Detailed derivation of analytical equations used for

response spectrum generation

Computers & Structures, Inc.
October 12, 2009
1 Preliminaries
Equation of motion for a single degree of freedom can be written as
m u +c u +ku = m u
g
(t) (1)
where u is displacement relative to the ground [3], [4].
Eqn. (1) can be written as modal equation
u + 2 u +
2
u = u
g
(t) (2)
where
=

k
m
(3)
and
=
c
2

km
< 1 (4)
Ground acceleration u
g
is assumed to be linear over a time step i:
u
g
(t) = u
g
(t
i1
) +
u
g
(t
i
) u
g
(t
i1
)
t
i
t
i1
(t t
i1
) (5)
Lets dene slope, s, of the acceleration within a time step as
s =
u
g
(t
i
) u
g
(t
i1
)
t
i
t
i1
(6)
and time step independent time variable, t, as
t = t t
i1
(7)
For each time step, Equations (2), (5), (6) and (7) can be combined into
the following dierential equation for u(t)
1
Response Spectrum Generation Computers & Structures, Inc.
u(t) + 2 u(t) +
2
u(t) = u
g
(t
i1
) st (8)
with initial conditions
u(t)
t=0
= u(t
i1
) (9)
u(t)
t=0
= u(t
i1
) (10)
The solution of Equation (8) with initial conditions (9) and (10) will be
obtained as a sum of homogeneous and particular solution
u = u
h
+u
p
(11)
2 Homogeneous Solution
2.1 General Solution
The solution of homogeneous second-order linear ordinary dierential equa-
tion with constant coecients
y

+ay

+by = 0 (12)
where a, b are real constants is readily available in the literature [1].
Let
1
and
2
be the roots of the characteristic equation
f() =
2
+a +b = 0 (13)
1. If
1
=
2
real, then
y = C
1
e

1
x
+C
2
e

1
x
(14)
2. If =
1
=
2
, then
y = (C
1
x +C
2
)e
x
(15)
3. If
1
= +i,
2
= i ( = 0), then
y = e
x
(C
1
cos(x) +C
2
sin(x)) = e
x
C cos(x +) (16)
2 bm, ok, October 12, 2009
Response Spectrum Generation Computers & Structures, Inc.
2.2 Problem Specic Solution
By comparison of Eqns. (1) and (12), the generic a, b coecients in Eqn. (12)
can we written in the problem specic variables and as
a = 2 (17)
b =
2
(18)
Then
a
2
4b = (2)
2
4
2
= 4
2
(
2
1) = 4
2
(1
2
) = 4
d
2
< 0 (19)
where

d
=
_
1
2
(20)
The solution of characteristic equation (13) has complex roots

1
=
2 + 2i
_
1
2
2
= +i
d
(21)

2
= i
d
(22)
and from the general solution can be rewritten using the problem
specic variables as
= (23)
=
d
(24)
Substituting Eqns. (23) and (24) into Eqn. (16) yields
u
h
(t) = e
t
(C
1
cos(
d
t) +C
2
sin(
d
t)) (25)
The constants C
1
and C
2
will be determined from the initial conditions.
3 Particular Solution
3.1 General Solution
Particular solution, y
p
, of dierential equation
y

+ay

+by = R(x) (26)

where a, b are real constants is also readily available in the literature [1].
For the special case of R(x) = P(x) = Ax
2
+Bx +C and b = 0
3 bm, ok, October 12, 2009
Response Spectrum Generation Computers & Structures, Inc.
y
p
=
1
b
_
P(x)
a
b
P

(x) +
a
2
b
b
2
P

(x)
_
(27)
For our case, A = 0, therefore
P(x) = Bx +C (28)
P

(x) = B (29)
P

(x) = 0 (30)
and Eqn. (27) reduces to
y
p
=
1
b
_
Bx +C
a
b
B
_
(31)
3.2 Problem Specic Solution
The a, b, B, and C constants can be written using problem specic notation
(see Eqns. (8), (17) and (18)) as
a = 2 (32)
b =
2
(33)
B = s (34)
C = u
g
(t
i1
) (35)
Then the general particular solution described by Eqn. (27) becomes
u
p
(t) =
1

2
_
st + ( u(t
i1
))
2

2
(s)
_
=
s

2
t +
1

2
_
2

s u
g
(t
i1
)
_
= Et +F (36)
where
E =
s

2
(37)
F =
1

2
_
2

s u
g
(t
i1
)
_
(38)
4 bm, ok, October 12, 2009
Response Spectrum Generation Computers & Structures, Inc.
4 Complete Solution
Eqns. (25) and (36) can be added to obtain complete solution as
u(t) = u
h
(t) +u
p
(t)
= e
t
(C
1
cos(
d
t) +C
2
sin(
d
t)) +Et +F (39)
The unknown coecients C
1
and C
2
will be determined from initial
conditions.
Derivative of Eqn. (39) is
u(t) = e
t
(C
1
cos(
d
t) +C
2
sin(
d
t)) +
e
t
(C
1

