Vous êtes sur la page 1sur 9

MECHANICS

RESEARCH COMMUNICATIONS

Mechanics Research Communications 33 (2006) 787–795


www.elsevier.com/locate/mechrescom

Estimating equivalent viscous damping ratio for RC


members under seismic and blast loadings
Binggeng Lu *, Pedro F. Silva
Department of Civil, Architectural and Environmental Engineering, University of Missouri-Rolla,
Rolla, MO 65409-0710, United States

Available online 30 May 2006

Abstract

It can be learned that the equivalent viscous damping (EVD) ratio as a function of displacement ductility is a crucial
parameter in the application of the displacement based method. A set of generalized expressions are proposed herein for
estimating the EVD ratio for individual and multiple reinforced concrete (RC) members. For individual RC members
under seismic loads, the EVD ratio was derived based on their hysteretic response and hysteretic energy dissipated under
fully reversed cyclic loading. Because of the nature of blast loads only the energy dissipated up to maximum displacement
was considered to estimate the EVD ratio for individual RC members under blast loads. For multiple degree of freedom
systems consisting of RC members, the EVD ratio was derived based on equating the total energy dissipated in the system
to the sum of the energy dissipated by its individual members. Analytical studies presented here and elsewhere indicate that
the EVD ratio is highly dependent on the damaged displacement ductility, which can be directly correlated to damage.
These analytical results along with the expressions to compute the EVD ratio for RC members under seismic and blast
loads are presented and discussed in this paper.
Ó 2006 Published by Elsevier Ltd.

Keywords: Blast loads; Displacement based method; Ductility; Equivalent viscous damping; Seismic loads

1. Introduction

One of the crucial parameters in applying the displacement based design (DBD) method for any given load-
ing type is the equivalent viscous damping (EVD) ratio. Most commonly used in seismic design, the EVD ratio
for individual reinforced concrete members is estimated from their hysteretic response under fully reversed
cyclic loading (Clough and Penzien, 1993). This procedure was implemented in this study to develop general-
ized expressions for estimating the EVD ratio for seismic loads. However, under blast loads the EVD ratio was
estimated using only the energy dissipated up to the first peak displacement.
For seismic loads, the hysteretic response of RC members was expressed in terms of variables that are cor-
related to the axial load applied on the member and the achieved displacement ductility. Next, the EVD ratio

*
Corresponding author. Tel.: +573 341 5852.
E-mail address: bl2t9@umr.edu (B. Lu).

0093-6413/$ - see front matter Ó 2006 Published by Elsevier Ltd.


doi:10.1016/j.mechrescom.2006.05.002
788 B. Lu, P.F. Silva / Mechanics Research Communications 33 (2006) 787–795

was obtained as a function of these variables. Analytical results indicate that the EVD ratio significantly
depends on the axial load and the displacement ductility. Accordingly, a decrease in the axial load tends to
increase the EVD ratio, and an increase in the displacement or curvature ductility also increases the EVD
ratio. These results are consistent with those reported in the literature (Clough and Penzien, 1993; Priestley
et al., 1995).
Several researchers have concluded that the loading pattern, numbers of cycles, and loading rate affect the
yield and ultimate displacement, and load capacity of reinforced concrete members (Lehman and Moehle,
2000; Otani, 1981). Since these properties are significantly different under seismic and blast loads, in this paper,
it is proposed that the EVD ratio for blast loads should also be different than those used in seismic design.
Other differences related to the response of RC members under seismic and blast loads that are likely to
affect the EVD ratio are: (1) under seismic load, a structure is subjected to multiple reversed cyclic accelera-
tions of increasing magnitude up to the peak ground acceleration and the structure exhibits corresponding
levels of damage during the excursion to the peak displacement cycle; and (2) under blast load, a structure
is initially undamaged and it is subjected to a rapid monotonic increase in the applied pressure wave (Bimha,
1996; Bulson, 1997). Based on the hysteretic response and the energy dissipated in a RC member under these
loading conditions, generalized expressions are proposed for estimating the EVD ratio for seismic and blast
loads as a function of the displacement ductility level for individual RC members.
Analytical results presented in this paper indicate that under seismic loads the EVD ratio is a function of
the applied axial load, but under blast loads the EVD ratio is not a function of the applied axial load. In some
cases, as reported in this paper, it can be shown that the EVD ratio is higher under blast loads than under
seismic loads. This is within conclusions recorded by other researchers that indicate the curvature ductility
capacity of RC members is higher under high strain rates typical of blast load effects than under low strain
rate effects typical of seismic load effects (Ngo et al., 2003). This needs to be further examined by experimental
evidence in order to validate the effect of the EVD ratio under high strain rate effects. In this paper, only ana-
lytical studies are presented and discussed.
Up to date no expression has been developed or recommendations have been proposed that can be used to
estimate the appropriate level of EVD ratio for the design/assessment of systems with multiple members that
display different hysteretic responses. As such, a mathematical procedure is proposed in this paper that can be
used to estimate the EVD ratio for these systems. For multiple members, the EVD ratio was derived based on
equating the total energy dissipated in the system to the sum of the energy dissipated by its individual RC
members. In this context the total energy refers primarily to the hysteretic energy dissipated in the plastic hinge
zones plus the recoverable elastic strain energy. Detailed mathematical development of this expression is also
presented in this paper.

