the introduction of avoidable errors. The increase in
computational effort, as compared to the SRSS method, is small compared to the
total computer time for a seismic analysis. The CQC method has a sound
theoretical basis and has been accepted by most experts in earthquake engineering.
The use of the absolute sum or the SRSS method for modal combination cannot be
justified.
In order for a structure to have equal resistance to earthquake motions from all
directions, the CQC3 method should be used to combine the effects of earthquake
spectra applied in three dimensions. The percentage rule methods have no
theoretical basis and are not invariant with respect to the reference system.
Engineers, however, should clearly understand that the response spectrum method is
an approximate method used to estimate maximum peak values of displacements and
forces and that it has significant limitations. It is restricted to linear elastic analysis
in which the damping properties can only be estimated with a low degree of
confidence. The use of nonlinear spectra, which are commonly used, has very little
theoretical background and should not be used for the analysis of complex three
dimensional structures. For such structures, true nonlinear time-history response
should be used as indicated in Chapter 19.
15.11 REFERENCES
1. J. Penzien and M. Watabe, "Characteristics of 3-D Earthquake Ground Motions,"
Earthquake Engineering and Structural Dynamics, Vol. 3, pp. 365-373, 1975.
2. E. L. Wilson, A. Der Kiureghian and E. R. Bayo, "A Replacement for the SRSS
Method in Seismic Analysis," Earthquake Engineering and Structural Dynamics,
Vol. 9, pp. l87-l92, 1981.
3. C. Menun and A. Der Kiureghian, “A Replacement for the 30 % Rule for
Muticomponent Excitation”, Earthquake Spectra, Vol. 13, Number 1, February
1998.
Multiple degrees of freedom structural dynamics
Example 1
For the building shown in Fig. 1, we are interested in the response in the direction
of the numeric reference axes. The lateral stiffness of each one of the stories is the
same and can be defined as k. The mass of the two lower stories is twice that of
the roof. The roof mass is defined as m.
U3 3rd story
m U3
2nd story
k
U2
2m U2
1st story k
U1
2m U1
k
C 2
1
A
↓ dof
m 0 0 U3
[ M ] = 0 2m 0 U 2
0 0 2m U1
↓ dof
k −k 0 U3
[ K ] = −k 2k −k U 2
0 −k 2k U1
m 0 0 U k −k 0 U 3 0
3
0 2m 0 U 2 + −k 2k −k U 2 = 0
0
0 2m U
0
1 −k 2k U1 0
We now proceed to find the solution of the free vibration response of the system for
different initial conditions. From Eq. (11) we have:
∆ = [ K ] − ωi 2 [ M ] = 0
k − ω2 m −k 0
∆= −k 2k − ω 2m
2
−k =0
0 −k 2k − ω 2m
2
∆ = 4m 3 ω6 − 12km 2 ω4 + 9k 2 mω2 − k 3 = 0
k 4 9 k2 2 1 k3
ω6 − 3 ω + ω − =0
m 4 m2 4 m3
A simple inspection of the equation tell us that ω2 = k/m is a root, and by using
synthetic division, we transform the characteristic equation into:
2 k
(
ω − m 4m ω − 8km ω + k m = 0
3 4 2 2 2
)
Solving the second-degree equation contained in the second term of the previous
equation, we obtain:
k
1.866
8km ± 64k m − 16k m
2
k 2
3 4
m
2 4
ω2 = 3
= 1 ± =〈
8m m 2 k
0.134
m
k k k
ω12 = 0.134 ω22 = ω23 = 1.866
m m m
Now, by using Eq. (13) we can obtain the vibration modes by going back to the
characteristic determinant:
{ }
[ K ] − ωr 2 [ M ] φ(r ) = {0} r = 1, 2 and 3
Replacing here the mass and stiffness matrices, we obtain the following set of
homogeneous simultaneous equations:
k − ωr2 m −k 0 φ(r3 ) 0
(r )
−k 2k − ωr2 2m −k φ 2 = 0
0 −k 2k − ωr2 2m φ1(r ) 0
Expanding the product, we see the system in the classical simultaneous equation
format:
(k − ω m ) φ
2
r
(r )
3 − kφ(r2 ) = 0
− kφ(r3 ) + ( 2k − ω 2m ) φ
2
r
(r )
2 − kφ1(r ) = 0
− kφ(r2 ) + ( 2k − ω 2m ) φ
2
r
(r )
1 = 0
