See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/301337059

Seismic Performance Assessment of a Code

Compliant Multistorey Building
Conference Paper June 2016

READS

154

3 authors, including:
Giuseppe Marcantonio Del Gobbo

Anthony Blakeborough

University of Oxford

University of Oxford

2 PUBLICATIONS 0 CITATIONS

68 PUBLICATIONS 868 CITATIONS

SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate,

letting you access and read them immediately.

SEE PROFILE

Available from: Giuseppe Marcantonio Del Gobbo

Retrieved on: 18 April 2016

Seismic Performance Assessment of a Code Compliant Multistorey Building

Giuseppe M. DEL GOBBO1, Martin S. WILLIAMS2, Anthony BLAKEBOROUGH3
1

University of Oxford, Oxford, UK, giuseppe.delgobbo@eng.ox.ac.uk

2
University of Oxford, Oxford, UK, martin.williams@eng.ox.ac.uk
3
University of Oxford, Oxford, UK, tony.blakeborough@eng.ox.ac.uk

Abstract: Recent earthquakes have revealed that conventional seismic design philosophy allows for
undesirable levels of nonstructural damage. Inadequate nonstructural seismic performance has led to
extensive repair costs and lengthy functional disruptions, as nonstructural systems comprise the majority of
building investments and are essential to building operations. This study investigates the seismic
performance of a Eurocode compliant multistorey building. A 16 storey steel building was designed to be
representative of modern seismically designed structures. This model was subjected to nonlinear time history
analyses representing two ground motion intensities in the finite element program OpenSees. The FEMA
P-58 performance assessment procedure was used to determine repair costs with a focus on nonstructural
systems. The seismic performance assessment determined that drift-sensitive and acceleration-sensitive
nonstructural systems would experience damage. The results indicate that significant repair costs and
downtime can be expected in modern multistorey structures following an earthquake. This suggests that
modern building standards do not accomplish resilience, the ability of a community to quickly recover after
an earthquake. Limitations of the Eurocode damage mitigation methodology were revealed as the prescribed
interstorey drift ratio limits did not prevent nonstructural damage. These results highlight the need for design
procedures which enhance nonstructural seismic performance. The results of this analysis provide a
benchmark on which to compare the effectiveness of alternative designs in improving nonstructural seismic
performance. This benchmark is valuable when evaluating retrofit alternatives for an existing building or
when selecting design options for new structures.
Keywords: Seismic Performance Assessment, Nonstructural Systems, FEMA P-58, Eurocode

1. Introduction
Buildings designed to modern structural codes are able to withstand earthquakes while undergoing
minimal to no structural damage. Although this suggests that a quick recovery is possible, this result is
not always attained. Recent earthquakes such as the 2010 Canterbury earthquake and the 2010 Chile
earthquake have demonstrated that buildings which experience minimal structural damage frequently
incur extensive damage to nonstructural systems (Dhakal 2010; Miranda et al. 2012). Nonstructural
systems include building contents, architectural components, and mechanical, electrical and plumbing
systems. These systems are essential to building operations and comprise the majority of building
investment as shown in Figure 1. Nonstructural damage causes lengthy functional disruptions and
accounted for several billion dollars of losses in 2010 alone (Fierro et al. 2011).
Attaining a target level of seismic performance mandates the harmonization of structural and
nonstructural performance levels. This study assesses the seismic performance of a 16 storey steel
building that is representative of conventional multistorey buildings constructed in seismic regions. The
case study provides a benchmark on which to evaluate retrofit alternatives for an existing building or when
selecting design options for new structures. The final aim of the project is to develop a viscous fluid damper
placement strategy that minimizes nonstructural repair costs and building service disruptions.

Building Investment

80%

Structural
Nonstructural
Contents

60%
40%
20%
0%
Office Building

Hotel

Hospital

Figure 1 Distribution of building investment for three building categories. Data

from Taghavi and Miranda (2003)

2. Case study building

A Eurocode compliant building design was created in order to evaluate the seismic performance of
conventionally designed structures (CEN 2010). The case study structure is a 16 storey steel office building.
The lateral load resisting system consists of concentrically braced frames located around the perimeter of
the structure. Seismic design was conducted through modal response spectrum analysis in SAP2000 (CSI
2013). The peak ground acceleration used for the design was 0.31g and the fundamental period of the
building (T1) is 2.34 seconds. An elevation and a plan view are shown in Figure 2. Complete building design
information can be found in a previous report (Del Gobbo 2014).

