Computers and Concrete, Vol. 20, No.

6 (2017) 719-729
DOI: https://doi.org/10.12989/cac.2017.20.6.719 719

Reliability assessment of RC shear wall-frame buildings

subjected to seismic loading

Ahmet Tukena, Mohamed A. Daheshb and Nadeem A. Siddiqui

Department of Civil Engineering, King Saud University, Riyadh 11421, Saudi Arabia

(Received May 22, 2016, Revised August 9, 2017, Accepted August 10, 2017)

Abstract. A considerable research is available on the seismic response of Reinforced Concrete (RC) shear wall-frame
buildings, but the studies on the reliability of such buildings, with the consideration of human error, are limited. In the present
study, a detailed procedure for reliability assessment of RC shear wall-frame building subjected to earthquake loading against
serviceability limit state is presented. Monte Carlo simulation was used for the reliability assessment. The procedure was
implemented on a 10-story RC building to demonstrate that the shear walls improve the reliability substantially. The annual and
life-time failure probabilities of the studied building were estimated by employing the information of the annual probability of
earthquake occurrence and the design life of the building. A simple risk-based cost assessment procedure that relates both the
structural life-time failure probability and the target reliability with the total cost of the building was then presented. The
structural failure probability (i.e., the probability of exceeding the allowable drift) considering human errors was also studied. It
was observed that human error in the estimation of total load and/or concrete strength changes the reliability sharply.
Keywords: serviceability limit state; RC shear wall-frame building; reliability; story drift; earthquake loading

1. Introduction procedure was then applied to a 3-D building with different

heights. The analytical lateral displacements matched
Numerous reinforced concrete (RC) shear wall-frame reasonably well with the SAP2000 results. Tuken and
buildings exist almost in all the big cities of the world Siddiqui (2011) proposed a simple-to-apply analytical
including the cities of Saudi Arabia. Many of these cities lie method based on “dual system” concept to determine the
in moderate to high seismic regions. The reliability amount of shear walls which can satisfy the strength,
assessment of such shear wall-frame buildings against stiffness and ductility requirements imposed by the seismic
earthquake, likely to occur in the life-time of the building, is codes on RC moment resisting frame buildings. The
necessary to assure a minimum safety level for the proposed methodology was then applied to a 10-storey RC
occupants’ lives and the property (Douglas et al. 2013, building containing shear walls. It was shown that the
Goulet et al. 2007, Haselton et al. 2011, Ulrich et al. 2014). amount of shear walls which is enough to satisfy the
In the past, researchers have studied the influence of shear strength requirements also fulfills the stiffness criteria (i.e.,
walls (or cantilever walls) on controlling the lateral story story drift limitation) required by the seismic codes.
drift of RC buildings subjected to earthquake forces (e.g., Burak and Comlekoglu (2013) evaluated the effect of
Paulay 1999, Priestley et al. 2007). Sezen et al. (2003) shear wall plan area to floor plan area ratio (shear wall
highlighted the structural deficiencies observed in the ratio) on the seismic response of RC buildings. They carried
damaged structures during 1999 Kocaeli, Turkey out nonlinear time-history analysis for 24 mid-rise building
earthquake. The deficiencies which were identified include models having shear wall ratios between 0.51 and 2.17
insufficient transverse reinforcement in RC columns, percent in both directions. In the analyses, seven different
strong-beam and weak-columns, soft and weak stories, poor earthquake records were used in the evaluation of the
quality construction, poor detailing in beam-column joint seismic performance of these buildings. They recommended
regions etc. They concluded that the buildings having shear that in order to control the lateral story drift, a minimum of
walls perform considerably well during the earthquake. 1.0% shear wall ratio has to be used in the design of mid-
Tuken (2004) proposed an analytical procedure to estimate rise buildings. They also observed that when the shear wall
the lateral displacement of a mixed (frame+shear wall) ratio is more than 1.5%, the effect of the shear wall on the
structure subject to earthquake forces. The analytical performance was insignificant. Lee and Haldar (2003a)
developed an efficient and accurate algorithm to study the
reliability of a steel frame and RC shear wall structural
Corresponding author, Professor system subject to static loading. The algorithm was then
E-mail: [email protected] extended to consider dynamic loading (including seismic
a
Assistant Professor loading) in another companion paper (2003b). The concept
E-mail: [email protected] integrates the first-order reliability method and the finite-
b
M.Sc. Student element method, resulting in a stochastic finite-element-
Copyright © 2017 Techno-Press, Ltd.
http://www.techno-press.org/?journal=cac&subpage=8 ISSN: 1598-8198 (Print), 1598-818X (Online)
720 Ahmet Tuken, Mohamed A. Dahesh and Nadeem A. Siddiqui

