Generalized Single-Degree-of-Freedom Systems: Chopra: Prentice-Hall PAGES JUL. 19, 2000 14:21 ICC Oregon (503) 221-9911
Generalized Single-Degree-of-Freedom Systems: Chopra: Prentice-Hall PAGES JUL. 19, 2000 14:21 ICC Oregon (503) 221-9911
Generalized Single-Degree-of-Freedom Systems: Chopra: Prentice-Hall PAGES JUL. 19, 2000 14:21 ICC Oregon (503) 221-9911
8
Generalized
Single-Degree-of-Freedom
Systems
PREVIEW
Consider, for example, the system of Fig. 8.1.1a, consisting of a rigid, massless bar sup-
ported by a hinge at the left end with two lumped masses, a spring and a damper, attached
to it, subjected to a time-varying external force p(t). Because the bar is rigid, its deflections
can be related to a single generalized displacement z(t) through a shape function ψ(x) as
305
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 306
p(t)
m1 m2 EI(x), m(x)
(a) (b)
c k
u(x,t) = ψ(x)z(t)
u(x,t) = xz(t)
z(t)
z(t)
x x
known, the displacements at all locations of the system are determined from Eq. (8.1.1).
This analysis procedure leads to the exact results for the system of Fig. 8.1.1a because the
shape function ψ(x) could be determined exactly but provides only approximate results
for the system of Fig. 8.1.1b because they are based on an assumed shape function.
The key step in the analysis outlined above that is new is the evaluation of the gen-
eralized properties m̃, c̃, k̃, and p̃(t) for a given system. Procedures are developed to
determine these properties for (1) assemblages of rigid bodies that permit exact evaluation
of the deflected shape (Section 8.2), and (2) multi-degree-of-freedom systems with dis-
tributed mass or several lumped masses which require that a shape function be assumed
that satisfies the displacement boundary conditions (Sections 8.3 and 8.4).
In this section the equation of motion is formulated for an assemblage of rigid bodies
with distributed mass supported by discrete springs and dampers subjected to time-varying
forces. In formulating the equation of motion for such generalized SDF systems, appli-
cation of Newton’s second law of motion can be cumbersome, and it is simpler to use
D’Alembert’s principle and include inertia forces in the free-body diagram. The dis-
tributed inertia forces for a rigid body with distributed mass can be expressed in terms of
the inertia force resultants at the center of gravity using the total mass and the moment of
inertia of the body. These properties for rigid plates of three configurations are presented in
Appendix 8.
Example 8.1
The system shown in Fig. E8.1a consists of a rigid bar supported by a fulcrum at O, with
an attached spring and damper subjected to force p(t). The mass m 1 of the part O B of the
bar is distributed uniformly along its length. The portions O A and BC of the bar are mass-
less, but a uniform circular plate of mass m 2 is attached at the midpoint of BC. Selecting
the counterclockwise rotation about the fulcrum as the generalized displacement and consid-
ering small displacements, formulate the equation of motion for this generalized SDF system,
determine the natural vibration frequency and damping ratio, and evaluate the dynamic re-
sponse of the system without damping subjected to a suddenly applied force po . How would
the equation of motion change with an axial force on the horizontal bar; what is the buckling
load?
Solution
1. Determine the shape function. The L-shaped bar rotates about the fulcrum at O.
Assuming small deflections, the deflected shape is shown in Fig. E8.1b.
2. Draw the free-body diagram and write the equilibrium equation. Figure. E8.1c
shows the forces in the spring and damper associated with the displacements of Fig. E8.1b,
together with the inertia forces. Setting the sum of the moments of all forces about O to zero
gives
L L L L L L 3L 3L L
I1 θ̈ + m 1 θ̈ + I2 θ̈ +(m 2 L θ̈)L + m 2 θ̈ + c θ̇ + k θ = p(t)
2 2 4 4 2 2 4 4 2
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 308
u(x′,t) = x′θ(t)
x′
u(x,t) = xθ(t)
θ θL
massless O
x
C
p(t)
•
mass = m2 L/4
L/8 (b)
•
3L/4
massless • • L/4
A O mass = m1
p(t)
•
c B Q
k
I2θ̈
L/2 L m2(L/4)θ¨
• • • Q
I1θ̈
O m2(Lθ¨)
k(3L/4)θ
(a) c(L/2)θ˙ m1(L/2)θ¨
(c)
Figure E8.1
where
m1 137 cL 2 9k L 2 L
m̃ = + m2 L 2 c̃ = k̃ = p̃(t) = p(t) (c)
3 128 4 16 2
z
δz
•
u
δu fo(ξ)
L fI (x,t)
EI(x) x
m(x) ξ
x
•
ug
Figure 8.3.1 (a) Tower deflections and virtual displacements; (b) inertia forces; (c) equiv-
alent static forces.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 310
We assume that the displacement relative to the ground can be expressed by Eq. (8.1.1).
The total displacement of the tower is
u t (x, t) = u(x, t) + u g (t) (8.3.1)
The shape function ψ(x) in Eq. (8.1.1) must satisfy the displacement boundary con-
ditions (Fig. 8.3.1a). For this tower, these conditions at the base of the tower are ψ(0) = 0
and ψ (0) = 0. Within these constraints a variety of shape functions could be chosen. One
possibility is to determine the shape function as the deflections of the tower due to some
static forces. For example, the deflections of a uniform tower with flexural rigidity E I due
to a unit lateral force at the top are u(x) = (3L x 2 − x 3 )/6E I . If we select the generalized
coordinate as the deflection of some convenient reference point, say the top of the tower,
then z = u(L) = L 3 /3E I , and
3 x2 1 x3
u(x) = ψ(x)z ψ(x) = 2
− (8.3.2a)
2L 2 L3
This ψ(x) automatically satisfies the displacement boundary conditions at x = 0 because
it was determined from static analysis of the system. The ψ(x) of Eq. (8.3.2a) may also
be used as the shape function for a nonuniform tower, although it was determined for a
uniform tower. It is not necessary to select the shape function based on deflections due to
static forces, and it could be assumed directly; possibilities are
x2 πx
ψ(x) =
2
and ψ(x) = 1 − cos (8.3.2b)
L 2L
The three shape functions above have ψ(L) = 1, although this is not necessary. The ac-
curacy of the generalized SDF system formulation depends on the assumed shape function
ψ(x) in which the structure is constrained to vibrate. This issue will be discussed later
together with how to select the shape function.
