Sunteți pe pagina 1din 24

EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS

Earthquake Engng Struct. Dyn. 2001; 30:14391462 (DOI: 10.1002/eqe.71)

A displacement-based seismic design procedure for RC


buildings and comparison with EC8
T. B. Panagiotakos and M. N. Fardis
Structures Laboratory; Department of Civil Engineering; University of Patras; Patras; Greece

SUMMARY
A procedure for displacement-based seismic design (DBD) of reinforced concrete buildings is described
and applied to a 4-storey test structure. The essential elements of the design procedure are: (a) proportioning of members for gravity loads; (b) estimation of peak inelastic member deformation demands in
the so-designed structure due to the design (life-safety) earthquake; (c) revision of reinforcement and
@nal detailing of members to meet these inelastic deformation demands; (d) capacity design of members
and joints in shear. Additional but non-essential steps between (a) and (b) are: (i) proportioning of
members for the ULS against lateral loads, such as wind or a serviceability (immediate occupancy)
earthquake; and (ii) capacity design of columns in Cexure at joints. Inelastic deformation demands in
step (b) are estimated from an elastic analysis using secant-to-yield member stiEnesses. Empirical expressions for the deformation capacity of RC elements are used for the @nal proportioning of elements
to meet the inelastic deformation demands. The procedure is applied to one side of a 4-storey test
structure that includes a coupled wall and a two-bay frame. The other side is designed and detailed
according to Eurocode 8. Major diEerences result in the reinforcement of the two sides, with signi@cant
savings on the DBD-side. Pre-test calculations show no major diEerence in the seismic performance of
the two sides of the test structure. Copyright ? 2001 John Wiley & Sons, Ltd.
KEY WORDS:

displacement-based design; deformation capacity; Eurocode 8; equal displacement rule; RC


buildings

1. INTRODUCTION
In displacement-based procedures for the seismic design of new structures or the evaluation of
existing ones, seismic displacements are the primary response variables for the design or the
evaluation. This means that design or acceptance criteria and capacity-demand comparisons
are expressed in terms of displacements rather than forces. Since their introduction in the early
1990s [1; 2], displacement-based concepts have found their way more into seismic evaluation

Correspondence to: M. N. Fardis, Department of Civil Engineering, University of Patras, P.O. Box 1424, University
Campus, Patras GR-26500, Greece.

Copyright ? 2001 John Wiley & Sons, Ltd.

Received 16 November 1999


Revised 15 July 2000 and 9 January 2001
Accepted 7 February 2001

1440

T. B. PANAGIOTAKOS AND M. N. FARDIS

or assessment of existing structures [36], than into the design of new ones [714]. For existing structures, application of displacement-based concepts is more straightforward in that the
geometry of the structure and the reinforcement are known and can be used as input to either
simple or advanced analysis procedures for the estimation of member inelastic displacement
and deformation demands throughout the structure, to be compared with the corresponding
deformation capacities [3; 4].
The state-of-the-art of displacement-based design (DBD) for new structures is reCected
in Appendix I (Tentative Guidelines for Performance-Based Seismic Engineering)Part B
(ForceDisplacement Approach) of the 1999 SEAOC Blue Book [15]. Appendix I refers to a
Direct DBD procedure and an Equal-displacement-based (EBD) one. The former, proposed
and advocated by Priestley and co-workers [712], uses a substitute elastic structure to relate
displacement demands to the eEective period at peak response. The EBD procedure uses
instead the equal displacement rule to relate peak displacements to the period of the cracked
elastic structure. These DBD procedures hold great promise for seismic design codi@cation
and practice, especially after feedback from their application to real cases is received and after
the code-speci@ed drift limits, global ductility factors and damping values are further re@ned
and rationalized.
In this paper, another DBD approach is brieCy described and exempli@ed through its application to a 4-storey dual test structure. It has evolved from earlier proposals of the authors [16; 17] for RC frames and diEers from most other DBD procedures (e.g. References
[711; 15]) mainly in that: (a) (displacement-based) seismic design is integrated with ultimate
limit state (ULS) and serviceability (SLS) design for other loads, such as factored gravity
and wind; and (b) local seismic displacement and deformation demands are used directly for
member proportioning and detailing, without conversion to strength demands and recourse to
force-based proportioning.
2. DESIGN PROCEDURE
2.1. Overview of the design procedure
The proposed DBD procedure comprises the following steps:
(1) Proportioning of longitudinal and transverse reinforcement of all members on the basis
of: (a) ULS and SLS design in Cexure and shear for factored gravity (and wind) loads
and for a serviceability (operational level) earthquake combined with the simultaneously acting (arbitrary-point-in-time) gravity loads; (b) capacity design of columns
in Cexure at beam-column joints and of all members in shear; and (c) minimum reinforcement for structures without earthquake resistance.
(2) Estimation of inelastic chord rotation demands at member ends under the life-safety
earthquake, through a 5 per cent-damped elastic analysis with yield point member
stiEness for antisymmetric bending.
(3) Revision of reinforcement and of detailing, so that member chord rotation capacities
exceed the (possibly factored) seismic demands computed in step 2.
(4) Capacity design of joints in shear and reevaluation of capacity design of: (a) columns
in Cexure at joints; and (b) of all members in shear, on the basis of the @nal member
longitudinal reinforcement from step 3.
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

DBD PROCEDURE FOR RC BUILDINGS

1441

In step 1, the usual sequence of seismic design is followed. Beams are designed for Cexure
considering minimum reinforcement and taking into account bar development within joints
for the selection of longitudinal bar sizes; beams are capacity designed in shear; columns are
capacity designed for Cexure at joints considering minimum reinforcement; columns are capacity designed for shear; walls are designed for Cexure considering minimum reinforcement;
walls are capacity designed in shear for overstrength over Cexural hinge formation at the base.
Capacity design of joints in shear is left for step 4.
Completion of step 1 is necessary for step 2, as the eEective stiEness at yielding of members depends on their longitudinal reinforcement. In step 3, normally only the compression
steel and the stirrups of members may need to be revised. An increase in the compression
reinforcement may require that the capacity design calculations in step 4 be repeated. Normally, changes in reinforcement from steps 3 and 4 will not have a major eEect on member
eEective stiEnesses used in step 2. Even if they do, any increase in member stiEnesses will
reduce chord rotation demands from step 2 and increase the safety margin provided by step
3 in the @rst round. Therefore, unless chord rotations demands from step 2 cannot be met in
step 3 through changes in the reinforcement but require also revision of member dimensions,
iterations are not needed.
Step 1 is similar to the current design of RC structures for nonseismic and (reduced) seismic
loads. The only diEerence is that the stringent detailing imposed by current seismic codes for
member ductility is replaced by normal detailing for non-earthquake-resistant structures. In
lieu of prescriptive detailing for member ductility the procedure provides in step 3 for explicit
member veri@cation against the seismic chord rotation demands.
Design for a serviceability earthquake in step 1a is not an essential part of the procedure,
especially if the structure is designed for wind. Its inclusion results in a two-level seismic
design procedure, in-between: (1) the current codi@ed seismic design (e.g. Reference [18])
based on: (a) ULS proportioning of members for a rare earthquake including a force reduction factor R and (b) drift limits for damage limitation under a frequent earthquake, and (2)
the full four-level seismic design of performance-based seismic engineering [4; 15; 19].
Capacity design of columns in Cexure at beam=column joints is not an essential component
of the proposed procedure. It can be replaced by explicit veri@cation of the deformation
capacity of columns under a very rare [19] or maximum considered [15] earthquake along
the lines of Section 2.3. In the absence of such an explicit veri@cation, capacity design of
columns in Cexure at joints is used here to ensure the formation of a controlled and stable
inelastic mechanism and prevent collapse under such large-magnitude events.
2.2. Estimation of inelastic displacement and deformation demands under
the design level earthquake
A procedure was proposed in Reference [20] for the estimation of peak inelastic chord rotation
demands at member ends in RC frame structures. It is similar to the coeNcient-method in
Reference [4] and the EBD procedure in Part B of Appendix I in Reference [15], as it
uses linear elastic analysis and the 5 per cent-damped elastic spectrum of the design level
earthquake. The analysis can be equivalent static, with lateral forces from a postulated
linear mode shape, or preferably multimodal (dynamic) with CQC combination of modal
contributions. Member stiEnesses, EI, are taken equal to the secant stiEness at yielding at both
member ends in antisymmetric bending. With such member stiEnesses the fundamental period
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

