You are on page 1of 24

THE STRUCTURAL DESIGN OF TALL AND SPECIAL BUILDINGS Struct. Design Tall Spec. Build.

20, 223246 (2011) Published online 16 July 2009 in Wiley Online Library (wileyonlinelibrary.com/journal/tal). DOI: 10.1002/tal.531

A case study of the structural responses of a tall building in Singapore subjected to close-in detonations
Bing Li*,, Tso-Chien Pan and Anand Nair
Protective Technology Research Center, Nanyang Technological University, Singapore

SUMMARY The response of tall buildings has been a major concern in metropolitan cities, especially with the recent surge in extreme activities targeted at structures with viable commercial values. This paper discusses a study carried out on the structural behaviour of a 2D frame, modelled to represent a tall building with ABAQUS. The model frame was subjected to a charge of the equivalent weight of 1 ton of TNT but placed at two varying cases of 5 and 10-m standoff distances. Plane-strain elements that incorporate the feature of material nonlinearity were utilized to model the structural components of the building and the simulated blast overpressures were obtained from the CONWEP software. The effects of large deformations of beams and columns corresponding to the short time loading duration depicted by the explosions were analysed from a local perspective. The extent of the damage is based on a local index dened as the ratio of curvatures. These local indices are consequently used to determine the possibility of disproportionate collapse of the frame from a global perspective. Finally, the provision of more ductile structural detailing is recommended to enhance the structural integrity of the building, increasing its resilience against blast attacks. Copyright 2009 John Wiley & Sons, Ltd.

1. INTRODUCTION Over the past 20 years, accidental or deliberate explosive incidents have revealed the vulnerability of civilian structures to the extreme dynamic loading caused by detonations from short standoff distances (Corley et al., 1998). These types of incidents are characterized by its great intensity and short duration. The collapse of the World Trade Center in New York City has brought the focus of many structural engineers to develop an understanding of the behaviour and resistance of structures when subjected to blast loadings. This unfortunate event has brought about an urgent need for an increased awareness in the behaviour of public buildings, especially commercial high-rise buildings, when placed within a blast environment. In this study, two case scenarios were created to understand the dynamic behaviour of a 30-storey reinforced concrete structure located in Singapore. The two cases involved a similar frame structure modelled to represent an actual 30-storey building and was subjected to a charge of the equivalent weight of 1 ton of TNT but placed at two case scenarios of 5 and 10-m standoff distances. The study was carried out to determine the damage levels sustained by the structure as well as the post-blast behaviour of the damaged building under gravity loading. A conclusion of the vulnerability of this typical building when subjected to such blast incidents is provided based on the ndings from the abovementioned analysis. Buildings in Singapore are normally designed for the onerous combination of wind and gravity loads. The inherent resistance of Singapore buildings is therefore derived from these design loads and the associated structural detailing. The results of this study have shown that RC buildings in Singapore would experience demands on strength and ductility that are greater than their inherent strength capacities and ductility factors for blast events that occur as close as from 10-m standoff distances. Therefore, in such cases, the potential damage to the structure and consequent loss of lives may be devastating.
* Correspondence to: Bing Li, Nanyang Technological University, Singapore 639798 E-mail: cbli@ntu.edu.sg Copyright 2009 John Wiley & Sons, Ltd.

224

B. LI, T-C. PAN AND A. NAIR

Substantial resiliency of a structural system is required in order to sustain the blast-induced load and displacement. The required resiliency can best be achieved via the continuity, redundancy and energy-absorbing capacity within the structural system and its components. This therefore points to a need for a systematic review of the provisions within the current building regulations and codes with respect to lateral load requirements for both wind, earthquake and re effects, which will collectively provide an inherent resiliency against the progressive collapse of a building structure when subjected to blast loads.

2. STRUCTURAL DESCRIPTION The model frame was to represent a 30-storey prestressed reinforced concrete structure with a main tower standing at the height of 120 m and was built in accordance to the Structural Use of Concrete design code (BS 8110, 1997). The structural system of the building was made up of frames with a core wall to provide lateral and vertical strength. Vertical loads (dead and live loads) were transferred to the columns and in turn to the ground via post-tensioned prestressed beams. The rst 10 story heights vary from 3 to 8.78 m and their design live loads vary from 2.5 to 5 kPa. Beyond the 11th storey, a standard storey height of 4.2 m was utilized in the model and their corresponding design live loads varied from 3.5 to 5 kPa. The dead loads for all the stories were computed based on the material density and geometries of the actual building and its components. The column dimensions on the rst three stories of the building vary from 1800 1200 to 2500 1150 mm. The column on the right side of the selected bay was slightly larger and modelled to be 2500 1150 mm. The column on the left bay was modelled as 1800 1200 mm. The typical beam dimensions are 1200 550 mm. As the effects of blast loading was not accounted for during the structural design of the actual building, an accurate assessment of the behaviour of the structure when placed within a blast environment is imperative. A 2D frame model was constructed to represent this actual building through the ABAQUS (Hibbit, Karlsson & Sorensen, Inc., 2001) software. Nonlinear nite element method (FEM) analysis was carried out on this model to gain an overview of the blast resistance of the actual building. The modelled frame is a two-bay structure with a core wall in the middle. The width of the web and two anges used to model the core wall is 5.4 and 0.4 m, respectively. The bays span 12.6 and 10.75 m, respectively, from left to right. An illustration of the modelled frame is depicted in Figure 1. Following the nonlinear nite element dynamic analysis, a damage index determination and post-blast behaviour analysis is carried out on the modeled frame structure.