d
(sin(
d
t)) +C
2

d
cos(
d
t)) +E (40)
Evaluating Eqn. (39) and (40) for initial conditions (9) and (10) yields
u(t
i1
) = e
0
(C
1
cos 0 +C
2
sin0) +F (41)
u(t
i1
) = e
0
(C
1
cos 0 +C
2
sin0) +
e
0
(C
1

d
(sin0) +C
2

d
cos 0) +E (42)
Simplify the above as
u(t
i1
) = C
1
+F (43)
u(t
i1
) = C
1
+C
2

d
+E (44)
Finally, express C
1
and C
2
C
1
= u(t
i1
) F (45)
C
2
=
u(t
i1
) +C
1
E

d
(46)
4.1 Summary
The nal solution is
u(t) = e
t
(C
1
cos(
d
t) +C
2
sin(
d
t)) +Et +F (47)
where
5 bm, ok, October 12, 2009
Response Spectrum Generation Computers & Structures, Inc.
C
1
= u(t
i1
) F (48)
C
2
=
u(t
i1
) +C
1
E

d
(49)
E =
s

2
(50)
F =
1

2
_
2

s u
g
(t
i1
)
_
(51)
s =
u
g
(t
i
) u
g
(t
i1
)
t
i
t
i1
(52)

d
=
_
1
2
(53)
t = t t
i1
(54)
5 Computer Algorithm
The formulation described by Eqn. (39) is applied recursively for each step,
i = 1 to n. In order to apply the algorithm and to obtain acceleration
response spectrum, we will also need to calculate velocity and acceleration
by obtaining rst and second derivatives of Eqn. (39).
5.1 Velocity
u(t) = e
t
(C
1
cos(
d
t) +C
2
sin(
d
t)) +
e
t
(C
1

d
(sin(
d
t)) +C
2

d
cos(
d
t)) +E (55)
5.2 Acceleration
u(t) =
2

2
e
t
(C
1
cos(
d
t) +C
2
sin(
d
t))
e
t
(C
1

d
(sin(
d
t)) +C
2

d
cos(
d
t))
e
t
(C
1

d
(sin(
d
t)) +C
2

d
cos(
d
t)) +
e
t
_
C
1

d
2
(cos(
d
t)) +C
2

d
2
(sin(
d
t))
_
=
2

2
e
t
(C
1
cos(
d
t) +C
2
sin(
d
t))
2e
t
(C
1

d
(sin(
d
t)) +C
2

d
cos(
d
t)) +
e
t
_
C
1

d
2
(cos(
d
t)) +C
2

d
2
(sin(
d
t))
_
(56)
5.3 Algorithm
The ground acceleration time history is given as a vector of acceleration
values that are uniformly spaced in time as shown in Fig. 1.
6 bm, ok, October 12, 2009
Response Spectrum Generation Computers & Structures, Inc.
i =
t = t t t t
1 2 3 n
1 2 3
n t
u
g
t
Figure 1: Measured ground acceleration discretized by linear segments
In the rst solution step, the initial displacement and velocity at time
t
1
= 0 are taken to be zero. Eqns. (47), (55) and (56) are applied to obtain
displacement, velocity, and acceleration the time t
2
.
In the subsequent steps, the displacement and velocity calculated for
the previous step are applied as initial conditions for the current step and
displacement, velocity, and acceleration at the end of the step are calculated.
After repeating the above procedure for all n steps, time histories of
displacements, velocities and accelerations are obtained. Spectral values are
determined as follows [2]:
Spectral displacement, S
d
Maximum absolute displacement obtained from
the time history.
Spectral velocity, S
v
Maximum absolute velocity obtained from the time
history.
Spectral acceleration, S
a
Maximum absolute acceleration obtained from
the time history.
Spectral pseudo-velocity, S
pv
Calculated from spectral displacement as
S
pv
= S
d
Spectral pseudo-acceleration, S
pa
Calculated from spectral displacement
as S
pa
=
2
S
d
References
[1] F. Bubenk, M. Pultar, and I. Pultarova. Matematicke vzorce a metody.
Vydavatelstv

CVUT, second edition, 1997. p. 224 (in Czech).
7 bm, ok, October 12, 2009
Response Spectrum Generation Computers & Structures, Inc.
[2] Anil K. Chopra. Dynamics of Structures, Theory and Applications to
Earthquake Engineering. Prentice Hall of India, New Delhi, 2nd edition,
2002. section 6-6, p. 208.
[3] Ray W. Clough and Joseph Penzien. Dynamics of Structures. McGraw-
Hill, New York, 2nd edition, 1993. 738 pages, ISBN 0-07-011394-7, sec-
tion 2-6, p. 26.
[4] Structural Wiki. Single degree of freedom damped free vi-
brations. Available at http://www.structuralwiki.org/en/ Sin-
gle degree of freedom damped free vibrations. accessed October 2009.
8 bm, ok, October 12, 2009