2. Estimating the equivalent viscous damping ratio – single members

According to the DBD method it is feasible to select a desirable damage performance criteria for a given
member or structure under a given hazard level (Priestley, 2000). This is an attractive feature of the DBD
method, as engineers can select the desired structural performance criteria before the design of any structural
member. Selection of performance goals can be related directly to displacement ductility levels, which in turn
can be associated with the EVD ratio (Hose et al., 2000). It is these relations between the structural perfor-
mance, displacement ductility, and EVD ratio that are necessary in applying the DBD method for the design
or assessment of seismic or blast resistant structures.
An analytical approach that can be used for estimating the EVD ratio, neq, as a function of the displace-
ment ductility, l, is (Clough and Penzien, 1993)
1 ED
neq ¼ ð1Þ
4p ES
According to this expression it is feasible to relate the EVD ratio to the displacement ductility, provided
that a well defined nonlinear model can be used to establish the response of a RC member. Based on the non-
linear model, the displacement ductility can be expressed as a function of ED and ES. In Eq. (1), the area
within the inelastic force-displacement response curve, ED, is a measure of the hysteretic damping or energy
B. Lu, P.F. Silva / Mechanics Research Communications 33 (2006) 787–795 789

dissipation capacity of the member, and ES depicts the recoverable elastic strain energy stored in an equivalent
linear elastic system. Subsequently, using Eq. (1) the EVD ratio is finally expressed as a function of the dis-
placement ductility.
In this paper, the generalized expression for estimating the EVD ratio, as a function of the displacement duc-
tility, was developed by using Eq. (1) in conjunction with the modified Takeda hysteretic model, which is
depicted in Fig. 1. This model is widely used to represent the cyclic response of reinforced concrete members
(Carr, 2000; Otani, 1981). However, similar expressions can be developed for other nonlinear hysteretic models.
Furthermore, an attractive feature of using the DBD method is that the complex hysteretic structural model
shown in Fig. 1 can be transformed into an equivalent linear elastic structural model, provided that the peak
displacement is nearly the same for both modeling conditions, namely, the nonlinear and the linearized system.
This transformation is often designated as the substitute structure approach (SSA) (Priestley et al., 1995).

2.1. EVD for Seismic Loads

In deriving expressions for the EVD ratio as a function of the displacement ductility, it is important to rec-
ognize that under seismic loads, any given structure is subjected to multiple reversed displacements of increas-
ing magnitude up to the peak displacement. An example of this observation is depicted in Fig. 2 for a far field
seismic event.
As shown in this figure, before the maximum displacement is reached in either direction, several reloading
and unloading cycles occur, leading to a successive reduction in the reloading and unloading stiffness before
the peak displacement is reached. This also indicates that under the loading cycle to the peak displacement
there exists increasing levels of structural damage that can be related to the hysteretic response of a RC mem-
ber. This phenomenon is well observed in the experimental testing of RC members under fully reversed cyclic
loading (Hose et al., 2000), which analytically can be well simulated by the modified Takeda hysteretic model.
As such, this model was selected for evaluation of the EVD ratio.
According to the modified Takeda hysteretic model (see Fig. 1), three parameters (a, b and r) are used to
completely represent the nonlinear hysteretic response of reinforced concrete members under fully reversed
cyclic loading. These parameters are as follows: a describes the unloading stiffness and it is in the range of zero
to 0.5, b describes the reloading stiffness and varies from zero to 0.6, and r describes the post-yielding stiffness.
It can be shown that increasing a decreases the unloading stiffness, and increasing b increases the reloading
stiffness. Assigning specific values to a and b, it is possible to duplicate the response of a reinforced concrete
member under different levels of axial load. This is exemplified next. Two nonlinear hysteretic representations
for reinforced concrete members under different levels of axial loads are shown in Fig. 3.
Referring to Fig. 3, point A corresponds to the beginning of a cycle and represents the accumulated damage
due to the previous progressive reversed cycles. Setting a and b equal to zero and 0.6, respectively, leads to the
hysteretic response shown in Fig. 3(a). This hysteretic response is typical of RC members subjected to low