3. BLAST LOADING MODELLING Blast loading is a type of extremely dynamic action on a structure that applies a great intensity of force over a short duration. It was noted that the variation of blast pressure on structures with respect to time can be modelled as an exponential function of time by Hyde (1991). The parameters required for this function to model the blast phenomenon are the peak value of the blast pressure and the duration of the blast. A scalar variable of scale distance, which is related to the equivalent weight of the explosive and the standoff distance correlate these two parameters as described by Hyde (1991). Theoretically, the weight of explosive could vary from relatively small to an extremely large amount. A larger weight of explosive would induce a larger blast peak pressure over a longer duration. This would result in more severe damage of the building. However, the weight of explosive that can be obtained in any particular region is limited. Reviews carried out on previous blast incidents that have occurred in recent years illustrate that most of the destructive deliberate terrorist attacks on public buildings were cases with moderate weight of explosives placed at short (<10 m) standoff distances. Based on the damage assessment carried out on those damaged buildings, it was found that explosions caused by moderate quantities of explosives placed at short standoff distances caused various damage to individual structural components as well as damage to structures globally. Blast resistance of a structure can be determined by studying the behaviour of the building when subjected to a typical value of explosive weight placed at various standoff distances. The variation in standoff
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

RESPONSES TO DETONATED BUILDING IN SINGAPORE

225

Figure 1. Elevation of frame model.

distances allows blast effects on structures at various peak blast pressures and durations to be determined. With these considerations, this study was carried out with an equivalent weight of 1 ton of TNT placed at two varied cases of standoff distances of 5 and 10 m. The function of blast pressure varying with time is generally reduced to be a triangular pulse to simplify the nonlinear inelastic structural dynamic response problem. The software CONWEP (Hyde, 1991), which is generally used in blast resistant design, was utilized to determine the value of the peak blast pressure and its respective duration for the two selected case scenarios. The distribution of blast pressure and its duration across the structural surface is plot in Figures 2 and 3. The variation of blast pressure with respect to time at ground level for both cases is plotted in Figures 4 and 5, where the blast pressure is denoted by P and the corresponding duration by t. It is evident that with the increase in standoff distance from 5 to 10 m, there is a drop in peak blast pressure from 39 640 to 8106 kPa. In contrast, the corresponding blast duration increases from 2.81 to 17.37 ms.

4. FEM MODELLING Nonlinear dynamic nite element analysis is needed to obtain actual structural dynamic response under blast conditions. The process of the analysis includes material modelling, corresponding parameters selection, damping calculation, boundary condition description, analytical step determination and loading denition. A cap plasticity model featured in ABAQUS (Hibbit, Karlsson & Sorensen, Inc., 2001) was selected to simulate the behaviour of concrete within the model for this study. An elastic perfectly plastic model, which is generally accepted to simulate reinforcing steel, was selected as it provides sufcient accuracy for the reinforced concrete structure analysis. The nite element code and material models have been veried with a simply supported RC beam and a slab subjected to blast loadings, and the numerically determined responses were similar to those obtained experimentally (Rong, 2005; Rong and Li, 2007).
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

226

B. LI, T-C. PAN AND A. NAIR

Figure 2. Blast loading peak pressure and loading duration of one-ton equivalent TNT placed at 10-m standoff distance.

Figure 3. Blast loading peak pressure and loading duration of one-ton equivalent TNT placed at ve-m standoff distance.

P (kPa) 8,106

17.37

t (ms)

Figure 4. Time history of blast pressure with one-ton equivalent TNT at 10-m standoff distance.
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

RESPONSES TO DETONATED BUILDING IN SINGAPORE

227

P (kPa) 39,640

2.81

t (ms)

Figure 5. Time history of blast pressure with one-ton equivalent TNT at ve-m standoff distance.