1442

T. B. PANAGIOTAKOS AND M. N. FARDIS

lies in the velocity-controlled part of the spectrum and the equal-displacement rule applies,
giving on the average good approximation of the peak inelastic chord rotations at member
ends.
Results in Reference [20] do not include an overall period-dependent correction as in References [4; 15], but multiplicative factors on chord rotations from the linear analysis, for
conversion to a mean (expected) value, Em , or to a 95 per cent-fractile, Ek; 0:95 , of peak
inelastic chord rotations. These factors are diEerent for beams and columns and depend on
whether the linear analysis is equivalent static or multimodal. They increase linearly from
the base to the top of the structure and are on the average around 1 for Em or 1.5 for Ek; 0:95 .
The diEerence between the mean and the 95 per cent-fractile reCects scatter due to: (a) details
of a ground motion conforming to a smooth elastic spectrum; and (b) model uncertainty due
to estimation of inelastic deformations from elastic analysis.
The secant-to-yield stiEness of a member in antisymmetric bending is
EI =

L My
6 y

(1)

in which L denotes the clear length of the member and My , y , the yield moment and the
corresponding chord rotation at the member end. For elements with unsymmetric section
(beams) two values of EI are calculated from Equation (1) at each end, one for positive
moment and another for negative. The slab width considered to be eEective as Cange on each
side of the beam and contributing with its reinforcement is taken to be equal to one-quarter
the beam span, L or half the distance to the nearest parallel beam, whichever is smaller. The
two values computed from Equation (1) at the two ends are averaged into a single EI-value
of the member.
The moment My at yielding of the tension steel can be computed from @rst principles. The
corresponding chord rotation, y , can be estimated from a semiempirical relation such as [21]
y =

fy2

y L

+ 0:0025 + 0:25db
6
Es (d d ) fc

(2)

In Equation (2)
y is the yield curvature (computed from @rst principles), h the depth of
the member, db the mean diameter of tension reinforcement, d d the distance between the
tension and compression steel and fy , Es and fc (all in MPa) the yield strength and elastic
modulus of longitudinal steel and the concrete strength. The second term accounts for the
eEects of shear and the last one reCects the end rotation due to reinforcement slip from its
anchorage beyond the member end.
It is clear from Equations (1) and (2) that knowledge of the longitudinal reinforcement
from step 1 of the design procedure is essential for the calculation of the eEective stiEness
of members for input to step 2.
The procedure was developed in Reference [20] on the basis of over one thousand non-linear
analyses of RC frame buildings from three to twelve storeys, all fairly regular in plan and
elevation. At beam-column joints the sum of column Cexural capacities was greater or slightly
less than that of beam capacities. The capacities of all members were controlled by Cexure
rather than shear and brittle failures were unlikely. Under these conditions, not uncommon in
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

DBD PROCEDURE FOR RC BUILDINGS

1443

earthquake-resistant buildings, signi@cant concentrations of inelasticity were not observed in


any single storey, even under motions exceeding the design earthquake by a factor of 2.
2.3. Detailing of RC members on the basis of inelastic deformation demands
In step 3, the longitudinal and transverse reinforcement at member ends is revised and detailed
so that the corresponding chord-rotation capacity, u , exceeds the demand from step 2. To
this end an expression is needed for the ultimate chord rotation of members in terms of their
geometric and mechanical characteristics (including the reinforcement). Well-known expressions based on section curvature and the notion of plastic hinge length (as e.g. in References
[13; 14] for walls), or empirical expressions for u , such as those in References [22; 23] may
be used to this end. At this instance, the proposed procedure employs the following expression for u , found to @t best a total of 878 tests to failure (242 monotonic, 636 cyclic tests)
of beam (288, unsymmetrically reinforced, without axial force), column (526, symmetrically
reinforced, with or without axial force) or wall (64) specimens [21]:

0:275  0:45

asl 
max(0:01; !2 ) 
Ls
fc
awall (0:2 )
1:1100!wx 1:3d
um = ast acyc 1 +
2:3
max(0:01; !1 )
h

(3)

with ast being the coeNcient for the steel of longitudinal bars, equal to 0.015 for ductile hotrolled steel, 0.0125 for heat-treated tempcore steel, or 0.008 for brittle cold-worked steel,
acyc the coeNcient for the type of loading, equal to 1 for monotonic and 0.6 for cyclic
loading, asl the coeNcient for the slip of longitudinal bars, equal to 1 if there is slip of the
longitudinal bars from the anchorage beyond the member end, or 0 if there is not, awall a
coeNcient, equal to 2=3 for shear walls or 1 for beams or columns,  = N=Ac fc the axial load
ratio, positive for compression, !1 ; !2 the mechanical reinforcement ratios, fy =fc , of the
tension and compression longitudinal steel not including any diagonal bars; (in walls all the
vertical web reinforcement is included as tension steel), fc uniaxial concrete strength (MPa),
Ls =h = M=Vh the shear span ratio at the member end, !wx = (Asx =bw sh )fyw =fc the mechanical
ratio of transverse steel parallel to the direction (x) of loading,  the con@nement eEectiveness
factor, equal to s n , with s = (1 sh =2b0 )2 and n = 1 (b0 =nh h0 + h0 =nb b0 )=3 for hoops with
nb legs or cross-ties parallel to side b0 of the con@ned core and nh legs or cross-ties parallel
to side b0 and d (%) the steel ratio of any reinforcement placed in each diagonal direction
of the member.
As it represents an average @t to the data, the ultimate chord rotation given by Equation (3)
is considered as an expected value and denoted by um . Owing to the large scatter, in the
veri@cation of chord rotations 5 per cent-fractile of the deformation capacity is used instead
of um . This lower characteristic value is:
uk; 0:05 = 0:4um

(4)

The proposed veri@cation of members at the life-safety performance level is


Ek; 0:95 6uk; 0:05
Copyright ? 2001 John Wiley & Sons, Ltd.