We now proceed to formulate the equation of motion for the tower. At each time instant
the system is in equilibrium under the action of the internal resisting bending moments and
the fictitious inertia forces (Fig. 8.3.1b), which by D’Alembert’s principle are
f I (x, t) = −m(x)ü t (x, t)
Substituting Eq. (8.3.1) for u t gives
f I (x, t) = −m(x)[ü(x, t) + ü g (t)] (8.3.3)
The equation of dynamic equilibrium of this generalized SDF system can be formulated
conveniently only by work or energy principles. We prefer to use the principle of virtual
displacements. This principle states that if the system in equilibrium is subjected to vir-
tual displacements δu(x), the external virtual work δW E is equal to the internal virtual
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 311
work δW I :
δW I = δW E (8.3.4)
The external virtual work is due to the forces f I (x, t) acting through the virtual displace-
ments δu(x):
L
δW E = f I (x, t)δu(x) d x
0
The internal virtual work is due to the bending moments M(x, t) acting through the cur-
vature δκ(x) associated with the virtual displacements:
L
δW I = M(x, t)δκ(x) d x
0
Substituting
M(x, t) = E I (x)u (x, t) δκ(x) = δ[u (x)]
where u ≡ ∂ 2 u/∂ x 2 gives
L
δW I = E I (x)u (x, t)δ[u (x)] d x (8.3.6)
0
The internal and external virtual work is expressed next in terms of the generalized
coordinate z and shape function ψ(x). For this purpose, from Eq. (8.1.1) we obtain
u (x, t) = ψ (x)z(t) ü(x, t) = ψ(x)z̈(t) (8.3.7)
The virtual displacement is selected consistent with the assumed shape function
(Fig. 8.3.1a), giving Eq. (8.3.8a), and the virtual curvature is defined by Eq. (8.3.8b):
δu(x) = ψ(x) δz δ[u (x)] = ψ (x) δz (8.3.8)
Substituting Eqs. (8.3.7b) and (8.3.8a) in Eq. (8.3.5) gives
L L
δW E = −δz z̈ m(x)[ψ(x)] d x + ü g (t)
2
m(x)ψ(x) d x (8.3.9)
0 0
Having obtained the final expressions for δW E and δW I , Eq. (8.3.4) gives
where
L
m̃ = m(x)[ψ(x)]2 d x
0
L
k̃ = E I (x)[ψ (x)]2 d x (8.3.12)
0
L
L̃ = m(x)ψ(x) d x
0
Because Eq. (8.3.11) is valid for every virtual displacement δz, we conclude that
m̃ z̈ + k̃z = − L̃ ü g (t) (8.3.13a)
This is the equation of motion for the tower assumed to deflect according to the shape func-
tion ψ(x). For this generalized SDF system, the generalized mass m̃, generalized stiffness
k̃, and generalized excitation − L̃ ü g (t) are defined by Eq. (8.3.12). Dividing Eq. (8.3.13a)
by m̃ gives
z̈ + 2ζ ωn ż + ωn2 z = −˜ ü g (t) (8.3.13b)
where ωn2 = k̃/m̃ and a damping term using an estimated damping ratio ζ has been in-
cluded. This equation is the same as Eq. (6.2.1) for an SDF system, except for the factor
L̃
˜ = (8.3.14)
m̃
Once the generalized properties m̃ and k̃ are determined, the system can be analyzed by
the methods developed in preceding chapters. In particular, the natural vibration frequency
of the system is given by
L
k̃ E I (x)[ψ (x)]2 d x
ωn =
2
= 0 L (8.3.15)
m̃ m(x)[ψ(x)]2 d x
0
The generalized coordinate response z(t) of the system to specified ground acceleration
can be determined by solving Eq. (8.3.13b) using the methods presented in Chapters 5 and
6. Equation (8.1.1) then gives the displacements u(x, t) of the tower relative to the base.
The next step is to compute the internal forces—bending moments and shears—in
the tower associated with the displacements u(x, t). The second of the two methods de-
scribed in Section 1.8 is used if we are working with deflected shape ψ(x) that is assumed
and not exact, as for generalized SDF systems. In this method internal forces are com-
puted by static analysis of the structure subjected to equivalent static forces. Denoted by
f S (x), these forces are defined as external forces that would cause displacements u(x).
Elementary beam theory gives
f S (x) = E I (x)u (x) (8.3.16)
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 313
The equivalent static forces of Eq. (8.3.19) are used to determine the peak values of the
response of a tower to earthquake ground motion. Comparing Eq. (8.3.13b) to Eq. (6.2.1)
for an SDF system and using the procedure of Section 6.7 gives the peak value of z(t):
˜
z o = ˜ D = A (8.3.20)
ωn2
where D and A are the deformation and pseudo-acceleration ordinates, respectively, of the
design spectrum at period Tn = 2π/ωn for damping ratio ζ . In Eqs. (8.1.1) and (8.3.19),
z(t) is replaced by z o of Eq. (8.3.20) to obtain the peak values of displacements and equiv-
alent static forces:
u o (x) = ˜ Dψ(x) ˜
f o (x) = m(x)ψ(x)A (8.3.21)
where the conventional subscript S has been dropped from f So for brevity.
The internal forces—bending moments and shears—in the cantilever tower are ob-
tained by static analysis of the structure subjected to the forces f o (x); see Fig. 8.3.1c. Thus
the shear and bending moment at height x above the base are
L L
Vo (x) = ˜
f o (ξ ) dξ = A m(ξ )ψ(ξ ) dξ (8.3.22a)
x x
L L
Mo (x) = (ξ − x) f o (ξ ) dξ = ˜ A (ξ − x)m(ξ )ψ(ξ ) dξ (8.3.22b)
x x
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 314
In particular the shear and bending moment at the base of the tower are
Vbo = Vo (0) = L̃ ˜ A Mbo = Mo (0) = L̃ θ ˜ A (8.3.23)
where L̃ was defined in Eq. (8.3.12) and
L
θ
L̃ = xm(x)ψ(x) d x (8.3.24)
0
This completes the approximate evaluation of the earthquake response of a system with
distributed mass and flexibility based on an assumed shape function ψ(x).
If the excitation were external forces p(x, t) instead of ground motion ü g (t), the equation
of motion could be derived following the methods of Section 8.3.2, leading to
m̃ z̈ + k̃z = p̃(t) (8.3.25)
where the generalized force
L
p̃(t) = p(x, t)ψ(x) d x (8.3.26)
0
Observe that the only difference in the two equations (8.3.25) and (8.3.13a) is in the exci-
tation term.
Derivation 8.1
The equivalent static forces from elementary beam theory, Eq. (8.3.17), are written as
We seek lateral forces f˜S (x, t) that do not involve the derivatives of M(x, t) and at
each time instant are in equilibrium with the internal bending moments; equilibrium is sat-
isfied globally for the system (but not at every location x). Using the principle of virtual
displacements, the external work done by the unknown forces f˜S (x, t) in acting through the
virtual displacement δu(x) equals the internal work done by the bending moments acting
through the curvatures δκ(x) associated with the virtual displacements:
L L
f˜S (x, t)δu(x) d x = M(x, t)δκ(x) d x (c)
0 0
This equation is rewritten by substituting Eq. (8.3.8a) for δu(x) in the left side and by using
Eq. (8.3.10) for the integral on the right side; thus
L L
2
δz f˜S (x, t)ψ(x) d x = δz z(t) E I (x) ψ (x) dx (d)
0 0
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 315
Utilizing Eq. (8.3.15) and dropping δz, Eq. (d) can be rewritten as
L
f˜S (x, t) − ωn2 m(x)ψ(x)z(t) ψ(x) d x = 0 (e)
0
Setting the quantity in brackets to zero gives
f S (x, t) = ωn2 m(x)ψ(x)z(t) (f)
where the tilde above f S has now been dropped. This completes the derivation of Eq. (8.3.19).