From the third equation, we can see that, in this case, the ratio between the
second unknown and the first unknown is:
φ(r2 ) 2k − ωr2 2m
=
φ1(r ) k
Now replacing the third equation into the second, we obtain the following ratio
between the third unknown and the first unknown:
( )
2
φ(r3 ) 2k − ωr2 2m
= −1
φ1|(r ) k2
These two ratios are fixed for any value of ωi2 . We now replace the values of
ωi2 obtained previously and the values of the unknowns are found for each case:
We may assign any arbitrary value to the φ1 term and thus from the obtained
ratios compute the other two values of the terms of the mode. We choose,
arbitrarily again, a value of one for φ1. By doing so, the modes are defined as:
2 −1 2
{φ }
(1)
= 1.732 {φ }
(2)
= 0 {φ }
(3)
= −1.732
1 1 1
1.732
0
-1.732
1 1
1
k k k
ω12 = 0.134 ω22 = ω23 = 1.866
m m m
We now change the normalization of the modes in such a way that they comply
with Eq. (15) to obtain orthonormal modes:
{φ } [M ]{φ } = 1
T
(r ) (r )
Mode 1
m 0 0 2 2 0.5774 m
1
{2 }
3 1 0 2m 0 3 = 12m ⇒ φ(1) = { }
12m
3 = 0.5000 m
0 0 2m 1 1
0.2887 m
Mode 2
Mode 3
m 0 0 2 2 0.5774 m
{ }
2 − 3 1 0 2m 0 − 3 = 12m ⇒ φ( 3) =
1
{ }
− 3 = −0.5000
12m
m
0
0 2m 1
1 0.2887
m
Example 5
3m
U6
3m
U5
3m
U4
3m
U3
U2 3m
3m
U1
6m
6m
6m
6m x
o
Fig. 9 - Example 5
The first step is to obtain the stiffness properties of the structure in the direction of
the ground acceleration. A rigid diaphragm scheme is employed; therefore, the
frames in that direction will have compatible lateral displacements. Since the
three frames in that direction have the same properties, once the lateral stiffness of
one frame is obtained it should be multiplied by three to obtain the lateral
stiffness of the whole structure in the direction of interest. The frame stiffness
matrix is modified to eliminate any axial deformations of the girders (to comply
with the rigid diaphragm condition), and the vertical deformations and joint
rotations are condensed. After performing all these operations, the lateral-load
The area of each floor slab is 12 m · 12 m = 144 m2. The total translational mass of
each story is m = 144 m2· 1780 kg/m2 = 256 Mg. The mass matrix of the buildings is:
↓ dof
256 0 0 0 0 0 U6
0 256 0 0 0 0
U5
0 0 256 0 0 0 U4
[ M ] = 0 0 0 256 0 0 U3
0 0 0 0 256 0 U2
0 0 0 0 0 256 U1
Matrix [γ] is in this case a single column vector having one in all rows, because all
the lateral degrees of freedom of the structure are parallel to the ground motion
acceleration. The dynamic equilibrium equations are:
256 0 0 0 0 0 U
6
0 256 0
0 0 0 U
5
0 0 256 0 0 0 U
4
0 0 0 256 0 0 U3
0 0 0 0 256 0 U
2
0 0 0 0
0 256 U
1
216.76 −306.77 105.49 −19.561 4.2822 −0.51088 U 6 1
−306.77 668.24 −475.14 137.94 −29.375
5.3857 U 5
1
105.49 −475.14 731.37 −493.23 159.60 −29.327 U 4 1
+103 × = − [ M ] x0
−19.561 137.94 −493.23 749.02 −494.47 145.71 U 3 1
4.2822 −29.375 159.60 −494.47 738.11 −515.90 U 2 1
−0.51088 5.3857 −29.327 145.71 −515.90 889.94 U1 1
Mode ω2 ω f T
(rad/s)² (rad/s) (Hertz) (s)
1 29.108 5.39 0.859 1.16
2 301.81 17.4 2.76 0.36
3 973.78 31.2 4.97 0.20
4 2494.3 49.9 7.95 0.13
5 4686.5 68.5 10.9 0.092
6 7113.8 84.3 13.4 0.075
6 6 6 6 6 6
5 5 5 5 5 5
4 4 4 4 4 4
3 3 3 3 3 3
2 2 2 2 2 2
1 1 1 1 1 1
0 0 0 0 0 0
-0.04 0.00 0.04 -0.04 0.00 0.04 -0.04 0.00 0.04 -0.04 0.00 0.04 -0.04 0.00 0.04 -0.04 0.00 0.04
Mode 1 Mode 2 Mode 3 Mode 4 Mode 5 Mode 6
(T1 = 1.16 s) (T2 = 0.36 s) (T3 = 0.20 s) (T4 = 0.13 s) (T5 = 0.092 s) (T6 = 0.075 s)
Fig. 10 - Example 5 - Structure vibration periods and modes
34.970
13.540
8.2331
{α} = [ Φ ] [ M ][ γ ] = 6.0279
T
4.4695
2.3861
In all six equations ξi = 0.05. Response of each one of the uncoupled equations is
obtained using Newmark’s Beta method. The first 10 seconds of response are
shown in the following graphs.