(5)

Earthquake Engng Struct. Dyn. 2001; 30:14391462

1444

T. B. PANAGIOTAKOS AND M. N. FARDIS

The value of  in Equation (5) can be related to the probability of member failure conditioned
on occurrence of the life-safety earthquake. For the present quanti@cation of the scatter in the
supply and demand values of chord rotation, u and E , this failure probability is approximately
equal to 1 per cent for  = 1:4 or 2 per cent for  = 1.
Exhaustion of deformation capacity at the level of the individual member does not necessarily imply failure of the system. For this reason veri@cation according to Equation (5) is
considered appropriate for the life-safety performance level, instead of the collapse prevention level.
Equation (5) needs to be veri@ed at each member end, separately for positive or negative
moments. Transverse gravity loads between the two ends of a beam aEect the veri@cation by
giving a (small) contribution to Em and by aEecting the value of the shear span, Ls = M=V ,
in Equation (3). Ls should be computed from the values of M and V at the member end due
to the superposition of gravity loads and seismic loading according to linear analysis.
It is clear from Equation (3) that if the member design from step 1 does not satisfy Equation
(5), in step 3 the designer may do one or more of the following: (a) For any type of element,
increase the con@ning reinforcement in the compression zone to increase !wx . (b) For elements
with unsymmetric cross-section and reinforcement, like beams and vertical elements with T-,
L- or channel-section, increase the amount of longitudinal reinforcement on the side which is
in compression when Equation (5) is violated, without increasing its tension reinforcement.
(c) For walls, reduce the amount of web vertical reinforcement between the edges of the
cross-section, to reduce the total tension reinforcement ratio !1 . (d) For short elements, like
coupling beams or short columns, add diagonal reinforcement at a steel ratio d (%). If these
measures are not enough to ful@l Equation (5), the designer may have to modify the crosssectional dimensions. This will change the shear span ratio, Ls =h, and (for vertical elements)
the axial load ratio,  = N=Ac fc . Note that such changes may require repeating the entire design
procedure, starting again from step 1.

3. APPLICATION TO A 4-STOREY DUAL TEST STRUCTURE


3.1. Description and design of the test structure
The proposed DBD procedure is applied to one side of the dual 4-storey structure shown in
Figure 1. Each side consists of a 3-bay frame with two shear walls, one with a 1:0m 0:5m
0:25 m L-shaped section and the other with a 1:0 m 0:25 m rectangular one, coupled through
a 0:45 m-deep and 1:0 m-long beam. The two columns of the frame have a depth of 0:4 m
and a width of 0:25 m. The test structure has two such frames at an axial distance of 4:0 m
and has been constructed for pseudodynamic (PsD) testing parallel to the frames at the ELSA
reaction wall facility of the European Commission in Ispra (I). These frames were designed
and tested (through additional masses) as if each one had a tributary slab width of 5:0 m,
instead of 2:0 m in the test structure. Nominal (5 per cent-fractile) strengths of 25 MPa for
concrete and 500 MPa for steel were speci@ed.
The two sides of the test structure were designed for unidirectional earthquake in the
direction of testing. One side was designed according to Eurocode 8 (EC8) [18] for a design
peak ground acceleration (PGA) of 0:4 g, soil class B (medium dense sands or medium stiE
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

DBD PROCEDURE FOR RC BUILDINGS

1445

Figure 1. Four-storey dual test structure (member cross-sectional dimensions in cm).

clays, see Figure 4 for spectral shape) and ductility class high (DCH). For this ductility class,
the force reduction factor R for dual structures dominated by coupled walls is taken in EC8
to be equal to 5.0. To achieve this value of R strict detailing rules and capacity design of
columns in bending at joints and of all elements in shear are prescribed in Reference [18].
As mentioned above, the design of the other side is according to the proposed procedure.
As the testing is within a framework of research on the structures designed according to
the Eurocodes, this design follows minimum reinforcement and detailing rules of Eurocode 2
(EC2) [24] for non-earthquake-resistant structures. Nevertheless, the procedure may be applied
along with any set of modern detailing rules for non-earthquake-resistant structures.
Tables IIII compare the detailing rules applied to the two sides (a dash in the last column
means that there is no provision for the corresponding rule).
To allow comparison of the two designs in step 3 the design level earthquake for which
Equation (5) is veri@ed is taken to be the same as in the EC8-design. Moreover, the serviceability earthquake in step 1 is taken to be equal to the design level earthquake divided
by R = 5. Then the ULS proportioning of members is performed for the same seismic load
eEects in both designs, corresponding to a base shear coeNcient of 0.2 from the spectrum, or
of 0.188 from the multimodal analysis. Moreover, the capacity design of columns in bending and of beams, columns and walls in shear is performed in both cases with the capacity
overstrength factors speci@ed in Reference [18] for ductility class H (1.25 for beams, 1.35
for columns, 1.25 for walls).
In both the EC8 design and step 1 of the proposed procedure the analysis uses uncracked
gross section stiEnesses and considers the length of beams within joints as rigid. With these
stiEness assumptions the fundamental period is equal to 0:51 s and the interstorey drift ratios
at the design level earthquake are fairly uniform: 0.51, 0.53, 0.5 and 0.47 per cent, from the
ground storey to the top. Despite the small depth of the coupling beam, the moment of the
couple of axial forces in the two walls according to the elastic analysis is about 1.1 times
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

Copyright ? 2001 John Wiley & Sons, Ltd.

if 2M=13Vc

0:8 d, 0:3m
6 mm
5dbL ; h=4; 24dbw , 0.15m

Transverse reinforcement (w)


sw; max outside critical regions
db; w critical regions
sw; max critical regions

Coupling beams
Diagonal reinforcement

8:5(1 + 1:2)fct =(1 + 0:5 =max )fy


8:5(1 + 1:2)fct =fy

7(1 + 1:2)fct =(1 +  =max )fy


7(1 + 1:2)fct =fy

1:25RMRd; ends =1 + Vsimply supp:beam


0

0:25As; bottom-span

Shear design
VEd , EQ
Vc critical regions, EQ

0.15
3.0

0:15; 50fct =fy


0:15 + 27fc =fy  =
308
0:25As; top-supports
0:25As; bottom-span
0:5As; top-supports

1:25RMRd; ends =1 + Vsimply supp:beam


full Vc

0:8 d, 0:3m

0:8 d, 0.3m

1:5h

DBD following EC2


& capacity design

2h

EC8-DCH

Critical end regions, length


Longitudinal bars (L)
min (%) tension Cange
max (%) critical regions
As; min top & bottom (mm2 )
As; min top-span
As; min top-supports
As; min cr.regions-bottom
db =hc -bar crossing joint for
column depth hc and  = N=Ac fc
(i) Interior joint:
(ii) Exterior joint:

Beam provisions

Table I. Minimum requirements of EC8 (DCH) and EC2=DBD; beams.

1446
T. B. PANAGIOTAKOS AND M. N. FARDIS

Earthquake Engng Struct. Dyn. 2001; 30:14391462

Copyright ? 2001 John Wiley & Sons, Ltd.

1:35RMRc =lc
full Vc

1:35RMRb 6 RMRc

6 mm; 0:25dbL
12dbL ; min(hc ; bc ); 0:3m
7:2dbL ; 0:6 min(hc ; bc ); 0:18m
0:4dbL
5dbL ; 0:25bo ; 0:1m
0.5
1:7875(fy =Es )(0:3 + 0:7Ag =Ao ) 0:0175
0.37

1.0
4.0
12 mm
3
150 mm

1:5 max(hc ; bc ); 0:6m; lc =5


(increased by 50% in lower two storeys)

EC8-DCH

1:35RMRc =lc
full Vc

1:35RMRb 6 RMRc

6 mm; 0:25dbL
12dbL ; min(hc ; bc ); 0:3m
7:2dbL ; 0:6 min(hc ; bc ); 0:18m

1.0
4.0
12 mm
3

DBD following EC2


& capacity design

The con@nement eEectiveness factor  equals s n , with s equal to (1 s=2bo )2 and n equal to 1 (bo =nh ho + ho =nb bo )=3 for hoops with nb legs or
cross-ties parallel to side bo of the core and nh legs or cross-ties parallel to side bo .