Example 8.2
A uniform cantilever tower of length L has mass per unit length = m and flexural rigidity E I
(Fig. E8.2). Assuming that the shape function ψ(x) = 1 − cos(π x/2L), formulate the equa-
tion of motion for the system excited by ground motion, and determine its natural frequency.
Solution
1. Determine the generalized properties.
L
π x 2
m̃ = m 1 − cos d x = 0.227 mL (a)
0 2L
L 2
π2 πx EI
k̃ = E I cos2 d x = 3.04 3 (b)
0 4L 2 2L L
L
πx
L̃ = m 1 − cos d x = 0.363 mL (c)
0 2L
The computed k̃ is close to the stiffness of the tower under a concentrated lateral force at the
top.
2. Determine the natural vibration frequency.
k̃ 3.66 EI
ωn = = 2 (d)
m̃ L m
√
This approximate result is close to the exact natural frequency, wexact = (3.516/L 2 ) E I /m,
determined in Chapter 16. The error is only 4%.
3. Formulate the equation of motion. Substituting L̃ and m̃ in Eq. (8.3.14) gives ˜ = 1.6
and Eq. (8.3.13b) becomes
z̈ + ωn2 z = −1.6ü g (t) (e)
•
ψ(x) = 1 - cos πx
2L
L
x
m, EI
•
Figure E8.2
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 316
Example 8.3
A reinforced-concrete chimney, 600 ft high, has a uniform hollow circular cross section with
outside diameter 50 ft and wall thickness 2 ft 6 in. (Fig. E8.3a). For purposes of preliminary
earthquake analysis, the chimney is assumed clamped at the base, the mass and flexural rigidity
are computed from the gross area of the concrete (neglecting the reinforcing steel), and the
damping is estimated as 5%. The unit weight of concrete is 150 lb/ft3 and its elastic modulus
E c = 3600 ksi.
Assuming the shape function as in Example 8.2, estimate the peak displacements, shear
forces, and bending moments for the chimney due to ground motion characterized by the
design spectrum of Fig. 6.9.5 scaled to a peak acceleration 0.25g.
Solution
1. Determine the chimney properties.
Length: L = 600 ft
Cross-sectional area: A = π(252 − 22.52 ) = 373.1 ft2
150 × 373.1
Mass/foot length: m= = 1.738 kip-sec2 /ft2
32.2
π
Second moment of area: I = (254 − 22.54 ) = 105,507 ft4
4
Flexural rigidity: E I = 5.469 × 1010 kip-ft2
3.66 EI
ωn = = 1.80 rad/sec
L2 m
2π
Tn = = 3.49 sec
ωn
24.5″
•
uo(x)
x Section a-a fo(x) Vo(x) Mo(x)
600′
a a
•• 2′-6″
•
• • 0 1.5
50′ 0 3
103kips 106kip-ft
(a) (b) (c) (d) (e)
Figure E8.3
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 317
3. Determine the peak value of z(t). For Tn = 3.49 sec and ζ = 0.05, the design
spectrum gives A/g = 0.25(1.80/3.49) = 0.129. The corresponding deformation is D =
A/ωn2 = 15.3 in. Equation (8.3.20) gives the peak value of z(t):
z o = 1.6D = 1.6 × 15.3 = 24.5 in.
4. Determine the peak displacements u o (x) of the tower (Fig. E8.3b).
πx
u o (x) = ψ(x)z o = 24.5 1 − cos in.
2L
5. Determine the equivalent static forces.
πx
˜
f o (x) = m(x)ψ(x)A = (1.6)(1.738) 1 − cos 0.129g
2L
πx
= 11.58 1 − cos kips/ft
2L
These forces are shown in Fig. E8.3c.
6. Compute the shears and bending moments. Static analysis of the chimney subjected
to external forces f o (x) gives the shear forces and bending moments. The results using
Eq. (8.3.22) are presented in Fig. E8.3d and e. If we were interested only in the forces at
the base of the chimney, they could be computed directly from Eq. (8.3.23). In particular, the
base shear is
Vbo = L̃ ˜ A = (0.363m L)(1.6)0.129g
= 0.0749m Lg = 2518 kips
This is 7.49% of the total weight of the chimney.
Example 8.4
A simply supported bridge with a single span of L feet has a deck of uniform cross section
with mass m per foot length and flexural rigidity E I . A single wheel load po travels across the
bridge at a uniform velocity of v, as shown in Fig. E8.4. Neglecting damping and assuming
the shape function as ψ(x) = sin(π x/L), determine an equation for the deflection at midspan
as a function of time. The properties of a prestressed-concrete box-girder elevated-freeway
connector are L = 200 ft, m = 11 kips/g per foot, I = 700 ft4 , and E = 576,000 kips/ft2 .
If v = 55 mph, determine the impact factor defined as the ratio of maximum deflection at
midspan and the static deflection.
Solution We assume that the mass of the wheel load is small compared to the bridge mass,
and it can be neglected.
1. Determine the generalized mass, generalized stiffness, and natural frequency.
πx π2 πx
ψ(x) = sin ψ (x) = − sin
L L2 L
L
πx mL
m̃ = m sin2 dx = (a)
0 L 2
L 2
π2 πx π4E I
k̃ = EI sin2 dx = (b)
0 L2 L 2L 3
k̃ π2 EI
ωn = = 2 (c)
m̃ L m
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 318
v
t
x m, EI L/v
L
• •
(b)
(a)
0 Time t, sec
0.5
(c)
1.0
1.5
Figure E8.4
This happens to be the exact value of the lowest natural frequency of the bridge because
the ψ(x) selected is the exact shape of the fundamental natural vibration mode of a simply
supported beam (see Section 16.3.1).
2. Determine the generalized force. A load po traveling with a velocity v takes time
td = L/v to cross the bridge. At any time t its position is as shown in Fig. E8.4a. Thus the
moving load can be described mathematically as
po δ(x − vt) 0 ≤ t ≤ td
p(x, t) = (d)
0 t ≥ td
where δ(x − vt) is the Dirac delta function centered at x = vt; it is a mathematical description
of the traveling concentrated load. From Eq. (8.3.26) the generalized force is
L
L
p̃(t) = p(x, t)ψ(x) d x = 0
po δ(x − vt) sin(π x/L) d x 0 ≤ t ≤ td
0 0 t ≥ td
po sin(πvt/L) 0 ≤ t ≤ td
=
0 t ≥ td
po sin(πt/td ) 0 ≤ t ≤ td
= (e)
0 t ≥ td
This generalized force is the half-cycle sine pulse shown in Fig. E8.4b.