5 0.3 max
max
4
3 0.2
2
0.1
1
η 1 η 2
0 t (s) 0.0 t (s)
(m) 0 1 2 3 4 5 6 7 8 9 10 (m) 0 1 2 3 4 5 6 7 8 9 10
-1
-2 -0.1
-3
-0.2
-4 min
-5 -0.3 min
(a) response for η1 (T1 = 1.16 s) (b) response for η2 (T2 = 0.36 s)
(e) response for η5 (T5 = 0.092 s) (f) response for η6 (T6 = 0.075 s)
The following table contains the response at selected instants, and the extreme
values obtained for each uncoupled degree of freedom during the first 10 s of
response.
t η1 η2 η3 η4 η5 η6
(s) (m) (m) (m) (m) (m) (m)
2.12 -2.005521 0.120653 0.022884 0.007405 0.003448 0.001047
2.16 -1.454167 0.131596 0.020979 0.009887 0.002831 0.001150
2.22 -0.313056 0.115723 0.009855 -0.005804 -0.002937 -0.001495
2.24 0.044580 0.071039 -0.010748 -0.012843 -0.004919 -0.001187
2.52 2.597408 0.177460 -0.054570 -0.012759 -0.000935 0.000215
2.58 2.305620 0.044364 0.008234 0.010581 -0.001491 0.000439
2.64 1.678734 -0.155214 0.022238 -0.017115 -0.001112 -0.000662
3.04 -3.664644 -0.169206 0.038901 -0.001547 0.000665 0.000280
3.08 -3.545856 -0.153292 -0.001900 0.003579 -0.001228 -0.000357
3.22 -1.871672 0.183567 0.047073 0.005055 0.001766 0.000137
4.58 2.840147 -0.284971 -0.010291 -0.001032 -0.001448 -0.000172
4.76 1.448789 0.295191 -0.005745 -0.002603 0.000228 0.000605
5.90 4.049463 0.020068 0.022552 -0.001531 -0.000257 -0.000036
max 4.049463 0.295191 0.047073 0.010581 0.003448 0.001150
min -3.664644 -0.284971 -0.054570 -0.017115 -0.004919 -0.001495
{U } = {φ } η (t)
(i ) (i )
i
Displacements of the structure for the same instant t, are obtained as the sum of
the individual contributions of each mode from:
{U} = [ Φ ]{η(t)}
To find the forces imposed by the ground motions at the same instant t = 3.08 s,
the stiffness matrix of the structure is multiplied by the displacement just
obtained for that instant:
{F} = [K E ]{U}
This operation can be made for each mode independently, thus obtaining the
contribution of the total internal forces caused by each one:
{F } = [K ] U
mod
E
(1)
U (2) " U (6) = F (1) F (2) " F (6)
−658.49
−636.50
−920.33
{F} = [ K E ]{U} = −1041.90
−663.43
−308.29
Base shear contributed by each mode, also in kN, at instant t = 3.08 s, is obtained
from:
−658.49
−636.50
−920.33
V = {1} {F} = {1 1 1 1 1 1}
T
= −4229.0
−1041.90
−663.43
−308.29
−658.49
−636.50
− 920.33
M = {h} {F} = {18 15 12 9 6 3}
T
= −46727
−1041.90
−663.43
−308.29
The same procedures can be used to obtain the response at any instant. If this is
performed systematically, results such as shown in Fig. 12 are obtained. There the
displacement response for the roof of the building is shown for the first 15 sec. of
the to the NS component of El Centro record. From this figure, it is evident that the
significant portion of the response is contributed solely by the first two modes,
with se second contributing marginally.