Shear design:
VEd seismic (EQ)
Vc for EQ

Capacity design at joints:

Transverse reinforcement (w)


dbw; min outside critical region
sw; max outside critical region
sw; max in splices
dbw; min critical regions
sw; max critical regions
(Asw =sw bc )(fy =fc ) critical regions
(Asw =sw bc )(fy =fc ) crit. regions
 = N=Ac fc 6

Longitudinal bars (L)


min (%)
max (%)
db; min
Minimum bars per side:
Dist. of bars restrained by hoops
or cross-ties 6

Critical end regions, height

Column provisions

Table II. Minimum requirements of EC8(DCH) and EC2=DBD; columns.

DBD PROCEDURE FOR RC BUILDINGS

1447

Earthquake Engng Struct. Dyn. 2001; 30:14391462

Copyright ? 2001 John Wiley & Sons, Ltd.

0:2; 0:5v
8 mm

0:3m

0.4
4.0

0.5
4.0

0:2; 0:5v
8 mm
bw =8
20dbh ; 0:2m

6 mm; 0:4dbL
9dbL ; 0:5bo ; 0:2m

6 mm; 0:4dbL
5dbL ; 0:25bo ; 0:1m
0.05
0:1375
(fy =Es )(0:3 + 0:7Ac =Ao ) 0:0175

0.2
4.0
8 mm

0:2dbv ; 0:2m

3
0.5
4.0
12 mm

3
1.0
4.0
12 mm

0.2, but such that v of entire wall section 0:4


0.4
8 mm
bw =8
0:2dbv ; 0:2m

bw

DBC (EC2& capacity design)

0:15lw ; 1:5bw

EC8-DCH

For the eEectiveness factor , see footnote of Table II. The required value of the wall curvature ductility factor, 
, is R2 for non-coupled walls and 0:8R2
for coupled; the eEective axial load ratio  of the boundary element can be taken equal to 0:75(NSd =2 + MSd =z)=Ac fc , where NSd ; MSd and z = 0:8lw are
the axial force, bending moment and internal lever arm of the entire wall cross-section.

Web:
Vertical bars (v)
v; min (%)
v; max (%)
dbv; min
dbv; max
sv; max
Horizontal bars (h)
h; min (%)
dbh; min
dbh; max
sh; max

Boundary elements:
1. In critical region:
Length lc from edge
Vertical reinforcement
Minimum bars per side
min over Ac = lc bw (%)
max over Ac = lc bw (%)
db; min
Con@ning hoops (w)
dbw; min
sw; max
(Asw =sw bc )(fy =fc )
(Asw =sw bc )(fy =fc )
2. Rest of wall
Vertical reinforcement
min (%)
max (%)

Wall provisions

Table III. Minimum requirements of EC8(DCH) and EC2=DBD; walls.

1448
T. B. PANAGIOTAKOS AND M. N. FARDIS

Earthquake Engng Struct. Dyn. 2001; 30:14391462

DBD PROCEDURE FOR RC BUILDINGS

1449

the sum of bending moments at the base of the individual walls, allowing for these walls to
qualify as coupled. According to the elastic analysis the walls take more than 70 per cent of
the seismic base shear.
In both designs the load combinations are according to EC2 [24] and EC8 [18], i.e. (a)
1:35 dead load+1:5 live load for factored gravity and (b) earthquake+dead load+0:3 live
load for the seismic combination.
The designs according to EC8 or step 1 of the DBD procedure follow the sequence outlined
in Section 2.1, as described in detail below:
(a) Beams were proportioned in Cexure for the envelope of the moments from the load
combinations: (i) of factored gravity loads; and (ii) of earthquake plus arbitrary-point-in-time
gravity loads (dead-load+0:3 live-load according to EC8). As the serviceability earthquake
of the DBD frame and the reduced by R = 5 design earthquake of the EC8 frame are the same,
longitudinal steel requirements for the beams of both frames are the same.
The upper limit to the diameter of beam bars passing through beam-column joints, as
determined by bar development within the joint (lines 11 and 12 in Table III) is 12 mm in
the EC8 frame and 14 mm in the DBD one. Bar development within the beam-wall joints sets
practically no limit to the beam bar diameter.
The minimum reinforcement ratio is 0:5fct =fy = 0:26 per cent for the EC8 beams, but 0.15
per cent for the DBD ones. So the minimum reinforcement is three 12 mm bars for the EC8
beams and two 12 mm beams for the DBD ones.
EC8 allows counting in the top beam reinforcement any slab bars which are parallel to the
beam and up to a distance bf from it equal to twice the slab depth hf on each side of the
beam, for exterior columns with transverse beams or interior ones without such beams, or
up to bf = 4hf on each side of the beam, for interior columns with transverse beams. In the
prototype structure, with the slab extending to both sides of the beams, this rule gives a slab
contribution to the top beam reinforcement of 560 or 200 mm2 at the face of the interior or
the exterior column, respectively, or of 170 mm2 at the face of the walls.
Over the interior column supports the minimum steel suNces as top reinforcement in both
designs, without any contribution from the slab reinforcement. Over beam supports at the
exterior column and at the interior wall the minimum reinforcement plus the above quantities
of slab steel barely meet the requirements for beam top reinforcement in the three lower
Coors. As a matter of fact, the reinforcement there meets the requirements with some de@cit
in the DBD frame, which has less minimum beam reinforcement than the EC8 one. The
minimum reinforcement suNces throughout the bottom of all beams, except those of the two
lower storeys of the DBD frame, where it needs to be increased to two 14 mm bars.
The minimum longitudinal reinforcement of two 14 mm bars in the EC8 design or two
12 mm bars in the DBD one suNces in the coupling beams. The maximum shear in the EC8
coupling beams (from the analysis or from capacity design, see (b) below, including 170 mm2
of slab steel in the calculation of the beam Cexural capacity), barely reaches the EC8 limit
for placement of bars along the two diagonals of the beam.
(b) Beam stirrups were proportioned for capacity design shears calculated on the basis of
overstrength Cexural capacities 1:25MRb at beam ends, plus transverse arbitrary-point-in-time
gravity loads. Beam Cexural capacities were determined including some slab reinforcement
according to EC8, i.e. up to a distance bf = 2hf on each side of the beam, except at the face
of interior columns, where bf = 4hf on each side of the beam can be taken.
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