3. Solve the equation of motion.
Equations (4.8.2) and (4.8.3) describe the response of an SDF system to a half-cycle sine
pulse. We will adapt this solution to the bridge problem by changing the notation from u(t) to
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 319
The response is given by Eq. (g) while the moving load is on the bridge span and by Eq. (h)
after the load has crossed the span. These equations are valid provided that ωn = πv/L or
Tn = 2L/v.
4. Determine the deflection at midspan.
πx
u(x, t) = z(t)ψ(x) = z(t) sin (i)
L
At midspan, x = L/2 and
L
u ,t = z(t) (j)
2
Thus the deflection at midspan is also given by Eqs. (g) and (h).
5. Numerical results. For the given prestressed-concrete structure and vehicle speed:
11
m= = 0.3416 kip-sec2 /ft2
32.2
E I = 576,000 × 700 = 4.032 × 108 kip-ft2
π2 4.032 × 108
ωn = = 8.477 rad/sec
(200)2 0.3416
Tn = 0.74 sec
Equation (k) is valid for 0 ≤ t ≤ 2.479 sec. The values of midspan deflections calculated
from Eq. (k) at many values of t are shown in Fig. E8.4c; the maximum deflection is u o =
( po /2400)(1.147). The static deflection is u(L/2) = po L 3 /48E I = po /2419. The ratio of
these deflections gives the impact factor: 1.156 (i.e., the static load should be increased by
15.6% to account for the dynamic effect).
Example 8.5
Determine the natural frequency of transverse vibration of the three-span, box-girder bridge
of Example 1.3. Therein several of the properties of this structure were given. In addition, the
second moment of area for transverse bending of the bridge deck is given: I y = 65,550 ft4 .
Neglect torsional stiffness of the bents.
Solution
1. Select the shape function. We select a function appropriate for a beam simply sup-
ported at both ends (Fig. E8.5):
πx
ψ(x) = sin (a)
L
This shape function is shown in Fig. E8.5.
2. Determine the generalized mass.
L
πx mL
m̃ = m sin2 dx = (b)
0 L 2
π 4 E Iy 3 3
= 3
+ kbent + kbent
2L 4 4
π 4 E Iy 3
= 3
+ kbent (c)
2
2L
k̃deck k̃bents
4. Determine numerical values for m̃ and k̃. From Example 1.3, the weight of the bridge
deck per unit length is 18.45 kips/ft.
1 18.45 × 375
m̃ = = 107.5 kip-sec2 /ft
32.2 2
•
x
L
• Figure E8.5
From Example 1.3, the stiffness of each bent in the longitudinal direction is kbent = 12,940
kips/ft. Because each column in the bent has a circular section, its second moment of area is
the same for bending in the longitudinal or transverse directions. Thus, the transverse stiffness
of each bent is also kbent = 12,940 kips/ft.
3 3
k̃bent = kbent = (12,940) = 19,410 kips/ft
2 2
k̃ 45,564
ωn = = = 20.59 rad/sec
m̃ 107.5
2π
Tn = = 0.305 sec
ωn
mN z
N uN δz fIN fNo
mj
j uj δuj fIj fjo
hj 2
m1
1 u1 fI1 f1o
ug
(a) (b) (c)
Figure 8.4.1 (a) Building displacements and virtual displacements; (b) inertia forces;
(c) equivalent static forces.
usually defined by a damping ratio estimated from experimental data for similar structures
(Chapter 11). Then the equation of motion is solved to determine the peak response—
displacements and internal forces—of the structure to earthquake ground motion.
We assume that the floor displacements relative to the ground can be expressed as
u j (t) = ψ j z(t) j = 1, 2, . . . , N (8.4.1a)
which in vector form is
u(t) = ψz(t) (8.4.1b)
where ψ is an assumed shape vector that defines the deflected shape. The total displace-
ment of the jth floor is
u tj (t) = u j (t) + u g (t) (8.4.2)
Before we can formulate the equation of motion for this system, we must define how the
internal forces are related to the displacements. This relationship is especially easy to de-
velop if the beams are rigid axially as well as in flexure and this “shear building” assump-
tion is adequate for our present objectives; however, realistic idealizations of multistory
buildings will be developed in Chapter 9. For this shear building idealization, the shear Vj
in the jth story (which is the sum of the shear in all columns) is related to the story drift
j = u j − u j−1 through the story stiffness k j :
Vj = k j j = k j (u j − u j−1 ) (8.4.3)
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 323
The story stiffness is the sum of the lateral stiffnesses of all columns in the story:
12E I
kj = (8.4.4)
columns
h3
where E I is the flexural rigidity of a column and h its height.
We now proceed to formulate the equation of motion for the shear building. At
each time instant the system is in equilibrium under the action of the internal story shears
Vj (t) and the fictitious inertia forces (Fig. 8.4.1b), which by D’Alembert’s principle are
f I j = −m j ü tj . Substituting Eq. (8.4.2) for u tj gives
f I j = −m j [ü j (t) + ü g (t)] (8.4.5)
As before, the principle of virtual displacements provides a convenient approach for
formulating the equation of motion. The procedure is similar to that developed in Section
8.3 for a beam. The external virtual work is due to the forces f I j acting through the virtual
displacements δu j :
N
δW E = f I j (t) δu j
j=1
The internal virtual work is due to the story shears Vj (t) acting through the story drifts
associated with the virtual displacements:
N
δW I = Vj (t)(δu j − δu j−1 ) (8.4.7)
j=1
Having obtained these expressions for δW E and δW I , Eq. (8.3.4), after dropping δz (see
Section 8.3.2), gives the equation of motion:
m̃ z̈ + k̃z = − L̃ ü g (t) (8.4.11)
where
N
N
N
m̃ = m j ψ j2 k̃ = k j (ψ j − ψ j−1 )2 L̃ = m j ψj (8.4.12a)
j=1 j=1 j=1
We digress briefly to mention that the preceding derivation would have been easier
if we used a matrix formulation, but we avoided this approach because the stiffness matrix
k and the mass matrix m of a structure are not introduced until Chapter 9. However, most
readers are expected to be familiar with the stiffness matrix of a structure, and as we shall
see in Chapter 9, the mass matrix of the system of Fig. 8.4.1 is a diagonal matrix with
m j j = m j . Using these matrices and the shape vector ψ = ψ1 ψ2 · · · ψ N T ,
Eq. (8.4.12a) for the generalized properties becomes
m̃ = ψ T mψ k̃ = ψ T kψ L̃ = ψ T m1 (8.4.12b)
where 1 is a vector with all elements unity. Equation (8.4.12b) for the generalized stiffness
is a general result because, unlike Eq. (8.4.12a), it is not restricted to shear buildings, as
long as k is determined for a realistic idealization of the structure.