Roof displacements
0.15
max
max = 0.14870 m
0.10
(1)
U6 0.05
Mode 1 0.00 t (s)
(m) -0.05 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
-0.10
min
min = -0.13456 m
-0.15
0.15
0.10
(2) 0.05
U6 max max = 0.0093399 m
Mode 2 0.00 min t (s)
(m) -0.05 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
min = -0.0096749 m
-0.10
-0.15
0.15
0.10
(3) 0.05
U6 max max = 0.0013730 m
0.00 t (s)
Mode 3 min
(m) -0.05 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
min = -0.0015917 m
-0.10
-0.15
0.15
0.10
(4) 0.05
U6 max max = 0.00035371 m
Mode 4
0.00 min t (s)
(m) -0.05 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
min = -0.00021867 m
-0.10
-0.15
0.15
0.10
(5) 0.05
U6 max max = 0.00004500 m
0.00 t (s)
Mode 5 min
(m) -0.05 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
min = -0.00006419 m
-0.10
-0.15
0.15
0.10 max = 0.0000089 m
(6) 0.05
U6 max
Mode 6 0.00 min t (s) min = -0.0000068 m
(m) -0.05 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
-0.10
-0.15
0.15 max max = 0.148729 m
0.10
Total U6 0.05
0.00 t (s)
(m)
-0.05 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
-0.10
min min = -0.128367 m
-0.15
Fig. 12 - Example 5 - Roof displacements from each mode and total response
Fig. 13 shows the variation of the base shear of the building during the first 15
sec. of response to the NS component of El Centro record.
Base shear
Fig. 14 shows the variation of overturning moment for the first 15 sec. of response
to the NS component of El Centro record.
Overturning moment
Example 6
Lets rework Example 5 using the displacement response spectra of the El Centro
record.
The results are the same up to the point where the dynamic equilibrium
equations were uncoupled.
The response for each of the uncoupled equations is obtained using the
displacement response spectra for the N-S component of the El Centro record. En la
Fig. 15 shows the spectrum and period fro each mode and the displacement read
from the spectrum for each period.
Damping ξ = 0.05
0.20
0.15
0.116 m
Sd
0.10
(m)
0.05
0.0218 m
0.00674 m
0.00285 m
0.00113 m 0.00
0.000720 m 0.0 0.5 1.0 1.5 2.0
Period T (s)
T T2 T1
T6 T5 T4 3
Fig. 15 - Example 6 - Displacement response spectrum for El Centro NS record
Mode Ti Sd(Ti,ξi)
(s) (m)
1 1.16 0.116
2 0.36 0.0218
3 0.20 0.00674
4 0.13 0.00285
5 0.092 0.00113
6 0.075 0.000720
Table 2 - Example 6 - Maximum displacement values for the uncoupled degrees of freedom
Mode
Sd(Ti,ξi) ( ηi )max = α i × Sd (Ti , ξ i )
αi
(m) (m)
1 34.970 0.116 4.0495
2 13.540 0.0218 0.29571
3 8.233 0.00674 0.055458
4 6.028 0.00285 0.017155
5 4.469 0.00113 0.0050639
6 2.386 0.000710 0.0017170
{U } = {φ } ( η )
(i )
mod
(i )
i max
These results can be computed for all the modes at the same time by introducing
the values of (ηi)max in the diagonal of a square matrix [Ηmod] y and performing
the operation:
( η1 )max 0 0 0 0 0
0 ( η2 )max 0 0 0 0
0 0 ( η3 )max 0 0 0
[ Η mod ] = 0 0 0 ( η4 )max 0 0
0 0 0 0 ( η5 )max 0
0
0 0 0 0 ( η6 )max
4.0495 0 0 0 0 0
0 0.29571 0 0 0 0
0 0 0.0055458 0 0 0
[ Η mod ] = 0 0 0 0.015155 0 0
0 0 0 0 0.0050639 0
0 0 0 0 0 0.001717
{U } (1)
mod {U }
(2)
mod {U }(3)
mod {U } (4)
mod {U }(5)
mod {U }
(6)
mod ↓ dof
0.148703 −0.009692 0.001618 −0.000355 0.000066 −0.000010 U6
0.136429 −0.003428 −0.000790 0.000557 −0.000163 0.000032
U5
0.115519 0.004295 −0.001915 0.000091 0.000144 −0.000050 U4