1450

T. B. PANAGIOTAKOS AND M. N. FARDIS

Over 1:5hb = 0:675 m-long critical end regions of the DBD beams, where the concrete
contribution to shear resistance Vc is neglected, 8 mm stirrups at 225 centres are required in
the two upper Coors or at 210 mm centres in the two lower Coors. This stirrup spacing is
required on both sides of the joint with the interior column, where the slab contribution to
the beam Cexural capacity is the largest and penalizes the capacity design shears the most.
This spacing could be increased a little at the other end of these DBD beams, but it was kept
the same at both ends for simplicity.
Over the 2hb = 0:9 m-long critical end regions of EC8 beams the antibuckling requirement
sw; max = 5dbL = 60 mm controls the stirrup spacing.
In-between the critical end regions the maximum stirrup spacing of 0:3 m controls in both
designs.
Stirrup spacing in the coupling beams of the DBD frame is controlled by the capacity design
shear force. It is equal to 120 mm in the two upper Coors and 105 mm in the other two. In
the coupling beams of the EC8 frame the stirrup spacing required to resist the capacity design
shear is 95 mm, but the antibuckling requirement sw; max = 5dbL = 50 mm controls. Since the
EC8 coupling beams were at the limit of requiring diagonal reinforcement, the stirrup spacing
was reduced to 50 mm in the coupling beam of the (most critical) 1st Coor.
The (@nal) beam reinforcement shown in Figure 2 diEers from what is given above only
in the coupling beams of the DBD frame, because there step 3 of the DBD procedure results
in more closely spaced stirrups and in placement of some diagonal bars.
(c) Columns were proportioned in Cexure, so that the sum of their factored Cexural capacities at beam column joints, RMRc , exceeds the sum of overstrength capacities of the beams,
1:35RMRb . The values of 1:35RMRb listed in Table IV are calculated on the basis of three
alternative considerations for the slab width bf contributing with its tension steel to the hogging Cexural capacity of the beam: (1) for bf according to the EC8 values quoted above and
considered on both sides of the beam (i.e. as in the prototype structure); (2) for bf according
to the more realistic estimate of the New Zealand code, i.e. equal to one-quarter of the beam
span Lb , but only on one side of the beam, i.e. as in the test structure; and (3) for bf equal
to the physical upper bound of one-half of the clear distance Ls of the two frames in the test
structure. Beam Cexural capacities resulting from consideration (1) are practically the same
as those from (2) at the interior column joint, but are 2025 per cent lower at the exterior
column joint. Beam Cexural capacities from the upper bound consideration (3) are 4050 per
cent higher than the ones from (2).
To meet the limitations on minimum number and maximum distance of bars around the
column section, a minimum of 10 or 8 bars is required in the EC8 and the DBD columns,
respectively. With these numbers 12 or 14 mm bars would be enough in the EC8 and the
DBD columns, respectively, to supply the minimum reinforcement ratio of 1 per cent. (The
minimum reinforcement ratio for EC2 columns is 0.3 per cent, but the 1 per cent value of
EC8 is considered here necessary for earthquake-resistant columns and adopted in the DBD
approach.) In the interior columns of the EC8 and the DBD frames ten 14 mm bars and eight
16 mm ones, respectively, are needed to meet or exceed the beam overstrength capacities
from considerations (1) and (2), for the minimum column axial force in the seismic load
combinations. Vertical reinforcement in the exterior column of the EC8 frame is the same as
in the interior one, because ten 12mm bars are barely enough in the weak direction of bending
against factored gravity loads in the test structure. At the interior column of the DBD frame
the bar diameter can be reduced to 14 mm at the 4th Coor, where capacity design of columns
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

1451

Figure 2. Beam reinforcement of the DBD and EC8 sides of the test structure.

DBD PROCEDURE FOR RC BUILDINGS

Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

Copyright ? 2001 John Wiley & Sons, Ltd.

99
236
284
301

RMRc (kNm)

117
244
262
277

RMRc (kNm)

272
272
272
272

bf = Lb =4(2)

228
228
246
246

bf = 4h(1)
f
221
221
240
240

bf = Lb =4(2)

1:35RMRb (kNm)

Interior column

278
278
278
278

bf = 4h(1)
f

1:35RMRb (kNm)

Interior column

302
302
319
319

bf = Ls =2(3)

DBD

376
376
376
376

bf = Ls =2(3)

91
186
193
199

RMRc (kNm)

109
222
228
233

RMRc (kNm)

150
150
150
150

bf = Lb =4(2)

95
95
95
95

bf = 2h(1)
f

125
125
125
125

bf = Lb =4(2)

1:35RMRb (kNm)

Exterior column

120
120
120
120

bf = 2h(1)
f

1.35 RMRb (kNm)

Exterior column

190
190
190
190

bf = Ls =2(3)

216
216
216
216

bf = Ls =2(3)

Beam capacities calculated for an eEective slab width equal to the EC8 value: bf = 4hf on each side of the beam at central columns; bf = 2hf at exterior
ones.
(2) Beam capacities calculated for an eEective slab width equal to: b = L =4 on one side of the beam (L = clear span of the beam).
f
b
b
(3) Beam capacities calculated for an eEective slab width equal to the upper bound value: b = L =2 on one side of the beam (L = span of slab between
s
s
f
beams).

(1)

4th
3rd
2nd
1st

Floor

4th
3rd
2nd
1st

Floor

EC8

Table IV. Comparison of the sum of column (factored) Cexural capacities RMRc , to the sum of beam (overstrength) capacities 1:35RMRb
at beam-column joints.

1452
T. B. PANAGIOTAKOS AND M. N. FARDIS

Earthquake Engng Struct. Dyn. 2001; 30:14391462

DBD PROCEDURE FOR RC BUILDINGS

1453

in Cexure is not required. The same reinforcement, of eight 14 mm bars, is enough for the
exterior column of the DBD frame. Exterior columns of both frames meet the capacity design
requirements at the joints, even in the upper bound case (3) of beam Cexural capacities.
(d) The transverse reinforcement of columns is proportioned for capacity design shear
forces computed on the basis of column Cexural overstrengths, 1:35MRc , at the maximum axial
compression in the seismic load combinations. Capacity design shears range from 105 kN in
the top storey to 155 kN in the 2nd storey of the DBD columns (the range is narrower in the
EC8 frame). These shears can be easily resisted by the minimum transverse reinforcement,
as the contribution of concrete to shear, Vc , ranges from 55 to 105 kN. In both frames hoop
spacing is controlled by the maximum spacing sw; max = 7dbL outside critical end regions in the
EC8 design, or sw; max = 7:2dbL in such regions or in splicing zones in the DBD one. Within
the critical end regions of EC8 columns hoops and cross-ties are controlled by the maximum
spacing of one-quarter of the minimum dimension of the con@ned concrete core, bo =4 = 45mm.
The column vertical and transverse reinforcement given above is the same as the @nal one
in Figure 3.
(e) Wall vertical reinforcement should resist design moments from a linear envelope of the
wall bending moments from the elastic analysis, shifted upwards by the wall length dimension
lw . Design moments are equal to 135 and 195kNm at the base of the L-shaped exterior wall or
of the interior rectangular wall, respectively (despite its smaller cross-section the interior wall
attracts larger seismic forces due to its connectivity with beams). The range of axial forces
for the earthquake plus arbitrary-point-in-time gravity load is 460 to 20 kN (tension) in the
L-shaped exterior walls and 505 to 270kN in the interior ones. The couple of wall axial forces
gives a moment of 370kNm at the base, larger than the sum of wall moments there, 330kNm,
implying signi@cant coupling action of the connecting beam despite its limited depth.
Throughout the height of the walls con@ned boundary zones are provided at the ends of the
section, according to the rules of Table III. A 500 mm 250 mm boundary zone is provided
over the full Cange of the L-shaped wall section in both designs. Boundary zones at the ends
of the rectangular webs of the walls have a length of lc = 1:5bw = 380 mm in the EC8 design
or of lc = bw = 250 mm in the DBD one. The vertical reinforcement of these boundary zones
was selected to meet the minimum requirements in Table III regarding maximum distance and
minimum number of bars and the minimum reinforcement ratio over the boundary zone itself
(1 or 0.5 per cent) and the entire section (0.4 per cent). As an exception, the intermediate
12 mm bar along the 250 mm side of the DBD wall section (normally needed as intermediate
vertical reinforcement in the beamwall joint in the weak direction of the wall) was removed;
the @ve 12mm bars of each 250mm 250mm boundary zone in the DBD walls were replaced
by four 14 mm corner bars, providing about the same section Cexural capacity. As another
exception, each boundary zone of the EC8 interior rectangular wall was provided with three
additional 12 mm bars, to make up (through dowel action) for a de@cit in resistance against
sliding shear according to EC8. This arrangement was preferred over the placement of two
additional bars through the wall base at each storey, at 45 .
Hoops are provided around corner bars of all boundary zones and cross-ties connect intermediate bars on opposite sides of boundary zones of the EC8 walls. In the critical 1st
storey of the EC8 walls the spacing of these hoops and cross-ties has to satisfy the minimum
requirements of Table III for con@nement of the core of the boundary zone and for prevention
of buckling of vertical bars. Anywhere else on the walls of both sides the spacing of hoops
and cross-ties is equal to the minimum dimension of the boundary zone, i.e. to bw = 250 mm.
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