based approach. The reliability of a steel frame with and seismic response of RC shear wall frames to assess the soil-
without the presence of RC shear walls was evaluated for foundation-structure interaction effects. They concluded
both serviceability and strength performance functions. The that the fixed-base assumption overestimates the design of
results were justified by using Monte Carlo simulations. the wall element and underestimates the design of the
The algorithm confirms quantitatively that the shear walls connected moment frame. Soleimani-Abiat et al. (2015)
are needed, particularly when the steel frame is weak, in investigated the effects of seismic load combinations on the
satisfying the serviceability requirement of lateral response behavior of slabs at their connection zones with
deflection. Wyadtowski et al. (2015) outlined a procedure to the shear walls. The results of the study revealed that layout
estimate the reliability index for a laterally loaded of shear walls substantially affects the magnification of
diaphragm wall against serviceability limit state. The forces at the shear wall-floor slab connections. Taleb et al.
reliability indices were obtained by creating two response (2012) studied the influence of opening ratios on the cracks
surfaces. One based on the maximum lateral top distribution and shear strength of RC structural walls. They
displacement of the wall and the other one using the tested four RC single span structural walls having various
maximum values of the bending moment. They also opening sizes and locations under lateral reversed cyclic
obtained the global reliability index by using the concept of loading. The results of the study showed that the shear
system reliability. Jeong et al. (2012) carried out the strength changes, depending on the loading direction, due to
fragility analyses to evaluate the relative seismic safety opening locations. Dahish et al. (2015) studied the influence
margins of multi-story RC buildings under varying ductility of shear walls in controlling the lateral response of the RC
level, input motion intensity, and configuration. The studied frame building by varying the shear wall thicknesses,
buildings were designed based on a seismic code. They height, configuration and opening locations. The earthquake
observed that the damage state probabilities of wall-frame was considered from one direction only while studying the
structures designed to high peak ground acceleration and effect of the first two parameters (i.e., thickness and height)
ductility levels do not achieve the most favorable safety as the building and shear wall arrangements were
objectives. A relationship was also proposed to quantify the symmetric along the two orthogonal directions. However, in
damage state probabilities of mid-rise RC buildings. the case of third and fourth parameters (i.e., shear wall
Martins et al. (2016) derived a vulnerability model in terms configuration and opening location), the earthquake was
of the ratio of repair cost to replacement cost for a given considered from the two directions separately as the shear
intensity level of ground shaking. They used the model to wall configuration and opening location was not symmetric
estimate economic losses due to seismic action. The results in the two orthogonal directions. The results of the study
of this study highlight important issues in the derivation of were useful for obtaining the optimum amount and
vulnerability functions, which are a fundamental component arrangement of shear walls for a given RC frame building
for an adequate seismic risk assessment. against a specified seismic loading.
Monteiro (2016) studied the seismic reliability of RC The above literature review shows that even though a
bridge structures using Latin Hypercube sampling considerable research is available on the static and dynamic
algorithm. He demonstrated the robustness and analyses of shear walls and shear-wall frame buildings, but
effectiveness of the Latin Hypercube algorithm in the studies on reliability analysis of shear wall-frame
estimating the probability of failure for both regular and buildings are still limited. Such reliability analysis of shear
irregular bridge configurations. Monteiro et al. (2016a) wall-frame buildings is necessary because many times (i)
computed the failure probability of existing bridges using the pure RC frames under seismic loading do not satisfy the
the results of a nonlinear dynamic analysis. They target safety (or reliability) requirements and (ii) designer
statistically characterized the different variables generally may be interested in knowing how much safety level has
considered in the seismic assessment procedure such as been improved by using a certain quantity of shear walls.
geometry, material properties, records of the earthquake, the Furthermore, how to incorporate the human errors in the
level of intensity etc. Failure probability was obtained structural reliability of shear wall-frame structure was not
through the probabilistic analysis of a safety indicator. The seen in the approachable references. A simple risk-based
safety indicator was defined as the difference between cost assessment procedure is also needed to relate both the
capacity and demand. A case study of seven bridge life-time failure probability and the target reliability with
configurations, with different (ir)regularity levels was the total cost of the building. In the present study, a detailed
considered along with many sets of earthquake records. The procedure for carrying out the reliability analysis of RC
simulation process was carried out using the Latin shear wall-frame buildings subjected to earthquake forces
Hypercube sampling algorithm. Through this study, they against lateral drift is provided. The influence of human
identified the vulnerable configurations and shown the errors in the two governing parameters, viz. concrete
importance of the variable detail level. In the other strength and story weight, on the reliability of shear wall-
companion paper, Monteiro et al. (2016b) made the use of frame buildings is also presented. The concrete strength was
the simplified procedures and addressed the performance of considered as one of the governing parameters where
a nonlinear static procedure by direct comparison with human errors are very common because of the involvement
nonlinear dynamic results. They also investigated the use of of human beings during mixing, curing or testing of the
different static analysis versions corresponding to different concrete. Similarly, story weight may be affected due to the
types of pushover load distributions. A comparison of the errors committed during the selection or calculation of loads
static and dynamic approaches was then carried out on a (out of the available guidelines e.g., codes). Finally, a
parametric basis. simple risk-based cost assessment procedure that relates
Marzban et al. (2014) carried out a case study on the both the life-time failure probability and the desired
 Reliability assessment of RC shear wall-frame buildings subjected to seismic loading 721