[ Umod ] = [Φ ][ Η mod ] = 0.084882 0.009854 −0.000280 −0.000592 −0.000017 0.000058
U3
0.049588 0.009914 0.001754 0.000118 −0.000124 −0.000054 U2
0.018061 0.004698 0.001397 0.000584 0.000181 0.000041 U1
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
0.00 0.05 0.10 0.15 -0.02 -0.01 0.00 0.01 0.02 -0.002 -0.001 0.000 0.001 0.002
Deflection (m) Deflection (m) Deflection (m)
mode 1 mode 2 mode 3
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
-0.0010 -0.0005 0.0000 0.0005 0.0010 -0.0002 -0.0001 0.0000 0.0001 0.0002 -0.00010 -0.00005 0.00000 0.00005 0.00010
Deflection (m) Deflection (m) Deflection (m)
mode 4 mode 5 mode 6
Fig. 16 - Example 6 - Maximum lateral displacements for each mode
Using the displacements just computed the story drift for each story and mode
could be computed as the algebraic difference of the displacement of two
consecutive stories. Drift is usually expressed as percentage of the inter-story
height.
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
0.00 0.50 1.00 1.50 -0.30 -0.20 -0.10 0.00 0.10 0.20 -0.10 -0.05 0.00 0.05 0.10
Drift (%h) Drift (%h) Drift (%h)
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
-0.04 -0.02 0.00 0.02 0.04 -0.015 -0.010 -0.005 0.000 0.005 0.010 -0.004 -0.002 0.000 0.002 0.004
Drift (%h) Drift (%h) Drift (%h)
To obtain the maximum modal lateral forces imposed on the structure by the
ground motions the stiffness matrix of the structure is multiplied by the modal
lateral displacements. Results are obtained in kN.
[ Fmod ] = [ K E ] Umod
(1) (2)
U mod (6)
" U mod = Fmod
(1) (2)
Fmod (6)
" Fmod
{F } {F } {F } {F } {F } {F }
(1)
mod
(2)
mod
(3)
mod
(4)
mod
(5)
mod
(6)
mod ↓ dof
1108.3 −748.9 403.3 −226.4 79.3 −18.6 F6
1016.2 −264.8 −196.9 355.8 −195.6 57.9
F5
860.2 331.8 −477.4 58.2 173.4 −91.0 F4
[ Fmod ] = [K E ][ Umod ] = 632.9 761.5 −69.8 −378.0 −20.2 105.1
F3
369.4 765.9 437.3 75.5 −148.2 −98.4 F2
135.1 363.0 348.2 372.7 217.3 74.1 F1
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
-1000 -500 0 500 1000 1500 -1000 -500 0 500 1000 1500 -1000 -500 0 500 1000 1500
Modal Forces Modal Forces Modal Forces
(kN) (kN) (kN)
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
-1000 -500 0 500 1000 1500 -1000 -500 0 500 1000 1500 -1000 -500 0 500 1000 1500
Modal Forces Modal Forces Modal Forces
(kN) (kN) (kN)
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
0 1500 3000 4500 -1500 -750 0 750 1500 -500 -250 0 250 500
Story shear Story shear Story shear
(kN) (kN) (kN)
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
-300 -150 0 150 300 -150 -75 0 75 150 -60 -30 0 30 60
Story shear Story shear Story shear
(kN) (kN) (kN)
It is the same value obtained for the first story when the story shears were
computed.
n
The overturning moment for each story is obtained from M (ij ) = ∑ (h
k = j+ 1
k − h j ) ⋅ Fj(i )
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
0 20000 40000 60000 -8000 -6000 -4000 -2000 0 -500 0 500 1000 1500 2000
Overturning moment (kN . m) Overturning moment (kN . m) Overturning moment (kN . m)
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
-1000 -500 0 500 -200 0 200 400 -100 -50 0 50 100
Overturning moment (kN . m) Overturning moment (kN . m) Overturning moment (kN . m)
This is the same result obtained for the overturning moment previously.
