Dynamic Loading and Design of Structures 1729068739
Dynamic Loading and Design of Structures 1729068739
Chapter 2
Analysis for dynamic loading
George D.Monolis
2.1 INTRODUCTION
The purpose of this chapter is analysis of structures that are subjected to time varying loads.
Despite the fact that the majority of civil engineering structures are built on the assumption
that all applied loads are static, there are exceptions which require a distinction between static
and dynamic loads to be made, as in earthquake engineering. All loads in nature are time
dependent. In many cases, however, loads will be applied to a structure in slowly varying
ways, which implies that static conditions can be assumed. The term slow here is quantified
through comparison with an intrinsic time of the structure, which is none other than its natural
period. Thus, a load varies slowly or is fast only in relation to the time required for the
structure to complete a full cycle of oscillation.
There is growing interest nowadays in the process of designing civil engineering structures
to withstand dynamic loads (Biggs, 1965; Craig, 1981; Bathe, 1982). As examples, we
mention (i) structures which house moving or vibrating equipment, (ii) bridges under traffic,
(iii) multistory structures subject to wind and (iv) the case of earthquake induced loads
(Clough and Penzien, 1993; Newmark and Rosenblueth, 1971). Essentially, dynamic analyses
focus on evaluation of time dependent displacements, from which the stress state of the
structure in question can be computed (Paz, 1997; Argyris and Mlejnek, 1991; Chopra, 1995).
The most basic pieces of information needed for this are the natural period, which is a
function of the structure’s mass and stiffness, and the amount of available damping (or,
equivalently, the amount of energy that can be absorbed by the structure).
Figure 2.1 (a) SDOF modelling of a single story frame for (b) horizontal, (c) vertical and (d) rotational
oscillations.
Figure 2.2 Various types of dynamic loads: harmonic, aperiodic, earthquake and long duration.
rotational, as shown in Figure 2.1(b)-(d). As expected, the SDOF oscillator is used extensively
for modelling structural systems, but it should be remembered that it is an approximate model
for anything else but the simple frame previously mentioned. Next, some typical dynamic
loads are shown in Figure 2.2, where we distinguish between periodic (both harmonic and
non-harmonic) and aperiodic (both short and long duration) loads.
Page 33
With reference to Figure 2.3, the equation of motion of the SDOF oscillator is
(2.1)
implying that the inertia, damping and restoring forces balance the applied force. Specifically,
M is the mass (kg), k is the stiffness (N/m), and c is the damping coefficient (N-sec/m).
Furthermore, y(t) is the displacement (m), ÿ(t) the velocity (m/ sec), ÿ(t) the acceleration
(m/sec2), F(t)=F1f(t) the externally applied force (N) withf(t) its dimensionless time variation.
Finally, dots denote time derivatives d/dt. Obviously, eqn (2.1) is a second order differential
equation that needs to be solved for the displacement y(t).
(2.2)
Thus, the oscillator undergoes free vibrations under the influence of an initial displacement
y(0)=y 0 and/or initial velocity ÿ(0)=ÿ0. The solution is simply
(2.3)
and implies a periodic, harmonic motion as shown in Figure 2.4. At this point, we respectively
define the circular frequency, the natural period and the frequency as
Page 34
Figure 2.4 Free vibration due to (a) initial displacement and (b) initial velocity.
follows:
(2.4)
(2.5)
(2.6)
(2.7)
We define as Dynamic Load Factor (DLF) the ratio of the dynamic displacement at
Page 35
any time instant to the displacement produced by static application of the load F(t)=F 1 as
(2.9)
(2.10)
The DLF is dimensionless and measures the amount by which the dynamic displacement in
the SDOF system exceeds its equivalent static one. Figure 2.5 plots the DLF for the suddenly
applied and maintained load case, where doubling of the response is observed at certain time
instances.
(2.11)
(2.12)
Thus, from eqn (2.3) the system experiences an instantaneous displacement y(t) equal to
(2.14)
Finally, the complete displacement history is evaluated by integrating from time t=0 to the
present time t as
(2.15)
(2.16)
then
(2.17)
If we finally add the effect of initial conditions at t=0, then we have a generel, closed form
expression for the dynamic displacement of the SDOF system in the form of Duhamel’s
integral as
(2.18)
(2.19)
(2.20)
where yst=F1/k. Figure 2.6 plots the above results for two cases, where we observe an intense
response when the duration of the load on the oscillator is large (td=1.2T). If the load is on the
oscillator for a short time (td=0.1T), the dynamic response is less than the static one.
(2.21)
and Figure 2.7 plots two cases, one with a rapid (tr =0.2T) and the other with a slow (tr =3.33T)
application. We note that the latter case produces a quasi-static response in the SDOF
oscillator. Finally, in Figure 2.8 we have the maximum value of the DLFmax as function of the
time ratio tr/T.
Page 38
Figure 2.8 Maximum value of the DLF as a function of rise time tr.
(2.22)
and its solution in terms of the DLF (with y st=F1/k) and for y0=y0=0 has the following form:
(2.23)
We observe that the oscillations comprise two parts, the free part with frequency ω and the
forced part with frequency . Also, an approximate maximum value of the DLF is obtained
when and , i.e.
(2.24)
(2.25)
Page 39
(2.26)
Thus, the dynamic displacement diverges, but only after a finite number of oscillations. Also,
Figure 2.9 plots the DLF for the case where we see that the total factor, despite being
the superposition of two harmonic functions, is no longer harmonic but only a periodic
function of time.
(2.27)
(2.28)
We also define the coefficient of damping and the damped natural frequency as follows:
(2.29)
(2.30)
(2.31)
(2.32)
The displacement is no longer a periodic function of time and the oscillator simply returns to
its original position without executing any vibrations. From the condition
(2.33)
(2.34)
It should be noted here that the coefficient of damping is seldom used nowadays, with
preference given to damping ratio . Obviously, the two coefficients are related as
The effect of damping on the natural frequency is minimal; for instance, a 10 per cent of
critical damping ratio yields . It is rare to find civil engineering structures
exhibiting anything close to critical damping, although many mechanical components (such as
shock absorbers) do.
Damping can be experimentally measured by tracing the logarithmic decrement (i.e. the log
of the difference between two consecutive peaks in a displacement versus time plot for free
vibrations). Referring to Figure 2.10, we have the logarithmic decrement as
Page 41
(2.36)
For instance, when per cent, d is equal to 0.2π and the ratio of two consecutive
peaks is exp(0.2π)=1.87. Thus, a damping ratio of 10 per cent reduces the dynamic
displacements by a factor of 0.534 during each vibration cycle.
(2.37)
The complete expression in terms of a Duhamel integral can be found through time
integration of the above impulse, to which the effect of initial conditions is subsequently
superimposed. Thus,
(2.38)
As a special case consider F(t)=F1; substitution of the time function f(t)=1 in eqn (2.38) gives
the solution for a suddenly applied and maintained load as
(2.39)
Comparing Figures 2.5 and 2.11 clearly shows the effect of damping in reducing the
Page 42
Figure 2.11 Forced vibration with damping due to suddenly applied and maintained load.
amplitude of the dynamic displacements and in bringing about, after some time, quasi-static
conditions.
(2.40)