Figure 2 Elevation and plan views of the case study building

A 2D model of the structure was created in the finite element program OpenSees in order to perform
nonlinear analyses (PEER 2015). A leaning column was employed to account for P- effects from the vertical
loads acting on gravity columns in the tributary plan area. Inherent damping was incorporated using mass
and tangent stiffness proportional Rayleigh damping of 5%. The first and third periods were used to
determine the Rayleigh damping parameters as these modes account for 92% of the effective mass. The first
period was elongated to account for the expected nonlinear brace buckling and prevent the generation of
artificial damping forces (Charney 2008). The braces were modelled using the procedure from Uriz et al.
(2008). Each brace was modelled using two elements with an initial imperfection of 0.1% at the midspan to
induce buckling in compression. A fictitious load producing 5% of the section yield moment was applied at
the midspan in order to prevent brace straightening and ensure buckling. The analytical model was verified
using cyclic loading experimental data from Black et al. (1980) as shown in Figure 3. The hysteretic behaviour
is captured, as brace buckling in compression and yielding in tension are accurately predicted.

Seismic Performance Assessment of a Code Compliant Multistorey Building

Force (kN)

1200
1000
800
600
400
200
0
-200
-400
-600
-800

Experimental,
Black et al. 1980
OpenSees Model

-40

-30

-20

-10
0
10
20
Brace Elongation (mm)

30

40

50

Figure 3 Comparison of experimental and analytical brace hysteretic behaviour

3. Ground motion scaling and selection

Nonlinear time history analyses were performed using recorded ground motion suites representing the
ultimate limit state (ULS) and the serviceability limit state (SLS). The ULS earthquake has a 10% probability of
exceedance in 50 years and the SLS earthquake has a 10% probability of exceedance in 10 years. The ground
motion records were selected and scaled following the Eurocode 8 requirements, with a factor of 0.5 used
to define the SLS spectrum (CEN 2013). Records were obtained from the PEER ground motion database (PEER
2013). A linear scale factor was applied to each record which minimized the mean squared error between
the ground motion spectrum and the target Eurocode spectrum over the period range of 0.2T1 to 2T1. A
maximum of one record was selected per earthquake and the scale factor was limited to 2. ULS ground
motions were constrained to have a magnitude greater than 5.5 to match the Eurocode Type 1 spectrum and
all ground motions had an average shear wave velocity appropriate for ground type C (CEN 2013). 25 ground
motions with the smallest mean squared error were selected for each suite. Figure 4 compares the ground
motion suite spectrum and the Eurocode 8 elastic response spectrum for both the ULS and SLS.
2

2
Eurocode 8 ULS
Suite Mean
Mean +
Mean -

1.5

Sa (g)

Sa (g)

1.5

Eurocode 8 SLS
Suite Mean
Mean +
Mean -

0.5

0.5

0
0

2
Period (s)

2
Period (s)

Figure 4 Comparison of the ground motion suite spectrum and the Eurocode 8 spectrum. ULS (left) and SLS (right)

4. Nonlinear time history analysis results

Structural response parameters used to characterise demands on structural and nonstructural systems are
peak floor absolute accelerations, floor velocities and interstorey drift ratios (IDR). These parameters were
determined from the time history analyses of the case study building. The results are shown in Figure 5 for
both the ULS and SLS. These plots include the mean values as well as the values from each individual ground
motion. Peak floor velocity results were omitted as the values were relatively constant (ULS mean of 4.2 m/s
Seismic Performance Assessment of a Code Compliant Multistorey Building

and SLS mean of 2.5 m/s) throughout the building for each analysis. The ULS results have a greater spread
than the smaller intensity SLS. This is expected due to the significant nonlinear behaviour exhibited by the
braces during large intensity earthquakes.
17
15
13

13

11

11

1
0

0.5
1
Acceleration (g)

1.5

16

16

13

13

10

10

Storey

Storey

SLS Results
Mean

15

Floor

Floor

17

ULS Results
Mean

7
4

ULS Results
Mean

1
0.0%

1.0%

2.0%

3.0%

0.5
1
Acceleration (g)