Dp
F b Dp

FY+
α
⎛Δ ⎞
Ko K U = Ko ⎜⎜ Y ⎟⎟
DY ⎝ ΔU ⎠

DU D

KU

FY−
rKo

Fig. 1. Modified Takeda hysteretic model.


790 B. Lu, P.F. Silva / Mechanics Research Communications 33 (2006) 787–795

Displacement Ductility, m
2

-1

-2
0 10 20 30
Time, t (sec)
Fig. 2. Time history of displacement ductility – seismic loads.

Fig. 3. Modified Takeda model – seismic loads. (a) Low axial load ratios (Low P) and (b) high axial load ratios (High P).

axial load levels and matches results reported in the literature (Priestley et al., 1995). On the other hand, select-
ing a to 0.5, and b to zero leads to the hysteretic response shown in Fig. 3(b). In this case, these values are more
efficient to represent a reinforced concrete member subjected to an axial load ratio of 5–10 percent (Silva and
Seible, 2001).
For both cases in Fig. 3, the total hysteretic energy dissipated during a fully reversed cycle, ED, (indicated
by the area inside the heavy lines), is bounded by the reloading and unloading stiffness representing the
selected cycle. In conjunction with the modified Takeda model presented in Fig. 1, the hysteretic energy dis-
sipated during the selected cycle, ED, is
ED ¼ K 0 D2Y ð½1 þ rð1  bÞðl  1Þf2l þ bðl  1Þ  la ½1 þ rðl  1Þg
2 2
þ rb2 ðl  1Þ  la ½1 þ rðl  1Þ Þ ð2Þ
Similarly, the recoverable elastic strain energy, ES, is:
1
ES ¼ K 0 lD2Y ½1 þ rðl  1Þ ð3Þ
2
Next, substituting Eqs. (2) and (3) into Eq. (1), and assuming that the external or structural damping ratio
is 0.05, the EVD ratio for seismic loads is:
1 ½1 þ rð1  bÞðl  1Þf2l þ bðl  1Þ  la ½1 þ rðl  1Þg
neq ¼ 0:05 þ
2p l½1 þ rðl  1Þ
!
2 2
rb2 ðl  1Þ  la ½1 þ rðl  1Þ
þ ð4Þ
l½1 þ rðl  1Þ
B. Lu, P.F. Silva / Mechanics Research Communications 33 (2006) 787–795 791

50 50
r =0
r =0.01
40 40 r =0.02

Dapming Ratio, x (%)

Dapming Ratio, x (%)


r =0.05
30 30

20 20
r =0
10 r =0.01 10
r =0.02
r =0.05
0 0
0 2 4 6 8 10 0 2 4 6 8 10
(a) Ductility Index, m (b) Ductility Index, m

Fig. 4. EVD ratios vs. ductility – seismic loads. (a) Low axial load ratios (Low P) and (b) high axial load ratios (High P).