reliability with the total cost of the building is proposed. A 2.1 Probability of limit state violation
few sensitivity analyses were also carried out to obtain the
results of practical interest. The probability of limit state violation can be obtained
by considering the probability models and statistics of all
the basic variables involved in the limit state function.
2. Formulation for reliability analysis There are several methods for estimating the probability of
failure such as First Order Reliability Method (FORM),
In order to carry out the reliability analysis of RC
Second Order Reliability Method (SORM), Monte Carlo
building, with and without shear walls, a limit state function Simulation Technique, Latin Hypercube Sampling Method,
which describes the failure criterion is needed. This LHSM (Monteiro 2016, Monteiro et al. 2016a, 2016b) etc.
function is negative or zero at the failure and it is positive In the present study, Monte Carlo Simulation technique was
when the structure is safe. Thus, the probability of failure preferred as this method, although computationally
(i.e., probability of limit state violation) can be defined as
expansive, is considered as the most accurate method
Pf  PG ( X)  0 (1) (Nowak and Collins 2012). As the authors had access to a
powerful workstation, they could perform a large number of
where G(X) is the limit state function and X is the vector of simulations, required in Monte Carlo Simulation, in a
random variables. Thus, for the serviceability criterion of reasonable time.
story drift, the limit state function can be expressed as
2.2 Annual probability of failure
G(X)  lim  Cd cal (2)
where δcal=calculated design story drift and δlim=allowable In order to obtain the annual probability of failure of the
story drift prescribed in Clause 10.12 of SBC 301 (2007), shear wall-frame structure, annual probability of occurrence
Cd=the deflection amplification factor, which is a random of the earthquake is required. For this purpose, the mean
variable and accounts for the inelastic behavior. occurrence rate of an earthquake in the selected site class is
The calculated design story drift (δcal) can be obtained required. Assuming that the number of earthquakes that
by the difference of the lateral deflections at the top and occur within a certain time interval follow a Poisson
bottom of the story under consideration (i.e., δi=yi-yi-1). In distribution, the annual probability of occurrence of an
the present study, the lateral deflections (yi) for shear wall- earthquake (or the annual probability of at least one
frame structure were estimated using the equation derived earthquake) can be estimated. Having known the annual
by Tuken and Siddiqui (2013) as given below probability of occurrence of the earthquake, the absolute
annual failure probability of the building Pfa can be
K.y( x)  A1s 2 cosh   A2 s 2 sinh  
estimated by
 1 4 k k3 k5 
2
1  2  pH    
 v   6 12 120  (3) Pfa  Pfn  Pann _ earthqke (8)
s 2 .p.k 2
.x  A3.x  A4
6.v 2 where Pfn=probability of failure when the building is
subjected to the earthquake; Pann_earthqke=annual probability
where, K=total stiffness of all shear walls and columns
of occurrence of the earthquake. The corresponding value
within the story along the axis considered=K (shear
of the annual reliability index βa can then be calculated as
walls)+ΣK (columns); H=height of the building; y=lateral
deflection of the building at a height of x from the base; a  1 Pfa  (9)
k=x/H; p=top intensity of triangular distributed lateral load;
s2=K/(v2GA); GA=the shear rigidity of the frame per unit
height (i.e., equivalent shear stiffness of the building); 2.3 Life-time probability of failure
v2=1+K/K0; K0=flexural rigidity of the structure in the
Assuming that the design life of the structure is Nd years
horizontal plane; ϕ=x/s. In the above equation, the
and the probability of failure in each year remains constant
coefficients A1, A2, A3 and A4 are defined as
and independent during lifetime, the probability of failure in
ps 2   λ 1   the entire life of the structure can be derived using Binomial
A1  1     sinh λ 
v cosh λ   2 λ 
2 (4) distribution as given below.

N 
PfL   d  Pfa (1  Pfa ) Nd 1 (10)
ps  λ 1 
2
 1 
A2   2    (5)
v 2 λ
2.4 Failure probability considering human errors
2
A3=−A2s and A4=−A1s (6)
If the probability of occurrence and non-occurrence of
where λ=H/s. human error are P(E1) and P(E 2) respectively and the
For pure frame, the above lateral deflection formula probability of structural failure (i.e., probability of
simplifies into exceeding the allowable drift) with and without this error
pH 2  k3  are Pf E1 and Pf E 2 respectively, then the total failure
y ( x)   k   (7)
2GA  3
probability Pf considering human error can be obtained by
722 Ahmet Tuken, Mohamed A. Dahesh and Nadeem A. Siddiqui

Fig. 1 Plan of the building without shear walls Fig. 2 Plan of the building with shear walls

using the theorem of total probability as below failure cost should be added ‘as is’ to the initial cost to get
the total cost. Thus, for all practical purposes, the failure
Pf  Pf E1  P( E1 )  Pf E2  P( E2 ) (11) cost must be multiplied by the failure probability of the
building to obtain the total cost. That is
where, E1, E2 are events that denote the occurrence and non-
CT  CI  PfLCF (15)
occurrence of human error and P(E1), P(E2) represent the
corresponding probabilities; Pf E1 , Pf E 2 are probabilities where, CT=the total cost of the building; CI=initial cost;
CF=failure cost; PfL=life-time failure probability. It is worth
of failure of the building under the condition of occurrence
mentioning that CI is a function of target (or desired)
and nonoccurrence of human errors respectively. Since E1,
reliability index as mentioned above and shown by Eq. (14).
E2 are mutually exclusive and collectively exhaustive
The failure cost can be expressed as a multiple of initial
events then
cost, that is
P(E1 )  P(E2 )  1 (12)
CF  CI (16)
Substituting, Pf E1  Pf 1 ; Pf E2  Pf 0 ; where γ is a multiplying factor. Substituting CI and CF from
Eqs. (14) and (16) respectively into Eq. (15), yield the
P( E1 )  P( E) and P( E2 )  1  P( E) , we have equation for the total cost as follows
Pf  (1  PE ) Pf 0  PE Pf 1 (13)  
CT  1  PfL T1  2T CI0  (17)

The above equation shows the dependence of the

2.5 Risk-based cost assessment
building’s total cost on failure probabilities and the selected
target reliability index βT. The above equation will be
The total cost of the building (for the purpose of the risk
employed to carry out the simple risk assessment of the
assessment or insurance) can be divided into two parts:
building.
initial cost and failure cost. In a probabilistic sense, the
initial cost depends on the target reliability index (or target
probability of failure) of the building in its life-time. In the
3. Numerical study and discussion of results
present study, the following equation was proposed to show
the dependence of initial cost over the target reliability
3.1 Description of the selected building
index (βT) of the building.