1.5
SLS Results
Mean

7
4
1
0.0%

IDR

1.0%

2.0%

3.0%

IDR

Figure 5 Peak absolute accelerations (top) and peak interstorey drifts (bottom) determined from the
time history analyses for the ULS (left) and SLS (right)

5. Structural and nonstructural performance assessment

The FEMA P-58 performance assessment procedure (ATC 2012) was used to evaluate the seismic
performance of the conventional building. Structural response parameters from the time history analyses
were used in combination with fragility functions to determine repair costs. Fragility functions indicate the
probability of exceeding a damage state at a given engineering demand parameter (EDP) value for a specific
component. Fragility functions are represented by lognormal cumulative distribution functions.
Nonstructural normative quantities corresponding to a commercial office building were considered along
with all structural components. Table 1 provides a summary of the critical fragility information used in the
case study from FEMA P-58 (ATC 2012).

Seismic Performance Assessment of a Code Compliant Multistorey Building

Table 1 Summary of critical structural and nonstructural system fragility information

System

EDP

Damage State

xm

Concentric Braced Frame

IDR

Brace buckling and yielding

Brace fracture or local buckling

0.01
0.0178

0.3
0.3

Glass Curtain Wall

IDR

Glass cracking
Glass falls from frame

0.01097
0.01254

0.45
0.45

Gypsum wall partition with metal studs

IDR

Minor cracking
Moderate cracking or crushing
Significant cracking or crushing

0.0021
0.0071
0.012

0.6
0.45
0.45

Suspended Ceiling, vertical and lateral

support

PFA

Minor tiles dislodgment

Moderate tile dislodgment and grid damage
Total ceiling collapse

0.35
0.55
0.8

0.4
0.4
0.4

Air Handling Unit

PFA

Equipment does not function

0.25

0.4

Desktop Electronics

PFA

Falls, does not function

0.4

0.5

xm - median, - standard deviation, IDR - interstorey drift ratio, PFA - peak floor acceleration

6. Seismic performance results

The results of the seismic performance assessment are expressed by repair costs in 2011 US dollars. These
are direct repair costs and do not include indirect costs due to building downtime as well as flooding due to
piping failure. Although these indirect costs are substantial, they are difficult to accurately calculate for a
general building and were deemed out of scope. The expected total repair costs for the case study building
subjected to a ULS or SLS earthquake are shown in Figure 6. The total repair costs are significant, as the ULS
and SLS mean are $10 million and $4 million respectively. The SLS repair costs are of particular concern.
Damage at this intensity should be limited to a point which does not compromise building serviceability.
However, the assessment results indicate that extensive repairs are required. It is probable that these repairs
will disrupt building function for an extended period of time. This suggests that modern building standards
do not accomplish earthquake resilience, the ability to quickly recover after an earthquake.
Repair Cost ($M)

14

ULS
SLS

12
10
8
6

4
2
0
Mean

10th
Percentile

Median

90th
Percentile

Figure 6 Total repair cost information for the ULS and SLS

Following discussion with professional cost consultants, the total cost of the case study building was
estimated to be approximately $20 million. The mean and 90th percentile ULS repair costs are then
approximately 50% and 65% the cost of a new building respectively. If repair costs exceed 40% of the building
cost, owners often elect to demolish and replace the existing building (ATC 2012). This guideline and the
performance assessment results imply it is probable that buildings designed to current standards may be
demolished and replaced following a ULS earthquake.
Mean repair costs were calculated for each floor in the building, with floor 1 as the ground floor and floor 17
representing the roof. Floor repair costs sorted based on the associated EDP are displayed in Figure 7. IDR
and acceleration comprise the majority of the repair costs for both the ULS and SLS. 53% of the ULS repair
Seismic Performance Assessment of a Code Compliant Multistorey Building

costs are IDR-sensitive and 40% are acceleration-sensitive, while IDR and acceleration each account for 42%
of the SLS repair costs. Conventional building codes control IDR but do not place limitations on acceleration.
Figure 7 shows that acceleration-sensitive damage is comparable to IDR-sensitive damage. This should be
reflected in the design procedures of structures.
1

1
IDR
Acceleration
Velocity

Repair Cost ($M)

0.8
0.7

IDR
Acceleration
Velocity

0.9

0.8
Repair Cost ($M)

0.9

0.6
0.5
0.4
0.3

0.7
0.6
0.5
0.4
0.3

0.2

0.2

0.1

0.1

0
1 2 3 4 5 6 7 8 9 1011121314151617
Floor

1 2 3 4 5 6 7 8 9 1011121314151617
Floor

Figure 7 ULS (left) and SLS (right) mean repair costs on each floor per demand parameter