T. B. PANAGIOTAKOS AND M. N. FARDIS

Figure 3. Reinforcement of vertical elements of DBD and EC8 sides of the test structure.

1454

Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

DBD PROCEDURE FOR RC BUILDINGS

1455

Figure 4. Five per cent-damped acceleration spectra and time-history of synthetic


input motion after Kobe=Port Island.

The vertical and hoop reinforcement determined as described above is shown in Figure 3
throughout the height of the EC8 walls and at the top storey of the DBD ones (in the other
three storeys the reinforcement of at least one DBD wall was modi@ed in step 3 of the
proposed procedure).
The vertical reinforcement determined from minimum measures provides signi@cant Cexural
overstrength. For the maximum compression=minimum compression (maximum tension) at the
wall base the factored Cexural capacity MRw is equal to 515=430 kNm or 605=585 kNm in the
L-shaped exterior or the rectangular interior walls of the EC8 design, giving corresponding
overstrength ratios with respect to the elastic seismic moments of 3:85=3:2 or 3:1=3:0, respectively. On the DBD side the corresponding values of MRw are 395=290kNm and 415=390kNm
for the L-shaped and the rectangular walls, respectively, while those of the overstrength ratio
are 2:95=2:15 and 2:1=2:0.
(f) Walls were designed in shear for capacity design shears determined by magnifying
the shears from the analysis by a factor which is derived according to EC8, neglecting the
coupling of the two walls, from 1.25 times the moment overstrength ratios quoted at the end
of the previous paragraph, plus a semi-empirical factor accounting for inelastic higher-mode
eEects. The @nal magni@cation factor applied to wall shear forces from the analysis is equal to
5.0 in both L-shaped exterior walls and to 4.3 or 3.2 for the rectangular interior walls of the
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

1456

T. B. PANAGIOTAKOS AND M. N. FARDIS

EC8 and the DBD side, respectively. After this magni@cation design shears in the 1st storey of
the L-shaped walls for the extremes of compression and tension are equal to 280 and 215 kN,
respectively, on the EC8 side or to 265 and 150 kN on the DBD side. In the rectangular
interior walls, which do not develop tensile axial forces under the seismic load combination,
design shear forces at the 1st storey are equal to 370 or 275 kN on the EC8 and the DBD
side, respectively. Design shears at the 2nd storey are 85 or 75 per cent respectively, of the
values above for the L-shaped exterior walls and the interior rectangular ones, respectively.
At all storeys of the EC8 and DBD walls the minimum horizontal reinforcement of 8 mm
bars at 160 mm centres (20 times the bar diameter), or at 200 mm centres (giving a web
reinforcement ratio of 0.2 per cent) was, respectively, provided. For a 45 -truss these reinforcements give factored shear capacities of 220 or 175 kN, respectively. These capacities
suNce against the design shear of the exterior L-shaped walls under tensile axial load, i.e.
when at the critical 1st storey the contribution of concrete to shear resistance, Vc , is neglected.
For compressive axial forces the factored shear capacity of concrete, Vc , is equal to 85 or
80 kN in the 1st storey of the EC8 or the DBD wall, respectively; then the total factored
shear capacity exceeds the design shear force only in the L-shaped exterior walls of the EC8
side. Everywhere else in the 1st storey under peak axial compression there is a shear strength
de@cit ranging from 19 per cent in the rectangular wall of the EC8 side, to 7:5 or 4 per cent
in the rectangular and L-shaped walls of the DBD side. Nevertheless, as the EC8 rules for
the shear force magni@cation in walls are deemed to be overly conservative, these de@cits
may be considered acceptable.
The vertical reinforcement in the webs of the DBD walls was selected to be the same as
the horizontal (i.e. equal to the minimum). The two 8 mm bars placed on each side of the
web of the EC8 walls between the two boundary zones provide 50 per cent more vertical
web reinforcement than the minimum. For the value of the shear span ratio M=Vlw prevailing
in these walls (around 1.6) wall axial compression is considered to assist the web vertical
reinforcement in shear; hence, there is a surplus of web vertical steel in all four walls,
especially in those of the EC8 side.
It is clear that both the EC8 design and that from step 1 of the DBD procedure are
not controlled by ULS proportioning for the serviceability earthquake. They are controlled
instead by capacity design, minimum reinforcement and detailing rules and sometimes by
factored gravity loads.
The changes in the reinforcement of the DBD side eEected within step 3 of the proposed
approach (with  = 1:2 in Equation (5)) to meet the chord rotation demands from step 2 are
the following:
(i) The spacing of stirrups in the coupling beams was reduced from 120 to 80 mm in the
two upper Coors and from 105 to 50 mm in the two lower ones, while two 20 mm bars
were placed along each diagonal of the coupling beam of the 1st Coor.
(ii) In the 1st storey of the interior rectangular wall and in the three lower storeys of the
exterior L-shaped ones, hoop spacing in the boundary zones was reduced from 250 to
60 mm and the diameter of the four corner bars in all boundary zones other than that
of the Cange of the L-shaped wall was increased from 14 to 16 mm.
The increase in the vertical reinforcement of the DBD walls increases their Cexural capacity
and in turn the overstrength ratio over the elastic moments at the base from the analysis for
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

1457

DBD PROCEDURE FOR RC BUILDINGS

Table V. Comparison of steel quantities per frame (kg) in the two designs.
Floor

Longitudinal

Transverse

EC8

DBD

DiEerence
(%)

4
3
2
1
Total

66
66
66
66
264

41
41
48
48
178

38
38
27
27
33

4
3
2
1
Total

74
87
87
128
376

59
80
80
118
337

20
8
8
8
10

4
3
2
1
Total

110
130
160
191
561

70
88
88
146
392

36
32
32
24
30

EC8

DBD

DiEerence
(%)