 
CI  T1  2T CI0 (14) A 10 story RC frame building was selected for the
present study. The detailing of the RC frame without shear
where α1 and α2 are factors that allow expected variation in walls satisfy the requirements of special RC moment
the initial cost (with target reliability index βT). The values frames, and with shear walls it satisfies the requirements of
of α 1 and α 2 are very much dependent on the type, the dual system (with special moment frames) as per Clause
importance, and location of the building. In the above 10.2 of SBC 301 (2007). The building is rectangular in plan
having a total height of 40 m with 7-bays in x-direction and
equation, C I 0 is the initial cost of the building
5-bays in the y-direction and a constant floor plan area of
corresponding to a fixed target reliability index value (e.g., 816 m2 at each story. The height of every story is equal to 4
βT=3.5). In order to derive expression for the total cost of m. The building is assumed to be fixed at the base. The
the building, the two extreme cases were considered: (i) if floors of the building are considered to act as rigid
likelihood of the building failure is zero, the failure cost diaphragms. All the columns, beams and slabs were
will be zero and thus failure cost will not be a part of the considered to be of the same sizes (columns: 500×500 mm;
total cost; (ii) if failure is ‘certain’ in the design life, the beams: 200×500 mm; slab thickness: 150 mm). Fig. 1
 Reliability assessment of RC shear wall-frame buildings subjected to seismic loading 723

Table 1 Basic random variables and their statistical values

Nominal
Random Variable Bias factor COV Distribution
value
Story weights, wi 8.0 kN/m2 1.05 0.15 Extreme Type I
Design spectral response acceleration at
0.664 0.78 0.16 Normal
short period, SDS
Design spectral response acceleration at
0.344 0.78 0.16 Normal
1-sec period, SD1
Modulus of elasticity of concrete, E 23500 MPa 1.00 0.18 Lognormal

Beam width 200 mm 1.00 0.05 Lognormal

Beam depth 500 mm 1.00 0.05 Lognormal

Column width 500 mm 1.00 0.05 Lognormal

Column depth 500 mm 1.00 0.05 Lognormal

Shear wall thickness 300 mm 1.00 0.05 Lognormal

Compressive strength of concrete, fc’ 25 MPa 1.00 0.10 Normal

Fig. 3 Acceleration response spectrum (SBC 301) Yield strength of steel, fy 420 MPa 1.00 0.08 Normal
Response modification factor, R
6.5 1.00 0.05 Normal
(for pure frame)
Response modification factor, R
6.5 1.00 0.05 Normal
shows the plan of the building. The frame is subjected to (for dual system)

dead, live, and horizontal seismic loads and it was stiffened Occupancy importance factor, I 1.0 1.00 0.05 Normal

with shear walls in x- and y-directions as shown in Fig. 2. Deflection amplification factor, Cd
5.5 1.00 0.05 Normal
(for pure frame)
The thickness of all the shear walls was 300 mm and the Deflection amplification factor, Cd
6.5 1.00 0.05 Normal
(for dual system)
ratio of horizontal web reinforcement of the wall to the
gross area of the wall, ρn, was 0.025 (minimum
reinforcement as per SBC 304).
The reliability of the above RC building containing
shear walls was studied in the present paper. The statistical
data and probability distributions of the selected random
parameters, required for the reliability analysis, are given in
Table 1. The probability distribution of different random
variables was taken from the various sources e.g., Lee and
Haldar (2003), Nowak and Collins (2012) and Dahesh et al.
(2014), Dahesh (2015). The variables shown in Table 1
were considered random as they have considerable
uncertainty in their values. All the other values were
assumed to be deterministic.
The total base shear was estimated using an acceleration
response spectrum, defined in Saudi Building Code (SBC
301). This spectrum is shown in Fig. 3. The acceleration
response spectrum reveals that the effective ground
acceleration is magnified by a factor of SDS, for natural
periods of To–Ts seconds. SDS and SD1, shown in Fig. 3, are
the nominal values of design spectral response accelerations
at short periods and at 1-sec period respectively. Their
typical values for Haql City, a seismically active region of Fig. 4 A comparison of analytical and ETABS 2013 results
the Kingdom of Saudi Arabia, are given in Table 1.
The nominal value of the story weight is taken to be 8.0
kN/m2. This weight includes self-weight of the slab, weight al. 2014, Siddiqui et al. 2009, Siddiqui 2011, Siddiqui et al.
of floor finishes and superimposed dead and live loads. 2014).
The probability of failure of the RC building frames
against strength limit state is, in general, very small as these 3.2 Validation of the structural drift formulation
frames (with or without shear walls) are substantially strong
against the maximum expected base shear (Lee and Haldar To validate the formulation, presented in Section 2, for
2003a, 2003b). It is due to this reason, in the present study; computing the lateral deflections (yi) the response of
the reliability analysis has been carried out only against analytical formulation for the shear wall-frame building was
serviceability limit state of lateral drift. The results of the compared with ETABS 2013 results and shown in Fig. 4.
numerical studies have been presented and discussed in the The other data required for obtaining the structural drift
following sections. In the reliability discussion, desired response are given in the second column of Table 1. It can
reliability was taken as 3.5. This is a typical value of the be observed from Fig. 4 that the two responses are
reliability index which is generally used to assure a reasonably close to each other. This gives a confidence on
desirable safety level in buildings and structures (Dahesh et the analytical formulation which was employed in the
724 Ahmet Tuken, Mohamed A. Dahesh and Nadeem A. Siddiqui