In this expression, all the three parameters (a, b and r) used to describe the nonlinear response of a rein-
forced concrete member are involved. Next, Eq. (4) was evaluated for different axial loads applied on a RC
member. As previously stated, for reinforced concrete members under low axial load levels (see Fig. 3(a))
the hysteretic response of the member can be represented by setting a to zero and b to 0.6, and Eq. (4) reduces
to
 
1 1:3  0:7r rðl  1Þ
neq ¼ 0:05 þ 1:3   0:7r  0:6 ð5Þ
p l 1 þ rðl  1Þ
Similarly, for members under high axial load levels, such that the hysteretic response of the member can be
represented by setting a to 0.5 and b to zero (see Fig. 3(b)), Eq. (4) reverts to
 
1 1r pffiffiffi
neq ¼ 0:05 þ 1  pffiffiffi  r l ð6Þ
p l
For members under these hysteretic conditions, an expression typically used in the literature for computing
the EVD ratio as a function of the displacement ductility was given by Thorkildsen et al. (1995), as
 
1 0:95 pffiffiffi
neq ¼ 0:05 þ 1  pffiffiffi  0:05 l ð7Þ
p l
This equation is a special case of Eq. (6) where the post-yield stiffness ratio, r, is 0.05. Eqs. (5) and (6) where
then plotted as a function of the displacement ductility, l, for different values of the post-yield stiffness ratio, r,
and results are presented in Fig. 4.
From Fig. 4, it can be seen that the EVD ratio reduces as r increases, and this reduction is significantly more
accentuated for higher values of l. In addition, it is clear that the EVD ratio is higher for members subjected
to low axial load levels. This statement follows observations recorded in Fig. 3, where it can be observed that
the energy dissipated under low axial loads is higher than the energy dissipated under high axial loads, which is
represented by the area inside the solid heavy lines. Since in both cases the recoverable elastic strain energy is
the same, this indicates that any RC member with a higher hysteretic energy dissipation capacity also has a
higher EVD ratio. This suggests that in applying the DBD method one must carefully select the appropriate
relation between neq and l based on the axial load present in the member.

2.2. EVD for Blast Loads

In order to develop the EVD ratio for blast loads the following was considered. As depicted in Fig. 5, blast
loads impart a response on a RC member that is mostly characterized by a rapid monotonic increase in the
displacement, which is consistent with results reported in the literature (Longinow and Alfawakhiri, 2003;
Mays and Smith, 1995).
792 B. Lu, P.F. Silva / Mechanics Research Communications 33 (2006) 787–795

Displacement Ductility, m
-1

-2

-3

-4

-5
0 1 2 3
Time, t (sec)

Fig. 5. Time history of displacement ductility – blast loads.

Under blast loads, the maximum displacement occurs during the first increasing segment and attenuates
rapidly thereafter. Unlike the seismic case, there are no unloading and reloading phases during the path up
to the peak displacement and, as a result, there is no previous accumulated damage under blast loads. As such,
for blast loads the EVD ratio was computed such that the energy dissipated in going from the system initial
condition to the peak displacement was the same for the nonlinear (see Fig. 6) and the linearized system.
As shown in Fig. 6, the total energy dissipated during the first increasing segment, EN, is the area under the
response function up to the displacement, Da, and is given by
1
EN ¼ K 0 D2y fl þ ðl  1Þ½1 þ rðl  1Þg ð8Þ
2
It can be shown that the energy dissipated in the corresponding linearized system, Eeq, is
1 p
Eeq ¼ K 0 D2y l½1 þ rðl  1Þ þ neq K 0 D2y l½1 þ rðl  1Þ ð9Þ
2 2
In Eq. (9) the first term is the area under the linearized system, and the second term is the area dissipated by
an equivalent viscous damping system. Since the total energy dissipated by the equivalent structure, Eeq, must
equal the energy dissipated by the nonlinear structure, EN, the EVD ratio was obtained by equating Eq. (8) to
Eq. (9). As before, setting the external or structural damping ratio to 0.05, the EVD ratio for blast loads can be
obtained as:
1 ð1  rÞðl  1Þ
neq ¼ 0:05 þ ð10Þ
p l½1 þ rðl  1Þ
Eq. (10) was plotted as a function of the displacement ductility, l, in Fig. 7. As such, selection of the proper
relation between neq and l is also relevant for the DBD method applied to the design or assessment of blast
resistant structures.

Fa
FY rKo
Ko
EN

D
0 DY Da = m DY

Fig. 6. Nonlinear system – blast loads.


B. Lu, P.F. Silva / Mechanics Research Communications 33 (2006) 787–795 793

40

Dapming Ratio, x (%)


30

20

r =0
10 r =0.005
r =0.01
r =0.05
0
0 2 4 6 8 10
Ductility Index, m

Fig. 7. EVD ratios vs. ductility – blast loads.