Beams and coupling beams


45
22
51
45
22
51
45
26
42
46
26
44
181
96
47
Columns
84
51
39
84
49
42
100
49
51
117
57
51
385
206
45
Walls
74
54
27
74
56
24
74
56
24
167
66
61
386
232
40

Total
EC8

DBD

DiEerence
(%)

111
111
111
112
445

63
63
74
74
274

43
43
33
34
38

158
171
187
245
761

110
129
129
175
543

30
25
31
29
29

184
204
204
358
950

124
144
144
212
624

33
29
29
41
34

maximum=minimum axial compression from 2:95=2:15 and 2:1=2:0 in the L-shaped exterior
wall and the interior rectangular one, respectively, to 4:2=2:15 and 3:0=2:85. The new Cexural
overstrength ratios are similar to those of the EC8 walls. The increase in Cexural overstrength
at the base increases also capacity design shears in the DBD walls. In the rectangular interior
wall the design shear increases to 350 kN at the 1st storey and to 265 kN at the 2nd. The
available shear strength of the 2nd storey is marginal, while that of the 1st storey is de@cient
by about 35 per cent. Correction of this de@ciency requires reduction of the spacing of the
horizontal reinforcement from 200 to 110 mm. Nevertheless, as the capacity design shears of
EC8 are considered overconservative, it was decided to keep the shear reinforcement in the
DBD walls unchanged.
Horizontal reinforcement in the joints, consisting of perimeter hoops alone, was determined
at this stage according to EC8 rules on the basis of the cross-sectional area of beam bars
crossing the joints. As the beam longitudinal reinforcement was not revised in step 3, proportioning of the joint reinforcement could have been accomplished at the end of step 1.
Indeed, as the longitudinal reinforcement in the EC8 and the DBD beams is similar, joint
reinforcement in both designs is the same.
The @nal reinforcement of the DBD design is compared in Figures 2 and 3 to that of
the EC8 design. Table V compares the weight of steel required for one frame according to
the two designs, per type of element and Coor and separately for the longitudinal and the
transverse reinforcement. For all types of elements the DBD frame requires signi@cantly less
steel than the EC8 one. Overall the diEerence amounts to one-third of the EC8 requirements.
The diEerence is largest in the transverse reinforcement, where the DBD frame requires almost
half the reinforcement of the EC8 design.
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

1458

T. B. PANAGIOTAKOS AND M. N. FARDIS

3.2. Nonlinear dynamic analyses of the test structure


In the actual 4-storey test structure one of the two frames has been constructed according to
the EC8 design and the other according to the DBD one. In the PSD test the two frames are
subjected to the same horizontal displacement time-histories at Coor levels.
Non-linear dynamic response analyses were performed for the structure as tested, i.e. with
the one side according to DBD and the other according to EC8, using as input @ve synthetic
motions conforming to 1.5 times the elastic spectrum of the design earthquake. Each motion was applied with a positive or negative sign (direction). The @ve motions, drawn from
Reference [25], have phasing and intensity envelopes after some well-known historic records,
namely that of Kobe=Port Island, Hollister (1961), San Fernando=Alhambra and Fairmont
Reservoir (1971) and Imperial Valley=El Centro (1940). The synthetic motion after Kobe
was selected for the PSD test. Its time-history and the corresponding 5 per cent-damped elastic
spectra are shown in Figure 4, including the 1.5 scale factor over the design earthquake.
The nonlinear analyses employ simple lumped-inelasticity member models, of the onecomponent type with bilinear skeleton curve and modi@ed-Takeda hysteresis laws. The modelling used has been presented in more detail elsewhere [26; 27]. Material strengths were
assumed to be equal to their expected values: fcm = fc + 8 MPa = 33 MPa and fym = 1:15fy =
575 MPa.
Figure 5 shows the average ratio of member peak inelastic chord rotations from the nonlinear dynamic analyses to the corresponding supply from Equation (3). As this ratio can be
considered as a damage ratio, it is presented as a percentage, with a value of 100 per cent
meaning exhaustion of expected member deformation capacity. Averaging is over the 10 nonlinear dynamic analyses, with the @ve synthetic motions applied in the positive and negative
directions at intensity 1.5 times the design motion. The same @gure presents the average and
the extreme values of the interstorey and top drift ratios obtained in the 10 analyses. Despite
its signi@cantly lower amount of steel the DBD frame is predicted to experience similar member damage levels as the EC8 one and to fare on the average better than the EC8 frame in
the coupling beams.
Figure 6 presents analysis results for the motion of Figure 4 applied in the PSD test. Floor
displacement time-histories and peak interstorey drifts from pre-test calculations based on
assumed mean material strengths consistent with the speci@ed nominal values, are compared
with test results. The test results may have been aEected by the damage inCicted to the test
structure by previous testing at the design motion intensity.
In the analyses summarized in Figures 5 and 6 the DBD-side coupling beams of the 2nd
Coor were predicted to fail under 1.5 times the design earthquake. At @rst sight this is
surprising, as for  = 1:2 the combination of Equations (4) and (5) seems to provide a safety
factor against attainment of Ek;0:95 equal to 1:2=0:45 = 2:67, which should not be exhausted
by the application of the design motion at a scale factor of 1:5. To investigate the origin
of this apparent discrepancy, the member inelastic chord rotations, E;in , from the non-linear
analyses of the response of the two structures designed herein (i.e. of the EC8 and the
DBD building) to 10 input motions (5 motions applied with a plus or minus sign), are
compared to the elastic values, Em , computed in step 2 of the proposed procedure according
to Reference [20]. Mean standard deviation ranges and the 95 per cent-fractile of the ratio
E;in =Em are shown in Figure 7. The m ! range and the diEerence of the 95 per centfractile from the mean reCect not only the inCuence of the details of the input motion, but
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

DBD PROCEDURE FOR RC BUILDINGS

1459

Figure 5. Average damage ratio (per cent) at member ends, computed as ratio of peak
inelastic chord rotation to capacity from Equation (3), and mean and maximum=minimum
drifts from non-linear dynamic analyses to 10 spectrum-compatible synthetic motions
with intensity 1.5 times the life-safety design motion.

also dispersion between the DBD and the EC8 design and the diEerent members of the same
storey.
The results in Figure 7 show that the method developed in Reference [20] and applied
herein in step 2 of the DBD procedure estimates well the inelastic chord rotation demands
in the members of the test structure except in the coupling beams. In walls and columns the
mean and the 95 per cent-fractile value in Figure 7(c) and (d) are not far from the values
in Reference [20] of 0.85 at ground level and 1.0 at the top for the mean, or 1.05 at ground
level and 1.65 at the top for Ek;0:95 (the only diEerence is in Ek;0:95 at the top storeys, with
no eEect though on the design and detailing of the present case). For the beams, the values
in Reference [20] of 1.2 at ground level and 1:25 at the top for the mean, or of 1.65 and 1.7,
respectively, for the 95 per cent-fractile, are not very diEerent from those in Figure 7(a). With
the exception of the @rst Coor, they are signi@cantly lower, though, from those of Figure 7(b)
for the coupling beams. This means that in the top three Coors the values of uk;0:95 in coupling
beams are underestimated by the procedure in Reference [20] by approximately 15 per cent.
This explains only partly the high damage ratio values predicted in the DBD coupling beams
at the second Coor of the test structure.
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

1460

T. B. PANAGIOTAKOS AND M. N. FARDIS

Figure 6. Computed and measured Coor displacement time-histories and interstorey drifts
and computed damage ratio (per cent) at member ends for test structure subjected to
Kobe-like motion at intensity 1.5 times the life-safety motion, Figure 4.