Equation (8.4.11) governs the motion of the multistory shear frame assumed to de-
flect in the shape defined by the vector ψ. For this generalized SDF system, the gener-
alized mass m̃, generalized stiffness k̃, and generalized excitation − L̃ ü g (t) are defined
by Eq. (8.4.12). Dividing Eq. (8.4.11) by m̃ and including a damping term using an
estimated modal damping ratio ζ gives Eqs. (8.3.13b) and (8.3.14), demonstrating that
the same equation of motion applies to both—lumped mass or distributed mass—
generalized SDF systems; the generalized properties m̃, k̃, and L̃ depend on the system,
of course.
The generalized SDF system can now be analyzed by the methods developed in preceding
chapters for SDF systems. In particular, the natural vibration frequency of the system is
given by
N
j=1 k j (ψ j − ψ j−1 )
2
k̃
ωn =
2
= N (8.4.13a)
j=1 m j ψ j
m̃ 2
be determined by solving Eq. (8.3.13b) using the methods of Chapters 5 and 6. Equation
(8.4.1) then gives floor displacements relative to the base.
Suppose that it is desired to determine the peak response of the frame to earthquake
excitation characterized by a design spectrum. The peak value of z(t) is still given by
Eq. (8.3.20), and the floor displacements relative to the ground are given by Eq. (8.4.1)
with z(t) replaced by z o :
u jo = ψ j z o = ˜ Dψ j j = 1, 2, . . . , N (8.4.14)
The equivalent static forces associated with these floor displacements are given by
Eq. (8.3.21b) modified for a lumped-mass system:
˜ j ψj A
f jo = m j = 1, 2, . . . , N (8.4.15)
Static analysis of the structure subjected to these floor forces (Fig. 8.4.1c) gives the
shear force Vio in the ith story and overturning moment at the ith floor:
N
N
Vio = f jo Mio = (h j − h i ) f jo (8.4.16)
j=i j=i
where h i is the height of the ith floor above the base. In particular, the shear and overturn-
ing moment at the base are
N
N
Vbo = f jo Mbo = h j f jo
j=1 j=1
N
L̃ θ = h j m j ψj (8.4.18)
j=1
Observe that these equations for forces at the base of a lumped-mass system are the same
as derived earlier for distributed-mass systems [Eqs. (8.3.23) and (8.3.24)]; the parameters
L̃, L̃ θ , and ˜ depend on the system, of course.
Example 8.6
The uniform five-story frame with rigid beams shown in Fig. E8.6a is subjected to ground ac-
celeration ü g (t). All the floor masses are m, and all stories have the same height h and stiffness
k. Assuming the displacements to increase linearly with height above the base (Fig. E8.6b),
formulate the equation of motion for the system and determine its natural frequency.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 326
Solution
1. Determine the generalized properties.
5
12 + 2 2 + 3 2 + 4 2 + 5 2 11
m̃ = m j ψ j2 = m = m
52 5
j=1
5
12 + 1 2 + 1 2 + 1 2 + 1 2 k
k̃ = k j (ψ j − ψ j−1 )2 = k =
52 5
j=1
5
1+2+3+4+5
L̃ = m j ψj = m = 3m
5
j=1
2. Formulate the equation of motion. Substituting for m̃ and L̃ in Eq. (8.3.14) gives
˜ = 15
11 and Eq. (8.3.13b) becomes
z̈ + ωn2 z = − 15
11 ü g (t)
where z is the lateral displacement at the location where ψ j = 1, in this case the top of the
frame.
3. Determine the natural vibration frequency.
k/5 k
ωn = = 0.302
11m/5 m
√
This is about 6% higher than ωn = 0.285 k/m, the exact frequency of the system determined
in Chapter 12.
Example 8.7
Determine the peak displacements, story shears, and floor overturning moments for the frame
of Example 8.6 with m = 100 kips/g, k = 31.54 kips/in., and h = 12 ft (Fig. E8.7a) due to
the ground motion characterized by the design spectrum of Fig. 6.9.5 scaled to a peak ground
acceleration of 0.25g.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 327
m
11.3 in. 32.45 kips 32.45 kips
•
m k
9.04 25.96 58.41 389
m k
5@12′
Figure E8.7
Solution
1. Compute the natural period.
31.54
ωn = 0.302 = 3.332
100/386
2π
Tn = = 1.89 sec
3.332
2. Determine the peak value of z(t). For Tn = 1.89 sec and ζ = 0.05, the design
spectrum gives A/g = 0.25(1.80/1.89) = 0.238 and D = A/ωn2 = 8.28 in. The peak value
of z(t) is
zo = 15
11 D = 11 (8.28)
15
= 11.3 in.
3. Determine the peak values u jo of floor displacements.
j
u jo = ψ j z o ψj =
5
Therefore, u 1o = 2.26, u 2o = 4.52, u 3o = 6.78, u 4o = 9.04, and u 5o = 11.3, all in inches
(Fig. E8.7b).
4. Determine the equivalent static forces.
˜ j ψj A =
f jo = m 11 mψ j (0.238g) = 32.45ψ j kips
15
Although the principle of virtual displacements provides an approximate result for the
natural vibration frequency [Eqs. (8.3.15) and (8.4.13)] of any structure, it is instructive
to obtain the same result by another approach, developed by Lord Rayleigh. Based on
the principle of conservation of energy, Rayleigh’s method was published in 1873. In this
section this method is applied to a mass–spring system, a distributed-mass system, and a
lumped-mass system.
When an SDF system with lumped mass m and stiffness k is disturbed from its equilibrium
position, it oscillates at its natural vibration frequency
√ ωn , and it was shown in Section 2.1
by solving the equation of motion that ωn = k/m. Now we will obtain the same result
using the principle of conservation of energy.
The simple harmonic motion of a freely vibrating mass–spring system, Eq. (2.1.3),
can be described conveniently by defining a new time variable t with its origin as shown
in Fig. 8.5.1a:
u(t ) = u o sin ωn t (8.5.1)
where the frequency ωn is to be determined and the amplitude u o of the motion is given by
Eq. (2.1.9). The velocity of the mass, shown in Fig. 8.5.1b, is
u̇(t ) = ωn u o cos ωn t (8.5.2)
The potential energy of the system is the strain energy in the spring, which is propor-
tional to the square of the spring deformation u (Eq. 2.3.2). Therefore, the strain energy is
u
uo
t=0 t′ = 0 t, t′ (a)
•
Tn
u̇
ωn uo
t, t′ (b)
Figure 8.5.1 Simple harmonic motion of a
freely vibrating system: (a) displacement;
(b) velocity.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 329
maximum at t = Tn /4 (also at t = 3Tn /4, 5Tn /4, . . .) when u(t) = u o and is given by
E So = 12 ku 2o (8.5.3)
This is also the total energy of the system because at this t the velocity is zero (Fig. 8.5.1b),
implying that the kinetic energy is zero.