Comparing Fig. 7 to Fig. 4, it is clear that the EVD ratios for members under seismic loads and low axial
loads levels are higher than for blast loads; however, under high axial load levels the EVD ratio is smaller for
seismic loads than for blast loads.
It has been observed that high rate strain effects have a marked influence on the mechanical properties of
concrete and steel. For example, the strength of both concrete and steel increases by 10 to 30% but no signif-
icant increase on the initial stiffness has been observed as the modulus of elasticity of both steel and concrete
remains insensitive to the loading rate (Mays and Smith, 1995). Ngo et al. (2003) have also reported an
increase in the ultimate strain of concrete but not of steel. Furthermore, Ngo et al. (2003) have conducted
moment curvature analysis of RC members under blast loads, and have shown an increase in the curvature
ductility capacity of RC members under high strain rates typical of blast loads.

2.3. Estimating the equivalent viscous damping ratio – multiple members system

In the previous two sections the EVD ratio for individual RC members was derived based on the hysteretic
response of RC members. Of equal interest is the amount of EVD ratio that should be used in the design or
assessment of structures with multiple members that are subjected to either different levels of displacement
ductility or axial loads; and, furthermore, display different hysteretic responses. In this section the system
EVD ratio is derived based on equating the total energy dissipated in the system to the sum of the energy
dissipated by its individual members. Total energy refers primarily to the hysteretic energy dissipated in the
plastic hinge zones plus the recoverable elastic strain energy. Detailed mathematical development of these
expressions is presented next.
For seismic case, considering a system with n number of members, and using Eq. (1), the hysteretic energy
dissipated by the i-th member, EDi, is
EDi ¼ 4pneqi ESi ð11Þ
Accordingly, the hysteretic energy dissipated by a complete system, EDS, with an equivalent damping ratio,
neqS, and the system, EDS, defined as the summation of all the individual members, are
Xn X n
EDS ¼ EDi ¼ 4pneqS ESi ¼ 4pneqS ES ð12Þ
i¼1 i¼1

Solving for neq  S in the equations above then


Pn
i¼1 neqi E Si
neqS ¼ P n ð13Þ
i¼1 E Si

Similarly from Eq. (3), the recoverable elastic strain energy of the ith member, ESi, is:
1
ESi ¼ K 0i li D2Yi ½1 þ ri ðl  1Þ ð14Þ
2
794 B. Lu, P.F. Silva / Mechanics Research Communications 33 (2006) 787–795

Next, stipulating that the ductility of the system is the same as the ductility of the member that reaches
yielding first, DY1, the nominal ductility level of the system, l, and of the ith member are
DS
l¼ ð15Þ
DY 1
DS
l¼ ð16Þ
DYi
If a structure consists of n members with different yield displacements and initial stiffness such that the yield
displacements of the individual members are organized in ascending order with: DY1 6 DY2 6    6 DYi: than,
substituting Eqs. (15) and (16) into Eq. (14), the EVD ratio for the system, neqS, can be estimated by
P h  i
neqi ki si 1 þ ri sli  1
neqS ¼ P h  i ð17Þ
ki si 1 þ ri sli  1

where the factors si ¼ DDYYi1 and ki ¼ KK010i were used in Eq. (17) for simplicity. Furthermore, if all the members
have the same initial stiffness, K0i and the same yield displacement, DYi, then the EVD ratio reduces to:
neqS ¼ neq1 ¼     ¼ neqi ¼    ¼ neqn ð18Þ

Although same procedures can be applied to blast loads mathematically, in the case of blast loads, the input
energy can be concentrated at a specific location and, depending on the natural period of the structure and
duration of the pressure wave, blast loads can produce mainly local damage to individual members (Ali,
2002; DoD, 2003; FEMA, 2003). Consequently, in many cases it may not be necessary to apply the total sys-
tem damping because one must also consider the location of the applied blast load, which will determine if
individual members or the entire structure will resist the applied blast load.