Table 2 Probability of structural failure when earthquake Table 3 Annual probability of structural failure
hits the structure Building frame Building frame
Building frame Building frame Building Frame without shear wall with shear wall
Building Frame without shear wall with shear wall Pfa βa Pfa βa
Pfn βn Pfn βn Earthquake in -3 -7
1.15 × 10 3.049 2.44 × 10 5.031
Earthquake in x-direction
2.35 × 10-2 1.986 5.00 × 10-6 4.417 Earthquake in
x-direction 1.41 × 10-3 2.987 3.22 × 10-7 4.977
Earthquake in -2 -6
y-direction
2.89 × 10 1.897 6.60 × 10 4.357
y-direction

reliability index of 3.5 which improves to above 4.0 due to

derivation of the limit state function and subsequently in the the presence of shear walls. The improvement of the
present reliability analysis. reliability due to shear walls can be attributed to the
increased lateral stiffness of the building. The shear walls
3.3 Reliability of the frame building with and without reduced the lateral drift of the building considerably and
shear walls thus improved the reliability substantially. Table 2 also
shows that the probability of failure of the building in the x-
Table 2 shows the results of the Monte Carlo simulation- direction is lesser than the probability of failure in the y-
based reliability analysis performed on the building with no direction. This can be attributed to the higher stiffness of the
shear walls (Fig. 1) and building with shear walls (Fig. 2). building in the x-direction compared to the y-direction. This
Two million simulations were used for carrying out the study thus clearly illustrates the beneficial effects of shear
reliability analysis of building frame containing no shear walls in improving the reliability by substantially reducing
walls and about 15 million simulations were employed for the lateral displacement at the top of the building frame.
carrying out the reliability analysis of building frames with This is worth mentioning that the trend obtained may
shear walls. A substantially higher number of simulations substantially modify if the building is irregular instead of
were used for the reliability analysis of the shear wall-frame regular (as considered in the present study). The irregularity
building because the expected probability of failure of the in the building can be owing to plan irregularity or vertical
shear wall-frame building was of the order of 10 -6. irregularity. Plan irregularity is due to the difference
The accuracy of Monte Carlo Simulation was evaluated between the center of mass and the center of resistance of
by computing the variance of the estimated probability of the building, whereas vertical irregularity is due to abrupt
failure Pf. The variance was computed by assuming (i) each changes in the geometry, strength, or stiffness of the
simulation cycle constitutes a Bernoulli trial and (ii) the structure from floor to floor.
number of failures in n trials follow a binomial distribution.
The variance of the probability of failure was then 3.4 Annual probability of failure of the RC shear wall-
approximately computed as frame building

Var ( Pf ) 
1  P P
f f
(18)
In order to obtain the annual probability of failure of the
shear wall-frame building the mean occurrence rate of the
n earthquake having the nominal values of design spectral
response accelerations at short periods, SDS, and at 1-sec
The statistical accuracy of the calculated failure
period SD1, as given in Table 1, is required. In the present
probability was estimated by computing its coefficient of
study, the mean occurrence rate was selected to be 0.05
variation (COV) as
which corresponds to a return period of 20 years (Al-Amri
1  P P f f
2014). Assuming that the number of earthquakes that occur
within a certain time interval follow a Poisson distribution
n (19) (Nowak and Collins 2012), the annual probability of
COV( Pf )  occurrence of an earthquake (or the annual probability of at
Pf
least one earthquake) was estimated as
The smaller the COV, the better the accuracy of the P  P(n  1)  1  P(n  0)  1 
t n e t
ann _ earthqke
estimated probability of failure is. n!
However, for all practical purposes, the number of 1-
0.05  10 e 0.051  0.0488
simulation cycles for which COV(Pf) approaches less than 0!
5% may be considered as an appropriate number of Here, t=time in years; λ=mean occurrence rate
simulation cycles (Nowak and Collins, 2012). In the present (=1/return period); n=number of earthquakes; and
study, around 1 to 15 million of Monte Carlo simulations P ann _ earthqke =annual probability of occurrence of the
were required to achieve the desired accuracy in the earthquake. Having known the annual probability of
estimation of the failure probabilities. occurrence of the earthquake and the nominal probability of
Table 2 clearly illustrates the advantage of using shear failure of the building (probability of failure when the
walls in improving the serviceability-reliability of the building is subjected to the earthquake, Pfn), the absolute
building. When there were no shear walls the reliability annual failure probability of the building was estimated
index of the building was substantially less than the desired using Eq. (8). The corresponding value of the annual
 Reliability assessment of RC shear wall-frame buildings subjected to seismic loading 725

Table 4 Life-time failure probability of the building

β
5
Building frame Building frame PE = 0.0167
PE = 0.0167
Building Frame without shear wall with shear wall
4
PfL βL PfL βL
Earthquake in x- -2 -5 PE = 1.0 3
5.42 × 10 1.605 1.22 × 10 4.220
direction PE = 1.0
Earthquake in y-
6.58 × 10-2 1.508 1.61 × 10-5 4.157 2
direction Total load
fc'
1

reliability index βa was then calculated using Eq. (9). The 0

results are presented in Table 3 which clearly illustrates that -0.5 -0.3 -0.1 0.1 0.3 0.5
due to the small annual probability of earthquake Deviation from error-free value

occurrence, the annual failure probability of the building is Fig. 5 Effect of human error in some of the governing
one order (i.e., one tenth) lesser than the nominal parameters on shear wall-frame building (considering
probability of failure. The reliability indices have also probability of human error occurrence, PE=1.0 and
improved significantly. In fact, even frame building which PE=0.0167)
had no shear walls has achieved reliability index up to 3.0.