The kinetic energy of the system is proportional to the square of the velocity of the
mass u̇ [Eq. (2.3.2)]. Therefore, the kinetic energy is maximum at t = 0 (also at t = Tn /2,
3Tn /2, . . .) when the velocity u̇(t) = ωn u o and is given by
E K o = 12 mωn2 u 2o (8.5.4)
This is also the total energy of the system because at this t , the deformation is zero
(Fig. 8.5.1a), implying that the strain energy is zero.
The principle of conservation of energy states that the total energy in a freely vi-
brating system without damping is constant (i.e., it does not vary with time), as shown by
Eq. (2.3.5). Thus the two alternative expressions, E K o and E So , for the total energy must
be equal, leading to the important result:
maximum kinetic energy, E K o = maximum potential energy, E So (8.5.5)
Substituting Eqs. (8.5.3) and (8.5.4) gives
k
ωn = (8.5.6)
m
This is the same result for the natural vibration frequency as Eq. (2.1.4) obtained by solving
the equation of motion.
Rayleigh’s method does not provide any significant advantage in obtaining the nat-
ural vibration frequency of a mass–spring system, but the underlying concept of energy
conservation is useful for complex systems, as shown in the next two sections.
As an illustration of such a system, consider the cantilever tower of Fig. 8.3.1 vibrating
freely in simple harmonic motion:
u(x, t ) = z o sin ωn t ψ(x) (8.5.7)
where ψ(x) is an assumed shape function that defines the form of deflections, z o is the
amplitude of the generalized coordinate z(t), and the natural vibration frequency ωn is to
be determined. The velocity of the tower is
u̇(x, t ) = ωn z o cos ωn t ψ(x) (8.5.8)
The maximum potential energy of the system over a vibration cycle is equal to its strain
energy associated with the maximum displacement u o (x):
L
E So = 1
2
E I (x)[u o (x)]2 d x (8.5.9)
0
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 330
The maximum kinetic energy of the system over a vibration cycle is associated with the
maximum velocity u̇ o (x):
L
EKo = 1
2
m(x)[u̇ o (x)]2 d x (8.5.10)
0
From Eqs. (8.5.7) and (8.5.8), u o (x) = z o ψ(x) and u̇ o (x) = ωn z o ψ(x). Substituting these
in Eqs. (8.5.9) and (8.5.10) and equating E K o to E So gives
L
E I (x)[ψ (x)]2 d x
ωn = 0 L
2
(8.5.11)
2
0 m(x)[ψ(x)] d x
This is known as Rayleigh’s quotient for a system with distributed mass and elasticity; re-
call that the same result, Eq. (8.3.15), was obtained using the principle of virtual displace-
ments. Rayleigh’s quotient is valid for any natural vibration frequency of a multi-degree-
of-freedom system, although its greatest utility is in determining the lowest or fundamental
frequency.
As an illustration of such a system, consider the shear building of Fig. 8.4.1 vibrating freely
in simple harmonic motion,
u(t ) = z o sin ωn t ψ (8.5.12)
where the vector ψ is an assumed shape vector that defines the form of deflections, z o is
the amplitude of the generalized coordinate z(t), and the natural vibration frequency ωn
is to be determined. The velocities of the lumped masses of the system are given by the
vector
u̇(t ) = ωn z o cos ωn t ψ (8.5.13)
The maximum potential energy of the system over a vibration cycle is equal to its strain
energy associated with the maximum displacements, uo = u 1o u 2o · · · u N o T :
N
2
E So = 1
k
2 j
u jo − u j−1,o (8.5.14)
j=1
The maximum kinetic energy of the system over a vibration cycle is associated with the
maximum velocities, u̇o = u̇ 1o u̇ 2o · · · u̇ N o T :
N
EKo = 1
m u̇ 2
2 j jo
(8.5.15)
j=1
What makes Rayleigh’s method especially useful for estimating a natural vibration fre-
quency of a system are the properties of Rayleigh’s quotient: First, the approximate fre-
quency obtained from an assumed shape function is never smaller than the exact value of
the lowest or fundamental natural frequency of the system. Second, Rayleigh’s quotient
provides excellent estimates of the fundamental frequency, even with a not-so-good shape
function.
Let us examine these properties in the context of a specific system, the cantilever
considered in Example 8.2. Its fundamental frequency can be expressed as ωn =
tower
αn E I /m L 4 [see Eq. (d) of Example 8.2]. Three different estimates of αn using three
different shape functions are summarized in Table 8.5.1. The second frequency estimate
comes from Example 8.2. The same procedure leads to the results using the other two
shape functions. The percentage error shown is relative to the exact value of αn = 3.516
(Chapter 16).
ψ(x) αn % Error
Consistent with the properties of Rayleigh’s quotient, the three estimates of the nat-
ural frequency are higher than its exact value. Even if we did not know the exact value, as
would be the case for complex systems, we could say that the smallest value, αn = 3.57, is
the best among the three estimates for the natural frequency. This concept can be used to
determine the exact frequency of a two-DOF system by minimizing the Rayleigh’s quotient
over a shape function parameter.
Why such a large error in the third case of Table 8.5.1? The shape function ψ(x) =
x 2 /L 2 satisfies the displacement boundary conditions at the base of the tower but violates
a force boundary condition at the free end. It implies a constant bending moment over
the height of the tower, but a bending moment at the free end of a cantilever is unrealistic
unless there is a mass at the free end with a moment of inertia. Thus, a shape function that
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 332
satisfies only the geometric boundary conditions does not always ensure an accurate result
for the natural frequency.
An estimate of the natural vibration frequency of a system obtained using Rayleigh’s
quotient can be improved by iterative methods. Such methods are developed in
Chapter 10.
The accuracy of the natural vibration frequency estimated using Rayleigh’s quotient de-
pends entirely on the shape function that is assumed to approximate the exact mode shape.
In principle any shape may be selected that satisfies the displacement boundary conditions
at the supports. In this section we address the question of how a reasonable shape function
can be selected to ensure good results.
For this purpose it is useful to identify the properties of the exact mode shape. In
free vibration the displacements are given by Eq. (8.5.7) and the associated inertia forces
are
f I (x, t) = −m(x)ü(x, t ) = ωn2 z o m(x)ψ(x) sin ωn t
If ψ(x) were the exact mode shape, static application of these inertia forces at each time
instant will produce deflections given by Eq. (8.5.7), a result that will become evident
in Chapter 16. This concept is not helpful in evaluating the exact mode shape ψ(x)
because the inertia forces involve this unknown shape. However, it suggests that an
approximate shape function ψ(x) may be determined as the deflected shape due to static
forces p(x) = m(x)ψ̃(x), where ψ̃(x) is any reasonable approximation of the exact mode
shape.