In Example 5 the step-by-step response of the building was obtained for the same
earthquake record used to compute the spectrum in this example, it is interesting
to make some comparisons of the results obtained in both cases. Table 6 lists the
values obtained in Example 5 and Example 6 for each of the uncoupled degrees of
freedom.
As can be seen, the results are essentially the same, and the differences obey to
precision rounding in the numerical procedures because the algorithm employed
to obtain the response is different from the one used to compute the spectrum. It
should be noted that the maximum values for each uncoupled degree of freedom
in Example 5 were obtained at different time instants It should also be noted that
the maximum value obtained from the spectrum in some cases correspond to the
maximum value and in some to the minimum obtained in the step-by-step
procedure, this is because the value carried by the spectrum is the absolute value.
The maximum lateral displacement of the roof obtained in Example 5 was 0.149
m. The algebraic sum of the values obtained for the 6 story in Example 6 is 0.140
th
m, and the sum of the absolute values is 0.160 m. The algebraic sum of the
modal response underestimates the value obtained using a time step-by-step
procedure and the sum of the absolute modal values overestimate it.
The maximum value for the base shear of the building obtained in Example 5
using a time step-by-step procedure was 4360 kN. The sum of the maximum
modal base shears obtained in Example 6 was 6170 kN. This value overestimates
the time step value by a factor of 1.4. In the time step procedure of Example 5 the
base shear is controlled by the first mode with the other modes contributing very
little when the first mode peak occurs. For the overturning moment at the base in
Example 5 a value of 54,400 kN · m was obtained In Example 6 the algebraic sum
of the maximum modal values is 54,800 kN · m, and the sum of the absolute
values is 56,690 kN · m. For the overturning moment, the contribution of the
higher modes is small in both examples.
Example 7
Apply the square root of the sum of the squares SRSS procedure to the results
obtained in Example 6. The use of the SRSS technique produces the following
results:
{U } = {φ } ( η )
(i )
mod
(i )
i max
as:
{U }(1)
mod {U }
( 2)
mod {U }
( 3)
mod {U }
(4)
mod {U }
(5)
mod {U }
(6)
mod ↓ dof
0.148703 −0.009692 0.001618 −0.000355 0.000066 −0.000010 U6
0.136429 −0.003428 −0.000790 0.000557 −0.000163 0.000032
U5
0.115519 0.004295 −0.001915 0.000091 0.000144 −0.000050 U4
[ U mod ] = [ Φ ][ Η mod ] = 0.084882 0.009854 −0.000280 −0.000592 −0.000017 0.000058
U3
0.049588 0.009914 0.001754 0.000118 −0.000124 −0.000054 U2
0.018061 0.004698 0.001397 0.000584 0.000181 0.000041 U1
We now apply the SRSS procedure to each of the row of previous matrix. For
example for the roof (6 story): th
This value compares fairly well with the value of 0.14873 m obtained from the
step-by-step procedure in Example 6. The result, in m, for all stories is:
↓ dof
0.14903 U6
0.13648
U5
0.11560 U4
{U } srss
= ±
0.08545 U3
0.05059 U2
0.01872 U1
A symbol ± has been introduced to remind that the values obtained from the SRSS
procedure may be either positive or negative.
The modal spectral story drifts are computed from the values shown in [Umod].