3.5 Life-time failure probability of the RC shear wall- present study, the effect of the human error was considered
frame building only in the two major parameters-one in the determination
of concrete strength and the other in the selection of load
Assuming that the design life of the structure is 50 years value as the human error in these two parameters are very
and the probability of failure in each year remains constant probable and common.
and independent during lifetime, the probability of failure in The positive deviation was considered for the load and
the entire life of the structure was estimated using Eq. (10). negative for the concrete strength to study the adverse
The results of the analysis are presented in Table 4 which effects of human error. Fig. 5 clearly illustrates that there is
shows that the life-time failure probability is substantially a sharp decrease in the reliability index with increasing
higher than the annual probability of failure. This is due to magnitude of deviation from error-free value. The results
the fact that the life-time failure probability can be treated presented in this figure are for the earthquake acting in the
as the failure probability of a series system, and in a series x-direction. A similar graph is expected if the earthquake
system, if any element fails, the system fails. Due to this hits the building in the y-direction. The probability of
reason system failure probability (i.e., life-time failure human error, PE=1.0, mentioned in the figure indicates that
probability) is much higher than the individual element’s how the reliability index will be affected when (certainly)
probability of failure (i.e., annual failure probability). there is a human error in the estimation of the total load
Table 4 clearly shows that the life-time reliability and/or concrete strength.
indices (obtained from life-time probabilities of failure Fig. 5 also shows the reliability index variation of the
using Eq. (9)) for a frame building without shear walls are structure when the human error is not certain but has a
much less than the desired reliability index value. However, probability of occurrence of 0.0167 (Melchers 2002). In this
with shear walls, the life-time reliability indices are higher case, the total reliability index of the structure is almost
than the target reliability index value of 3.5. constant with the variation of deviation from error-free
As the influence of earthquake on reliability is similar in value. It is due to the low probability of occurrence of the
x- and y-directions, in the following studies, the results will human error. Mathematically, if the probability of
be presented for earthquake acting in x-direction only. occurrence of human error is PE, and the probability of
structural failure without and with this error is Pf0 and Pf1
3.6 Influence of human error on reliability of RC shear respectively, then the total failure probability can be
wall-frame building obtained by Eq. (13). As an extreme case if PE=0, then
Pf=Pf0 and if PE=1, Pf=Pf1. Thus, if there is a small
The human error was thought to cause the deviation in probability of human error occurrence, there will be a very
nominal values from their error-free values by a certain little change in the overall probability of failure of the
amount. The nominal values presented in Table 1 are building. It is due to this reason, in Fig, 5, there is no
assumed to be the values which are free from human error. significant fall in the reliability index when nominal values
Due to the human error, the nominal values deviate. This deviate from the error-free value with a probability of
deviation could be either positive or negative. The positive occurrence of human error as 0.0167.
deviation indicates an increase in the nominal value while
negative deviation shows a decrease. 3.7 Risk-based cost assessment of RC shear wall-
There is a possibility of human error at many stages frame building
from design to construction (Epaarachchi and Stewart 2004,
De Haan 2012). Human error involvement is very much In the present study employing Eq. (14) through Eq.
expected in the determination of concrete strength, steel (17) with their selected coefficient values α1, α2 and γ as 2.0,
yield strength, loads, and workmanship etc. However, in the -0.5 and 1.5 respectively, the initial cost and the total cost of
726 Ahmet Tuken, Mohamed A. Dahesh and Nadeem A. Siddiqui

Table 5 Earthquake parameters considered for sensitivity

analysis
Ss S1 SMS SDS SM1 SD1
Fa Fv
Site Class (g) (g) (g) (g) (g) (g)
A 1.0 0.4 0.8 0.8 0.800 0.533 0.320 0.213
B 1.0 0.4 1.0 1.0 1.000 0.667 0.400 0.267
C 1.0 0.4 1.0 1.4 1.000 0.667 0.560 0.373
D 1.0 0.4 1.1 1.6 1.100 0.733 0.640 0.427
E 1.0 0.4 0.9 2.4 0.900 0.600 0.960 0.640

Fig. 6 Variation of initial and total costs of shear-wall frame

building with target reliability index (a measure of risk)

the shear wall-frame building were estimated for different

values of target reliability index and the results are shown in
Fig. 6. This figure illustrates that as the target reliability
index increases, the initial cost, as well as the total cost of
the building, rises sharply. The difference between the
initial cost and final cost is not much due to small life-time
failure probability of the building. The two curves match
exactly when PfL is zero. This is very well expected because (a)
when life-time failure probability is zero, the failure cost
becomes zero which makes the total cost same as the initial
cost. The curve clearly illustrates that when the selected
level of risk or probability of failure is high (i.e., when
target reliability index is small), the cost is substantially
small, but when the selected level of risk or probability of
failure is small (i.e., when target reliability index is high)
cost is substantially high. The curve also shows that when
target reliability index is 3.5, the initial cost is CI 0 and

the total cost is little higher than CI 0 . It is due to the fact

that C I0 is the initial cost corresponding to the target (b)

reliability index=3.5. It is worth mentioning that the values Fig. 7 Effect of site class on probability of failure and
of coefficients α1 and α2 depend on how the initial cost was reliability of building frame
assumed to vary with target reliability index. On the other
hand, the value of coefficient γ depends on how the initial
and the failure costs are related to each other. The selected its poor lateral stiffness. However, when the shear walls are
values of α1, α2 and γ may change as the above-mentioned provided there is a dramatic decrease in the failure
dependence change. probability due to increased lateral stiffness of the building.
Fig. 7 shows that for the site class A to D the decrease in the
3.8 Sensitivity analysis failure probability due to provided amount of shear walls
makes the building as reliable as desired (i.e. reliability
3.8.1 Effect of seismic design category on structural index becomes greater than 3.5). However, for the site class
reliability E the present quantity of shear wall is not sufficient as
In this study, the same building was assumed to exist in reliability index is less than the target reliability index. This
different sites i.e., site class A through E (SBC 301) and its indicates that more shear walls are required to increase the
reliability index and the probability of failure were reliability of the building to the desired level.
obtained. In this analysis, the earthquake parameters related
to different site classes are tabulated in Table 5 and using 3.8.2 Effect of story weight on reliability of RC shear
these data the reliability analysis was carried out for wall-frame building
building frame without and with shear walls and the results The effect of story weight on reliability index of shear-
obtained are shown in Fig. 7 which shows that building wall frame building was studied to obtain the results of
without the shear wall is having a substantially high design interest. For this purpose, the parameters were
probability of failure for the site classes C through E due to selected in such a way that (β – βT)2≈0. Here β and βT are
 Reliability assessment of RC shear wall-frame buildings subjected to seismic loading 727