In general, this procedure to select the shape function involves more computational
effort than is necessary because, as mentioned earlier, Rayleigh’s method gives good accu-
racy even if the shape function is not so good. However, the preceding discussion does sup-
port the concept of determining the shape function from deflections due to a selected set of
static forces. One common selection for these forces is the weight of the structure applied
in an appropriate direction. For the cantilever tower it is the lateral direction (Fig. 8.6.1a).
This selection is equivalent to taking ψ̃(x) = 1 in p(x) = m(x)ψ̃(x). Another selection
includes several concentrated forces as shown in Fig. 8.6.1b.
The displacement and force boundary conditions are satisfied automatically if the
shape function is determined from the static deflections due to a selected set of forces.
This choice of shape function has the additional advantage that the strain energy can be
calculated as the work done by the static forces in producing the deflections, an approach
that is usually simpler than Eq. (8.5.9). Thus, the maximum strain energy of the system
associated with the forces p(x) in Fig. 8.6.1a is
L
E So = 12 p(x)u(x) d x
0
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 333
•
p(x) = gm(x) pj
EI(x)
m(x)
L xj u(x)
x
p1
•
(a) (b) (c)
Equating this E So to E K o of Eq. (8.5.10) with u̇ o (x) = ωn u o (x) and dropping the subscript
“o” gives
L
p(x)u(x) d x
ωn = L0
2
(8.6.1)
2
0 m(x)[u(x)] d x
This equation with p(x) = po (i.e., uniformly distributed forces) appears in the AASHTO
(American Association of State Highway and Transportation Officials) code to estimate
the fundamental natural frequency of a bridge.
For p(x) = gm(x) in Fig. 8.6.1a, Eq. (8.6.1) becomes
L
m(x)u(x) d x
ωn = g L0
2
(8.6.2)
2
0 m(x)[u(x)] d x
Similarly, the maximum strain energy of the system associated with deflections u(x)
due to the forces of Fig. 8.6.1b is
E So = 12 p j u(x j )
Equating this E So to E K o of Eq. (8.5.10) gives
p j u(x j )
ωn = L
2
(8.6.3)
2
0 m(x)[u(x)] d x
Although attractive in principle, selecting the shape function as the static deflections
due to a set of forces can be cumbersome for the nonuniform (variable E I ) tower shown in
Fig. 8.6.1a. A convenient approach is to determine the static deflections of a uniform (con-
stant EI) tower of the same length, and use the resulting shape function for the nonuniform
tower. The reader is reminded, however, against doing complicated analysis to determine
deflected shapes in the interest of obtaining an extremely accurate natural frequency. The
principal attraction of Rayleigh’s method lies in its ability to provide a useful estimate of
the natural frequency from any reasonable assumption on the shape function that satisfies
the displacement and force boundary conditions.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 334
mN
N pN pN mNg uN
mj
j pj mjg uj
m1
1 p1 m1g u1
The concept of using the shape function as the static deflections due to a selected set
of forces is also useful for lumped-mass systems. Three sets of forces that may be used
for a multistory building frame are shown in Fig. 8.6.2. The maximum strain energy of the
system associated with deflections u j in the three cases is
N
N
E So = 12 p N u N E So = 1
2
pj u j E So = 12 g mj uj
j=1 j=1
m(x)g m(x)g
(a) (b)
m(x)g
u(x) u(x)
Figure 8.6.3 Shape functions resulting from self-weight applied in appropriate direc-
tions.
Example 8.8
Estimate the natural frequency of a uniform cantilever beam assuming the shape function
obtained from static deflections due to a load p at the free end.
Solution
1. Determine the deflections. With the origin at the clamped end,
p
u(x) = (3L x 2 − x 3 ) (a)
6E I
2. Determine the natural frequency from Eq. (8.6.3).
L3
p j u(x j ) = pu(L) = p 2 (b)
3E I
j
L L
p2 11 p 2 m L 7
m(x)[u(x)]2 d x = m (3L x 2 − x 3 )2 d x = (c)
0 (6E I )2 0 420 (E I )2
Substituting Eqs. (b) and (c) in Eq. (8.6.3) gives
3.57 EI
ωn = 2
L m
This is the first frequency estimate in Table 8.5.1.
Example 8.9
Estimate the fundamental natural frequency of the five-story frame in Fig. E8.9 assuming the
shape function obtained from static deflections due to lateral forces equal to floor weights
w = mg.
Solution
1. Determine the deflections due to applied forces. The static deflections are determined
as shown in Fig. E8.9 by calculating the story shears and the resulting story drifts; and adding
these drifts from the bottom to the top to obtain
w
uT = 5 9 12 14 15T
k
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 336
Figure E8.9
FURTHER READING
Rayleigh, J. W. S., Theory of Sound, Dover, New York, 1945; originally published in 1894.
The inertia forces for a rigid bar and rectangular and circular rigid plates associated with
accelerations ü x , ü y , and θ̈ of the center of mass O (or center of gravity) are shown in
Fig. A8.1. Each rigid body is of uniform thickness, and its total mass m is uniformly
distributed; the moment of inertia I O about the axis normal to the bar or plate and passing
through O is as noted in the figure.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 337
ü y
θ̈ mL2 θ̈
12
ü x mL2
IO =
O müx O 12
müy
(a) (b)
b/2 b/2
• • •
•
m(b2 + d2) θ̈
d/2 12 mR2 θ̈
m(b2 + d2) 2 mR2
IO = IO =
•
12 O O 2
müx müx
d/2 müy müy
R
•
• •
(c) (d)
PROBLEMS
8.1 Repeat parts (a), (b), and (c) of Example 8.1 with one change: Use the horizontal displace-
ment at C as the generalized coordinate. Show that the natural frequency, damping ratio, and
displacement response are independent of the choice of generalized displacement.
8.2 For the rigid-body system shown in Fig. P8.2:
(a) Formulate the equation of motion governing the rotation at O.
(b) Determine the natural frequency and damping ratio.
(c) Determine the displacement response u(x, t) to p(t) = δ(t), the Dirac delta function.
p(t)
Mass = m
Rotational
•
k L/8
•
• • •
L/2 L/2 L/8 L/8
• • • Figure P8.2
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 338
8.3 Solve Problem 8.2 with one change: Use the vertical displacement at the center of gravity of
the square plate as the generalized displacement. Show that the results are independent of the
choice of generalized displacement.
8.4 The rigid bar in Fig. P8.4 with a hinge at the center is bonded to a viscoelastic foundation,
which can be modeled by stiffness k and damping coefficient c per unit of length. Using the
rotation of the bar as the generalized coordinate:
(a) Formulate the equation of motion.