4. CONCLUSIONS
The key elements of the proposed procedure for displacement-based seismic design of RC
frame-wall dual building structures are the following: (i) a preliminary proportioning of member reinforcement on the basis of the ULS for factored gravity loads, capacity design of all
elements in shear and detailing provisions for non-earthquake-resistant design; and (ii) @nal
Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

DBD PROCEDURE FOR RC BUILDINGS

1461

Figure 7. Mean standard deviation range and 95 per cent-fractile value of the ratio of inelastic chord
rotation demand to value from elastic multimodal analysis with 5 per cent damping, for the EC8 and
DBD structures subjected to 10 motions with intensity 1.5 times the life-safety design motion.

detailing of members and possible revision of member proportioning to meet the inelastic
chord rotation demands of the design (life safety) earthquake, as these are determined from
the 5 per cent-damped spectrum and a linear analysis with secant-to-yield member stiEnesses.
ULS design of members for a frequent or occasional serviceability earthquake and capacity
design of columns in Cexure at joints may be added to element (i) of the procedure. The
main diEerence between a structure designed along these lines and one which follows current
design codes such as EC8 is that the reinforcement which is additional to that proportioned
for non-seismic loads is concentrated only where it is needed to meet the seismic deformation demands in the DBD structure, rather than being placed indiscriminately in all members
according to prescriptive detailing rules. This diEerence may allow overall savings in steel
which amount to about 33 per cent for the 4-storey dual test structure considered in this paper.
Despite these savings, according to pre-test calculations the performance of the DBD structure
under a ground motion representative of the maximum considered earthquake (namely up
to 1.5-times the 475 year motion) is not inferior to that of an EC8-design.
ACKNOWLEDGEMENTS

This research is partly supported by the European Commission, within its Training and Mobility for
Researchers (TMR) network ICONS (Innovative seismic design concepts for new and existing structures), Contract No. FMRX-CT96-0022 (DG 12 - RSRF). The construction and PSD testing of the
4-storey dual structure at the ELSA Laboratory of the JRC in Ispra (I) is supported by the European
Commission under a TMR-Large-Scale Facilities contract to the JRC.
REFERENCES
1. Moehle JP. Displacement-based design of RC structures subjected to earthquakes. Earthquake Spectra 1992;
8(3):403 428.
2. Moehle JP. Displacement-based seismic design criteria. Proceedings of the 11th World Conference Earthquke
Engineering, Acapulco, Mexico, 1996.

Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

1462

T. B. PANAGIOTAKOS AND M. N. FARDIS

3. New Zealand National Society for Earthquake Engineering. The assessment and improvement of the structural
performance of earthquake risk buildings. Draft for General Release, 1996.
4. ATC. NEHRP Guidelines for the seismic rehabilitation of buildings. FEMA Report No. 273. Applied Technology
Council, for the Building Seismic Safety Council and the Federal Emergency Management Agency, Washington,
DC, 1997.
5. Priestley MJN. Displacement-based seismic assessment of reinforced concrete buildings. Journal of Earthquake
Engineering 1997; 1(1):157192.
6. Fardis MN. Seismic assessment and retro@t of RC structures. Invited State-of-the-Art Lecture. 11th European
Earthquake Engineering Conference, Paris, 1998.
7. Priestley MJN. Myths and fallacies in earthquake engineeringconCicts between design and reality. Proceedings
of the T. Paulay Symposium: Recent Developments in Lateral Force Transfer in Buildings, La Jolla, CA,
1993.
8. Kowalsky, MJ, Priestley MJN, MacRae GA. Displacement-based design of RC bridge columns in seismic
regions. Earthquake Engineering and Structural Dynamics 1995; 24(12):16231643.
9. Calvi GM, Kingsley GR. Displacement-based seismic design of multi-degree-of-freedom bridge structures.
Earthquake Engineering and Structural Dynamics 1995; 24(9):12471266.
10. Calvi GM, Pavese A. Displacement based design of building structures. In Proceedings of the 5th SECED
Conference on European Seismic Design PracticeResearch and Application, Elnashai AS (ed.). Balkema:
Rotterdam, 1995; 127132.
11. Priestley MJN. Calvi GM. Concepts and procedures for direct displacement-based design and assessment. In
Seismic Design Methodologies for the Next Generation of Codes. Fajfar P, Krawinkler H (eds). Balkema:
Rotterdam, 1997; 171182.
12. Priestley MJN. Displacement-based approaches to rational limit states design of new structures. Closing Lecture,
11th European Earthquake Engineering Conference, Paris, 1998.
13. Wallace JW. Seismic design of RC structural walls. Part I: new code format. Journal of Structural Engineering
ASCE 1995a; 121(1):7587.
14. Wallace JW. Seismic design of RC structural walls. Part II: applications. Journal of Structural Engineering
ASCE 1995b; 121(1):88100.
15. SEAOC. Recommended lateral force requirements and commentary. Structural Engineers Association of
California, Sacramento, 1999.
16. Fardis MN, Panagiotakos TB. Displacement-based design of RC buildings: proposed approach and application.
In Seismic Design Methodologies for the Next Generation of Codes. Fajfar P, Krawinkler H (eds). Balkema:
Rotterdam; 1997:195206.
17. Panagiotakos TB, Fardis MN. Deformation-controlled earthquake resistant design of RC buildings. Journal of
Earthquake Engineering 1999; 3(4):495518.
18. CEN. Design provisions for earthquake resistance of structures. Part 1: General rules. Part 11: Seismic actions
and general requirements for structures; Part 12: general rules for buildings; Part 13: speci@c rules for various
materials and elements. ENV1998-1-1, 12 and 13, Brussels, 1994.
19. SEAOC. Vision 2000, Performance Based Seismic Engineering of Buildings. Structural Engineers Association
of California, Sacramento, 1995.
20. Panagiotakos TB, Fardis MN. Estimation of inelastic deformation demands in multistorey RC frame buildings.
Earthquake Engineering and Structural Dynamics 1999; 28:501528.
21. Panagiotakos TB, Fardis MN. Deformations of RC members at yielding and ultimate. ACI Structural Journal,
2001; 98(2).
22. Park YJ, Ang AMS. Mechanistic seismic damage model of reinforced concrete. Journal of Structural
Engineering ASCE 1985; 111(4):722739.
23. Park YJ, Ang AH-S, Wen YJ. Damage-limiting aseismic design of buildings. Earthquake Spectra 1987; 3:1.
24. CEN. Design of concrete structures. Part 1: general rules and rules for buildings. ENV1992-1-1, Brussels, 1991.
25. Pecker A. Generaccprogramme de generation d accelerogrammes. Geodynamique et Structure Report, Paris,
February 1994.
26. Fardis MN, Panagiotakos TB. Seismic design and response of bare and in@lled reinforced concrete buildings.
Part I: bare structures. Journal of Earthquake Engineering 1997; 1(1):219256.
27. Fardis MN, Bousias SN, Franchioni G, Panagiotakos TB. Seismic response and design of RC structures with
plan-eccentric masonry in@lls. Earthquake Engineering and Structural Dynamics 1999; 28:173191.

Copyright ? 2001 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2001; 30:14391462

S-ar putea să vă placă și