The EDP sorted costs in Figure 7 have a similar distribution for both the ULS and SLS. Floor 1 has a large value
and the majority is acceleration-sensitive. This is due to a concentration of acceleration-sensitive
nonstructural components located on the ground floor such as heating, ventilating and air conditioning
(HVAC) equipment. Floor 17 exhibits minimal damage as there are few components located on the roof of
the building. The remaining floors are office occupancy and have similar nonstructural and structural
quantities. These floors exhibit a similar level of acceleration-sensitive repair costs. This is expected as the
mean acceleration results from the time history analyses are comparable for floors 2 to 16 as shown in Figure
5. The difference in repair costs at these floors is dominated by IDR-sensitive costs, with the peak at floor 13.
The distribution of IDR-sensitive costs correlates with the mean IDR results from the time history analyses as
shown in Figure 5.
The mean repair costs are arranged by structural and nonstructural fragility groups in Figure 8. For each
fragility group, the repair cost is a function of the number of components in the group, the replacement cost
per component, and the damage state experienced by each component. Fragility groups which noticeably
contribute to repair costs are structural components, glass curtain wall cladding, wall partitions, suspended
ceilings, HVAC equipment, and office electronics and equipment. Negligible repair costs are due to stairs,
access flooring, elevators, piping, the fire sprinkler system and electrical systems.

Repair Cost ($M)

2.5
ULS
SLS

2
1.5
1
0.5
0

Fragility Group

Figure 8 Mean repair costs arranged by fragility group for the ULS and SLS
Seismic Performance Assessment of a Code Compliant Multistorey Building

Considering the ULS, the structural system accounts for 26% of the total repair cost and has the largest repair
cost of the fragility groups. The majority of this cost is due to brace damage. Structural damage is expected
for the ULS, as conventional seismic design relies on the structure sustaining substantial inelastic
deformations. These large inelastic deformations are concentrated in members designed to dissipate seismic
energy in a controlled manner. Although the inelastic deformations cause irreparable damage in the
structural members, the life safety of building occupants is ensured during a major earthquake. Nonstructural
systems are often omitted or treated in an oversimplified manner during building design. However, 74% of
the total repair cost can be attributed to these systems. This demonstrates the need to consider
nonstructural systems during structural design. Attaining a target level of seismic performance mandates
the harmonization of structural and nonstructural performance levels.
Considering the SLS, 87% of the repair costs are due to nonstructural systems. Office equipment such as
desktop electronics is the largest contributor, followed by wall partitions and suspended ceilings. This
indicates that significant repair costs and downtime can be expected in modern multistorey structures
following a serviceability level earthquake. These results highlight the importance of considering
nonstructural systems when designing buildings in order to allow for a rapid return to building occupancy.

7. Eurocode nonstructural provision results

The case study exposed the limitations of the Eurocode 8 damage mitigation methodology. The most
stringent requirement to protect nonstructural systems is a maximum allowable IDR of 1% (CEN 2013). The
case study determined that this IDR limit is inadequate to prevent damage to drift-sensitive nonstructural
systems. Figure 9 demonstrates this finding using fragility functions for a gypsum wall partition. Damage
states one, two and three correspond to minor, moderate, and significant cracking respectively. If the 1%
drift limit is respected, a wall partition is still expected to sustain damage. The probability of exceeding
damage state one is approaching 100%, while damage state two is over 75% and damage state three is over
30%. These results highlight the need for further review of the Eurocode nonstructural provisions.
Probability of Damage

100%
80%
60%
40%

Minor Cracking
Moderate Cracking
Significant Cracking
Eurocode 8 Limit

20%
0%
0.0%

0.5%

1.0%
Interstorey Drift

1.5%

2.0%

Figure 9 Wall partition fragility functions and the Eurocode 8 drift limit for nonstructural protection