(b) Determine the natural vibration frequency and damping ratio.
L/2 L/2
• • • Figure P8.4
p(t)
Massless
Mass = m
k k
8.6 Solve Example 8.3 assuming the deflected shape function due to lateral force at the top:
3 x2 1 x3
ψ(x) = −
2 L2 2 L3
The shear forces and bending moments need to be calculated only at the base and mid-
height. (Note that these forces were determined in Example 8.3 throughout the height of the
chimney.)
8.7 A reinforced-concrete chimney 600 ft high has a hollow circular cross section with outside
diameter 50 ft at the base and 25 ft at the top; the wall thickness is 2 ft 6 in., uniform over
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 339
the height (Fig. P8.7). Using the approximation that the wall thickness is small compared
to the radius, the mass and flexural stiffness properties are computed from the gross area of
concrete (neglecting reinforcing steel). The chimney is assumed to be clamped at the base,
and its damping ratio is estimated to be 5%. The unit weight of concrete is 150 lb/ft3 , and its
elastic modulus E c = 3600 ksi. Assuming that the shape function is
πx
ψ(x) = 1 − cos
2L
where L is the length of the chimney and x is measured from the base, calculate the following
quantities: (a) the shear forces and bending moments at the base and at the midheight, and (b)
the top deflection due to ground motion defined by the design spectrum of Fig. 6.9.5, which
is scaled to a peak acceleration of 0.25g.
25′
• •
•• 2′-6″
•
600′
•
• •
50′ Figure P8.7
3 x2 1 x3
ψ(x) = −
2 L2 2 L3
8.9 Solve Problem 8.7 for a different excitation: a blast force varying linearly over height from
zero at the base to p(t) at the top, where p(t) is given in Fig. P8.9.
p(t)
p, kips/ft
4
t, sec
0.1 0.25
Figure P8.9
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 340
8.10– Three-story shear frames (rigid beams and flexible columns) in structural steel (E =29,000 ksi)
8.11 are shown in Figs. P8.10 and P8.11; w = 100 kips; I = 1400 in4 ; and the modal damping
ratios ζn are 5% for all modes. Assuming that the shape function is given by deflections due
to lateral forces that are equal to the floor weights, determine the floor displacements, story
shears, and overturning moments at the floors and base due to ground motion characterized
by the design spectrum of Fig. 6.9.5 scaled to a peak ground acceleration of 0.25g.
•
12′ EI EI 12′ EI/3 EI/3
w w
•
•
12′ EI EI 12′ 2EI/3 2EI/3
w w
•
•
12′ EI EI 12′ EI EI
•
•
24′ 24′
• • • •
8.12– Solve Problems 8.10 and 8.11 using the shape function given by deflections due to a lateral
8.13 force at the roof level.
8.14 A five-story frame with rigid beams shown in Fig. P8.14a is subjected to ground acceleration
ü g (t); k j are story stiffnesses. Assuming the displacements to increase linearly with height
above the base (Fig. P8.14b), formulate the equation of motion for the system and determine
its natural frequency. Determine the floor displacements, story shears, and floor overturning
moments due to ground motion characterized by the design spectrum of Fig. 6.9.5 scaled to a
peak ground acceleration of 0.25g.
Rigid beams
50 kips
kj, kips/in. 1
•
12′ 100
100
4/5
•
12′ 150
100
3/5
•
12′ 150
100
2/5
•
12′ 200
100
1/5
•
12′ 200
•
8.15 Solve Problem 8.14 using the shape function obtained from static deflections due to lateral
forces equal to floor weights.
8.16 Solve Problem 8.14 using the shape function given by deflections due to a lateral force at the
roof level.
8.17 Determine the natural vibration frequency of the inverted L-shaped frame shown in Fig. P8.17
using the shape function given by the deflections due to a vertical force at the free end. Neglect
deformations due to shear and axial force.
• L • Mass = m
2m m
k 2k
•
L
• •
L
1
ψr
•
8.18 (a) By Rayleigh’s method determine the natural vibration frequency of the rigid bar on two
springs (Fig. P8.18) using the shape function shown. Note that the result involves the unknown
ψr . Plot the value of ωn2 as a function of ψr .
(b) Using the properties of Rayleigh’s quotient, determine the exact values of the two vibration
frequencies and the corresponding vibration shapes.
8.19 The umbrella structure shown in Fig. P8.19 consists of a uniform column of flexural rigidity
E I supporting a uniform slab of radius R and mass m. By Rayleigh’s method determine the
natural vibration frequency of the structure. Neglect the mass of the column and the effect of
axial force on column stiffness. Assume that the slab is rigid in flexure and that the column is
axially rigid.
R
Mass = m • •
•
L
•
Figure P8.19
8.20 By Rayleigh’s method determine the natural vibration frequency of the uniform beam shown
in Fig. P8.20. Assume that the shape function is given by the deflections due to a force applied
at the free end.
Chopra: Prentice-Hall
PAGES JUL. 19, 2000 14:21
ICC Oregon (503) 221-9911
Page 342
EI m(x) = m
2L L
• • • Figure P8.20
8.21 By Rayleigh’s method determine the natural vibration frequency of transverse vibration of the
three-span, box-girder bridge of Example 8.5. Assume that the shape function is given by
the deflections due to uniform force p(x) = 1 applied in the transverse direction. Neglect
torsional stiffness of the bents.
8.22 Repeat Problem 8.21 using a simpler approach in which the deflections are assumed to be
u(x) = u o sin(π x/L), where u o is the midspan deflection due to uniform force p(x) = 1
applied in the transverse direction.
8.23 Repeat Problem 8.21 using a simpler approach in which the deflections are assumed to be
u(x) = u o ψ(x), where u o is the midspan deflection due to uniform force p(x) = 1 applied
in the transverse direction and
x 3 x 4
16 x
ψ(x) = −2 +
5 L L L
Note that ψ(x) is the deflected shape of a simply supported beam without bents subjected to
transverse force p(x) = 1.
8.24 Repeat Problem 8.21 with one change: consider torsional stiffness of the bents.
8.25 A simply supported bridge with a single span of L feet has a deck of uniform cross section
with mass m per foot length and flexural rigidity E I . An infinitely long, uniformly distributed
force po per foot length (that represents a very long train) travels across the bridge at a uniform
velocity v (Fig. P8.25). Determine an equation for the deflection at midspan as a function of
time. Neglect damping and assume the shape function as ψ(x) = sin(π x/L).
po
∞ v
m, EI
L
• • Figure P8.25
8.26 A pulsating force p(t) = po cos ωt travels across the bridge of Fig. P8.25 at a uniform velocity
of v, as shown in Fig. P8.26. Determine an equation for the deflection at midspan as a function
of time. Neglect damping and assume the shape function as ψ(x) = sin(π x/L).
p(t)
v
m, EI
L
• • Figure P8.26