increase of story weight, base shear increases which in turn

increases the story drift. Consequently, the probability of
reaching to the limiting drift value increases; thus,
reliability decreases or probability of serviceability failure
increases. Fig. 8(b) shows that the present building is
reliable to the desired extent provided the story weight is
approximately less than or equal to 10 kN/m2. Beyond this
story weight, reliability will sharply decrease to a value less
than 3.0. However, this optimum value is approximately 9.5
kN/m2 for achieving the life-time reliability index of 3.5.
(a)
3.8.3 Effect of shear wall ratio
The effect of shear wall ratio i.e., shear wall plan area to
floor plan area ratio on the reliability of shear wall-frame
building was studied by varying the nominal thickness of
the shear walls. Fig. 9(a) shows that as the shear wall ratio
is increasing, reliability is continuously increasing. This is
because with the increase of shear wall ratio, building
stiffness increases which in turn decreases the story drift.
Consequently, the probability of reaching to the limiting
drift value decreases; thus, reliability increases or
(b) probability of serviceability failure decreases. Fig. 9(b)
Fig. 8 Variation of reliability index β with story weight shows that the present building is reliable to the desired
extent provided the shear wall ratio is approximately more
than or equal to 0.8%. Beyond this shear wall ratio, the
reliability of the building will be more than 3.5. However,
this value is approximately 1.0% for the life-time reliability
index of 3.5.

4. Conclusions

The following conclusions can be drawn from the

present reliability study of a RC shear wall-frame building
against serviceability limit state.
i. Shear walls play an important role in reducing the risk
or improving the reliability of the building under seismic
(a) excitation. Even a small quantity of shear walls (1% or less
of floor plan area) can improve the reliability of RC frame
building dramatically. It was observed that an RC frame
building (without shear walls) whose life-time reliability is
substantially smaller than the desired/target reliability index
value of 3.5, achieves the desired reliability level in the
presence of a small quantity of shear walls.
ii. The annual failure probability of the studied shear
wall-frame building is approximately one order (i.e., one
tenth) lesser than the nominal probability of failure (i.e., the
probability of failure when the building is subjected to the
earthquake) of the building.
iii. There is a sharp change in the reliability of the
studied shear wall-frame building due to human error
(b)
involvement in the estimation of total load and/or concrete
Fig. 9 Variation of reliability index β with shear wall ratio strength. However, when the probability of occurrence of
(%) the human error is small, the reliability of the building is
almost unaffected.
iv. As the target reliability index increases, the initial
the actual and target reliability index values. (β–βT)2≈0 cost, as well as the total cost of the building, rises sharply.
indicates that the reliability of the shear wall-frame building The curves of initial and failure cost match exactly when
is almost equal to the target reliability value. Fig. 8(a) the life-time probability of failure is assumed to be zero.
shows that as the story weight is increasing reliability is v. For site class A to D the decrease in the failure
continuously decreasing. This is due to the fact that with the probability due to shear walls is to an extent that the studied
728 Ahmet Tuken, Mohamed A. Dahesh and Nadeem A. Siddiqui