8. Conclusions and future work

The seismic performance assessment of the conventional design determined that drift-sensitive and
acceleration-sensitive nonstructural systems experience substantial damage. Significant repair costs and
downtime can therefore be expected in modern multistorey structures following an earthquake. Large repair
costs were observed for both the SLS and ULS scenarios. These direct economic losses are compounded by
large indirect costs due to downtime and disruptions. The performance assessment results imply it is
probable that buildings designed to current standards may be demolished and replaced following a ULS
earthquake due to the high repair costs. The high level of repair costs was particularly concerning for the SLS
scenario, as damage should be limited during this more frequent event. This suggests that modern building
standards do not accomplish resilience, the ability of a community to quickly recover after an earthquake.
Seismic Performance Assessment of a Code Compliant Multistorey Building

Conventional building codes control IDR but do not place limitations on acceleration. It was shown that
acceleration-sensitive damage is comparable to IDR-sensitive damage. This should be reflected in the design
procedures of structures. Limitations of the Eurocode damage mitigation methodology were also revealed
as the prescribed IDR limit did not prevent nonstructural damage. These results highlight the need for design
procedures which enhance nonstructural seismic performance. These changes are required in order to
minimize the societal impacts of earthquakes and accelerate community recovery.
The results of this analysis provide a benchmark on which to compare the effectiveness of alternative designs
in improving nonstructural seismic performance. This benchmark is valuable when evaluating retrofit
alternatives for an existing building or when selecting design options for new structures. Fragility groups
which contribute significantly to repair costs and those with negligible impacts were identified. Additional
studies conducted may only include the significant fragility groups in order to decrease the computational
time needed. The final aim of the project is to develop a viscous fluid damper placement strategy that
minimizes nonstructural repair costs and building service disruptions. It has been shown that there is a need
and opportunity for this solution. Following the completion of the benchmark examination, viscous fluid
dampers will be incorporated into the building model. The capability of viscous fluid dampers to improve
nonstructural seismic performance will be measured by comparing the performance assessment results to
the conventional benchmark results. Damper placement optimization will then be explored. Expectations are
shifting in modern earthquake engineering, as clients are requesting a rapid return to occupancy after an
earthquake event. This research contributes towards meeting these new expectations.

References
Applied Technology Council (ATC) (2012). FEMA P-58 Seismic Performance Assessment of Buildings. Federal
Emergency Management Agency, USA.
Black, R.G., Wenger, W.A., Popov, E. (1980). Inelastic Buckling of Steel Struts Under Cyclic Load Reversal.
Report No. UCB/EERC-80/40, University of California Berkeley, USA.
Charney, F. (2008). Unintended Consequences of Modeling Damping in Structures. J. Struct. Eng., 134 (4),
581-592.
Computers and Structures Inc. (CSI) (2013). SAP2000 V15.2.1. Berkeley, USA.
Del Gobbo, G.M. (2014). Improving the Seismic Performance of Nonstructural Systems using Passive Energy
Dissipation Devices. Transfer of Status Report, University of Oxford, UK.
Dhakal, R.P. (2010). Damage to Non-Structural Components and Contents in 2010 Darfield Earthquake. Bull.
New Zeal. Soc. Earthq. Eng., 43 (4), 404-411.
European Committee for Standardization (CEN) (2010). Eurocode Basis of structural design. Belgium.
European Committee for Standardization (CEN) (2013). Eurocode 8: Design of structures for earthquake
resistance Part 1: General rules, seismic actions and rules for buildings. Belgium.
Fierro, E.A., Miranda, E., Perry, C.L. (2011). Behavior of Nonstructural Components in Recent Earthquakes.
2011 Architectural Engineering Conference, Oakland - USA, 369-377.
Miranda, E., Mosqueda, G., Retamales, R., Pekcan, G. (2012). Performance of Nonstructural Components
during the 27 February 2010 Chile Earthquake. Earthq. Spectra, 28 (S1), 453-471.
Pacific Earthquake Engineering Research Center (PEER) (2013), PEER NGA-WEST 2 Ground Motion
Database, http://ngawest2.berkeley.edu/site, accessed in February 25, 2016.
Pacific Earthquake Engineering Research Center (PEER) (2015). OpenSees V2.4.6. University of California
Berkeley, USA.
Taghavi, S., Miranda, E. (2003). Response Assessment of Nonstructural Building Elements. PEER Report
2003/05, University of California Berkeley, USA.
Uriz, P., Filippou, F.C., Mahin, S.A. (2008). Model for Cyclic Inelastic Buckling of Steel Braces. J. Struct. Eng.,
134 (4), 619-628.
Seismic Performance Assessment of a Code Compliant Multistorey Building