Using Eq. (113) the following result are obtained:
{∆ } (1)
mod {∆ }
(2)
mod {∆ }
( 3)
mod {∆ }
(4)
mod {∆ } (5)
mod {∆ } (6)
mod ↓ story
0.012274 −0.006264 0.002408 −0.000912 0.000229 −0.000042 6
0.020920 −0.007723 0.001125 0.000466 −0.000307 0.000082
5
0.030627 −0.005559 −0.001635 0.000683 0.000161 −0.000108 4
[ ∆ mod ] = 0.035304 −0.000060 −0.002034 −0.000710 0.000107 0.000112
3
0.031517 0.005216 0.000358 −0.000465 −0.000305 −0.000095 2
0.018061 0.004698 0.001397 0.000584 0.000181 0.000041 1
↓ story
0.0140 0.47%h 6
0.0223 0.74%h
5
0.0312 1.04%h 4
{∆ }
srss
=
0.0354 m = 1.18%h
3
0.0320 1.07%h 2
0.0188 0.62%h 1
Now, for the sake of discussion, lets compute erroneously the story drift from
lateral displacements already combined, {Usrss}. The following are the results for
story drift as a percentage of the story height (%h) thus computed:
↓ story
0.42%h 6 ⇐ wrong result
0.70%h
5 ⇐ wrong result
1.00%h 4 ⇐ wrong result
{∆} = 1.16%h
3 ⇐ wrong result
1.06%h 2
0.62%h 1
The maximum modal spectral forces were obtained for each mode in Example 6
multiplying the stiffness matrix by the modal spectral displacements of each
mode, obtaining there the following forces in kN:
[ Fmod ] = [ K E ] Umod
(1) ( 2)
U mod (6)
" U mod = Fmod
(1) ( 2)
Fmod (6)
" Fmod
{F } {F } {F } {F } {F } {F }
(1)
mod
(2)
mod
( 3)
mod
(4)
mod
(5)
mod
(6)
mod ↓ dof
1108.3 −748.9 403.3 −226.4 79.3 −18.6 F6
1016.2 −264.8 −196.9 355.8 −195.6 57.9
F5
860.2 331.8 −477.4 58.2 173.4 −91.0 F4
[Fmod ] = [K E ][ Umod ] = 632.9 761.5 −69.8 −378.0 −20.2 105.1
F3
369.4 765.9 437.3 75.5 −148.2 −98.4 F2
135.1 363.0 348.2 372.7 217.3 74.1 F1
The maximum credible modal spectral story shear may be obtained from Eq.
(115)
p
V (i )
j = ∑ Fk(i )
k=j
Applying, for example, the SRSS procedure to the second story, we obtain:
↓ story
1417.6 6
2369.8
5
3080.3 4
{V }srss
= ±
3640.1 3
4080.2 2
4327.6 1
The base shear, in kN, was obtained in Example 6 for each mode as:
1108.3 −748.9 403.3 −226.4 79.3 −18.6
1016.2 −264.8 −196.9 355.8 −195.6 57.9
860.2 331.8 −477.4 58.2 173.4 −91.0
{Vmod } = {1} [ Fmod ] = {1 1 1 1 1 1} 632.9 761.5 −69.8 −378.0 −20.2
T
105.1
369.4 765.9 437.3 75.5 −148.2 −98.4
135.1 363.0 348.2 372.7 217.3 74.1
= {4122.1 1208.5 444.6 257.9 106.1 29.1}
(1) ( 2) ( 3) (4) (5) (6)
Vmod Vmod Vmod Vmod Vmod Vmod
The overturning moment for each story and mode is obtained using Eq. (119):
n
M (ij ) = ∑ ( h
k = j+ 1
k − h j ) ⋅ Fj(i )
Now using the SRSS procedure for example to the fourth story:
↓ story
0.0 6
4252.9
5
11201.3 4
{M }srss
= 20070.6 3
30348.8 2
41722.9 1
53865.8
0
and
( 53833.1) + ( −932.7 ) + (1616.3 ) + ( 130.7 ) + ( 155.3 ) + ( 19.2 )
2 2 2 2 2 2
M srss =
= 53865.8 kN ⋅ m
These forces, in kN, are computed using Eq. (123) using the story shears obtained
by using the SRSS procedure:
↓ story
1417.6 1417.6 6
2369.8 − 1417.6 951.9
5
3080.3 − 2369.8 710.5 4
{F }
E
=
3640.1 − 3080.3 = ± 559.8
3
4080.2 − 3640.1 440.3 2
4327.6 − 4080.2 247.6 1
The overturning moment, in this case, is slightly larger than the one obtained
using the SRSS procedure with the modal spectral overturning moments.
In Example 5 the step-by-step response of the system to the El Centro record was
computed, in Example 6 the individual modal spectral responses were computed
for the spectrum of the same record — thus permitting the computation of the
absolute maximum spectral response —, and in Example 7 the SRSS procedure was
applied to the results obtained in Example 6. Now some comparisons can be
made between the results of the three examples.
For this case the match between the step-by step analysis values and the values
obtained using the SRSS procedure is reasonable good.