RC building (with shear walls) becomes as reliable as buildings. I: Assessment of ductile moment frames”, J. Struct.
desired. However, for site class E the present quantity of Eng., 137(4), 481-491.
shear wall is not sufficient and thus more shear walls are Jeong, S., Mwafy, A.M. and Elnashai, A.S. (2012), “Probabilistic
required. On the other hand, the provided quantity of shear seismic performance assessment of code-compliant multi-story
walls is enough for the studied building if it is located in RC buildings”, Eng. Struct., 34, 527-537.
site class A through D. Lee, S.Y. and Haldar, A. (2003a), “Reliability of frame and shear
wall structural systems I: Static loading”, J. Struct. Eng.,
vi. The studied building (with the shear walls) is reliable
129(2), 224-232.
to the desired extent provided that the total story weight is Lee, S.Y. and Haldar, A. (2003b), “Reliability of frame and shear
approximately less than or equal to 10 kN/m2. However, wall structural systems II: Dynamic loading”, J. Struct. Eng.,
this optimum value is approximately 9.5 kN/m2 for 129(2), 233-240.
achieving the life-time reliability index of 3.5. Martins, L., Silva, V., Marques, M., Crowley, H. and Delgado, R.
vii. The studied building is reliable to the desired extent (2016), “Development and assessment of damage-to-loss
provided the shear wall ratio is approximately more than or models for moment-frame reinforced concrete buildings”,
equal to 0.8%. Beyond this shear wall ratio, the reliability Earthq. Eng. Struct. Dyn., 45(5), 797-817.
of the building is more than 3.5. However, this value is Marzban, S., Banazadeh, M. and Azarbakht, A. (2014), “Seismic
performance of reinforced concrete shear wall frames
approximately 1.0 % for the life-time reliability index of
considering soil-foundation-structure interaction”, Struct. Des.
3.5. Tall Spec. Build., 23(4), 302-318.
Melchers, R.E. (2002), Structural Reliability: Analysis and
Prediction, 2nd Edition, Wiley, New York, U.S.A.
Acknowledgments Monteiro, R. (2016), “Sampling based numerical seismic
assessment of continuous span RC bridges”, Eng. Struct., 118,
The work presented in this paper was funded by the 407-420.
Deanship of Scientific Research, Research Centre, College Monteiro, R., Delgado, R. and Pinho, R. (2016a), “Probabilistic
of Engineering, King Saud University, Riyadh, Saudi seismic assessment of RC bridges: Part I-uncertainty models”,
Structures, 5, 258-273.
Arabia through research grant number 3/437.
Monteiro, R., Delgado, R. and Pinho, R. (2016b), “Probabilistic
seismic assessment of RC bridges: Part II-nonlinear demand
prediction”, Struct., 5, 274-283.
References Nowak, A.S. and Collins, K.R. (2012), Reliability of Structures,
2nd Edition, Taylor and Francis Group, CRC Press, U.S.A.
Al-Amri, A.M.S. (2014), Mechanism of Earthquakes and Paulay, T. (1999), “A simple seismic design strategy based on
Seismicity Modeling, Ministry of Municipal and Rural Affairs- displacement and ductility compatibility”, Earthq. Eng. Eng.
Riyadh, Saudi Arabia. Seismol., 1(1), 51-67.
Burak, B. and Comlekoglu, H.G. (2013), “Effect of shear wall area Priestley, M.J.N., Calvi, G.M. and Kowalsky, M.J. (2007),
to floor area ratio on the seismic behavior of reinforced concrete Displacement-Based Seismic Design of Structures, IUSS Press,
buildings”, J. Struct. Eng., 139(11), 1928-1937. Pavia, 670.
Dahesh, M.A. (2015), “Structural reliability of RC shear wall- SBC 301 (2007), Loading and Forces Requirements, Saudi
frame buildings subjected to earthquake loading”, M.Sc. Building Code, Riyadh, Saudi Arabia.
Dissertation, King Saud University, Riyadh, Saudi Arabia. SBC 304 (2007), Concrete Structures Requirements, Saudi
Dahesh, M.A., Tuken, A. and Siddiqui, N.A. (2014), “Effect of Building Code, Riyadh, Saudi Arabia.
shear wall in improving the reliability of RC frame buildings Sezen, H., Whittaker, A.S., Elwood, K.J. and Mosalam, K.M.
subjected to seismic loading”, Proceedings of the International (2003), “Performance of reinforced concrete and wall buildings
Conference on Construction Materials and Structures, during the August 17, 1999 Kocaeli, Turkey earthquake, and
Johannesburg, South Africa, November. seismic design and construction practice in Turkey”, Eng.
Dahesh, M.A., Tuken, A. and Siddiqui, N.A. (2015), “Controlling Struct., 25(1), 103-114.
the earthquake induced lateral displacement of RC buildings Siddiqui, N.A. (2011), “Influence of fluidity on reliability of SCC
using shear walls: Parametric study”, Arab. J. Geosci., 8(11), produced using local Saudi materials”, Arab. J. Sci. Eng., 36(2),
9913-9927. 203-214.
De Haan, J. (2012), “The design of a human reliability assessment Siddiqui, N.A. and Khan, F.H. and Umar, A. (2009), “Reliability
method for structural engineering”, M.Sc. Dissertation, Delft of underground concrete barriers against normal missile
University of Technology, Delft, the Netherlands. impact”, Comput. Concrete, 6(1), 79-93.
Douglas, J., Ulrich, T. and Negulescu, C. (2013), “Risk-targeted Siddiqui, N.A. and Khateeb, B.M.A., Almusallam, T.H., Al-
seismic design maps for mainland France”, Natur. Haz., Salloum, Y.A., Iqbal, R.A. and Abbas, H. (2014), “Reliability
65(3),1999-2013. of RC shielded steel plates against the impact of sharp nose
Epaarachchi, D.C. and Stewart, M.G. (2004), “Human error and projectiles”, J. Imp. Eng., 69, 122-135.
reliability of multistory reinforced-concrete building Siddiqui, N.A. and Khateeb, B.M.A., Almusallam, T.H., and
construction”, J. Perform. Constr. Facilit., 18(1), 12-20. Abbas, H. (2014), “Reliability of double-wall containment
Goulet, C.A., Haselton, C.B., Mitrani-Reiser, J., Beck, J.L., against the impact of hard projectile”, Nucl. Eng. Des., 270,
Deierlein, G.G., Porter, K.A. and Stewart, J.P. (2007), 143-151.
“Evaluation of the seismic performance of a code-conforming Soleimani-Abiat, M. and Banan, M. (2015), “Seismic behavior of
reinforced-concrete frame building-from seismic hazard to RC building by considering a model for shear wall-floor slab
collapse safety and economic losses”, Earthq. Eng. Struct. Dyn., connections”, Comput. Concrete, 16(3), 381-397.
36(13), 1973-1997. Taleb, R., Bechtoula, H., Sakashita, M., Bourahla, N. and Kono, S.
Haselton, C.B., Liel, A.B., Deierlein, G.G., Dean, B.S. and Chou, (2012), “Investigation of the shear behaviour of multi-story
J.H. (2011), “Seismic collapse safety of reinforced concrete reinforced concrete walls with eccentric openings”, Comput.
 Reliability assessment of RC shear wall-frame buildings subjected to seismic loading 729

Concrete, 10(4), 361-377.

Tuken, A. (2004), “Analysis and assessment of seismic drift of
reinforced concrete mixed (shear wall-frame) structures”,
Technol., 7(4), 523-532.
Tuken, A. and Siddiqui, N.A. (2011), “A simplified analytical
procedure to determine the amount of shear walls in reinforced
concrete buildings”, Proceedings of the 7th International
Conference on Analytical Models and New Concepts in
Concrete and Masonry Structures, Krakow, Poland, June.
Tuken, A. and Siddiqui, N.A. (2013), “Assessment of shear wall
quantity in seismic-resistant design of reinforced concrete
buildings”, Arab. J. Sci. Eng., 38(10), 2639-2648.
Ulrich, T., Negulescu, C. and Douglas, J. (2014), “Fragility curves
for risk-targeted seismic design maps”, Bullet. Earthq. Eng.,
12(4), 1479-1491.
Wyjadtowski, M., Pula, W. and Bauer, J. (2015), “Reliability of
diaphragm wall in serviceability limit states”, Arch. Civil Mech.
Eng., 15(4), 1129-1137.

CC