3. Conclusions

A set of generalized expressions were developed in this study that can be used to estimate the EVD ratio as
a function of the displacement ductility for individual and multiple reinforced concrete members subjected to
seismic loads and different levels of axial load. These expressions were developed based on the hysteretic
response of reinforced concrete members under reversed cyclic loading using the modified Takeda hysteretic
model; however, similar expressions can be developed for other hysteretic models. Similarly, an expression for
the EVD ratio for individual RC members under blast loads was derived based on equating the energy dissi-
pated by the nonlinear model to the energy dissipated by an equivalent damped linear elastic model.
Analytical studies were conducted using the derived expressions that showed that the EVD ratio for seismic
loads is not only directly related to the displacement ductility level, but also highly dependent on the axial load
level. However, under blast loads the EVD ratio is a function of the displacement ductility only.
Results showed that the EVD ratios for members under seismic loads and low axial loads levels are higher
than for blast loads. However, under high axial load levels the EVD ratio is smaller for seismic loads. As such,
selection of the proper relation between the EVD ratio and the displacement ductility must consider the axial
load and loading type.

References

Ali, M.M., 2002. Protective Design of Concrete Buildings Under Blast Loading. In: Proceedings of the 7th International Conference on
Structures Under Shock and Impact – SUSI VII. WIT Press, Wessex Institute of Technology, UK, pp. 23–33, May.
Bulson, P.S., 1997. Explosive Loading of Engineering Structures. E&FN Spoon an imprint of Chapman & Hall, London, UK, pp. 236.
Bimha, R., 1996. Response of Thin Circular Plates to Blast Loading, Master Thesis, University of Cape Town, Cape Town, South Africa,
1996. pp. 260–267.
Carr, A.J., 2000. Ruaumoko: Program for Inelastic Dynamic Analysis, User’s Manual. Department of Civil Engineering, University of
Canterbury, Christchurch, NZ, pp. 234.
Clough, R.W., Penzien, J., 1993. Dynamics of Structures, second ed. McGraw Hill Inc., New York, NY, pp. 635.
B. Lu, P.F. Silva / Mechanics Research Communications 33 (2006) 787–795 795

Department of Defense, (DoD) 2003. DoD Minimum Antiterrorism Standards for Buildings, Uniform Facilities Criteria: UFC 4-010-01,
October 2003, pp. 53.
Federal Emergency Management Agency, (FEMA) 2003. FEMA 426- Reference Manual to Mitigate Potential Terrorist Attacks Against
Buildings. FEMA-426, December 2003.
Hose, Y.D., Silva, P.F., Seible, F., 2000. Performance evaluation of concrete bridge components and systems under simulated seismic
loads. EERI Earthquake Spectra 16 (2), 413–442.
Lehman, D.E., Moehle, J.P., 2000. Performance-based seismic design of reinforced concrete bridge columns. In: Twelfth World
Earthquake Engineering Conference, New Zealand, January 2000, pp. 215–223.
Longinow, A., Alfawakhiri, F., 2003. Blast resistant design with structural steel. Modern Steel Construction (October), 61–65.
Mays, G.C., Smith, P.D., 1995. Blast Effects on Buildings. Thomas Telford Service Ltd., London, E14 4JD, 121 pp.
Ngo, T.D., Mendis, P.A., Teo, D., Kusuma, G., 2003. Behaviour of high-strength concrete columns subjected to blast loading. In:
Proceedings of the inaugural International Conference on Advances in Structures: Steel, Concrete, Composite and Aluminium
(ASSCCA’03), Sydney, Australia.
Otani, S., 1981. Hysteresis models of reinforced concrete for earthquake response analysis. J. Faculty Engineering 36 (2), 125–159.
Priestley, M.J.N., 2000. Performance based seismic design. In: Proceedings of the 12th World Conference on Earthquake Engineering,
Auckland, New Zealand, 2000, State of the Art Paper No. 2831, pp. 325–346.
Priestley, M.J.N., Seible, F., Calvi, M., 1995. Seismic Design and Retrofit of Bridges. John Wiley & Sons, Inc., New York, NY, September,
pp. 672.
Silva, P.F., Seible, F., 2001. Seismic performance evaluation of CISS piles. ACI Structural Journal 98 (1), 36–49.
Thorkildsen, E., Kowalsky, M., Priestley, M.J.N., 1995. Use of lightweight concrete in seismic design of California bridges. In:
Proceedings, International Symposium on Structural Lightweight Aggregate Concrete, Sandefjord, Norway, June 1995.

Vous aimerez peut-être aussi