Table 1. Concrete and steel material properties.


Concrete properties Elastic modulus Ec Ultimate compression strength fc Poissons ratio m Friction angle J Material cohesion Cap eccentricity parameter Initial cap yield surface Transition surface radius Parameter K Density rc Steel properties Elastic modulus Es Yield stress fy Ultimate stress fu Ultimate strain eu Mass density rs 29 200 MPa 30.5 MPa 0.2 30 10.25 MPa 0.23 0.02 0.5 1 2500 kg/m3 210 000 MPa 460 MPa 575 MPa 0.01 7800 kg/m3

The concrete and steel properties are listed in Table 1. The stressstrain curve is assumed to be bilinear, representing elasto-plastic behaviour with linear isotropic hardening. The expressions proposed by Malvar and Crawford (1998) as illustrated in Figure 6 were utilized to obtain the strain rate sensitivity due to dynamic loads. An increase of 20% in the strength and modulus of elasticity was incorporated accordingly to allow for the effects of strain rate. The response of structures when subjected to blast loading is highly dependent on damping. Rayleigh damping as dened by Clough and Penzien (1993) was utilized to determine the damping parameters for this study. The rst and second circular frequencies were respectively determined to be 2.0545 and 6.6865 rad/s through mode analysis of the actual frame structure. Subsequently, the two damping parameters were calculated to be 0.157 and 0.01144 by respectively substituting the circular frequencies into the Rayleigh damping expression. The bottom ends of the columns and walls of the actual building are perfectly xed to the ground. A 2D solid continuum element with four nodes dened as CPE4R in ABAQUS (Hibbit, Karlsson & Sorensen, Inc., 2001), was selected to mesh the frame concrete. As it would be too complicated to model each discrete bar of reinforcing steel for the 30-storey building, smeared rebar elements were modelled in accordance with the actual reinforcement ratio used in the building. There were two stages considered during the analysis of the structures response to blast loadings. The service loads were initially imposed on to the structure prior to the occurrence of the explosion. The intensity and distribution of the stresses and strains induced by these service loads will denitely
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

228

B. LI, T-C. PAN AND A. NAIR

Figure 6. DIF for reinforcements (Malvar and Crawford, 1998).

inuence structural behaviour under blast conditions. Thus, static nonlinear nite element analysis was initially performed to assess the structure. The loads imposed on to the structure up to this stage were the live, dead and prestressed loads. Live loads of 5 kPa were imposed on storeys 110. Subsequent oors were imposed with a live load of 6 kPa. Dead loads were dened as 5 kPa for all storeys. Two types of equivalent forces were utilized to represent the effect of prestressing in the reinforcement bars due to the difculty in incorporating the effect of the prestressed steels directly into the structures nite element model. The balance force, uniformly distributed along the prestressed member, and a compressive axial pressure at both ends of the member would represent the effect of prestressing. The balance force is determined in accordance with the following equation. q= 8 Pe L2 (1)

where, q is the balance force, e is the height of string, P is the applied pretension force and L is the length between contra-exure points. The second stage of the analysis involved determining the dynamic response of the structure when the blast loading was induced. In this stage, dynamic nite element analysis of the structural response was required.

5. DYNAMIC RESPONSE 5.1. Case 1: 10-m standoff distance The distribution of the structural deformation with time when an explosion with an equivalent weight of 1 ton of TNT placed at a standoff distance of 10 m based on the FEM analysis is illustrated in Figure 7. The rst plot in Figure 7 shows the initial static stress experienced by the structure when subjected to the dead load, live load and balance force of the pretension. The following pictures in the same gure illustrate the displacement response of the structure at selected time steps when it was subjected to the blast loading. It was observed that at the time step of 13.4 ms, the column at the bottom experienced a maximum displacement of 32.7 mm. Plastic deformation is mainly concentrated on the bottom column and at the beams of the second and third storeys. However, the occurrence of this
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

RESPONSES TO DETONATED BUILDING IN SINGAPORE

229

Figure 7. Deformation contour at various time steps of case 1 (10-m standoff).

plastic deformation was not simultaneous. Severe damage in the column occurred at approximately 13 ms, while the peak damage in the beams occurred slightly later at 16 ms. Structural response of buildings when subjected to blast loadings can be generally divided into two phases. They are the forced and free vibration phases. The former occurs during the time of blast impulse. After the end of the blast duration, no external dynamic force is applied onto the structure
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

230

B. LI, T-C. PAN AND A. NAIR

Figure 7. Continued

and it enters into a free vibration phase. The approximate blast duration for this case scenario was 20 ms. It was noted that during the forced vibration phase of this case scenario, the deformations were localized on the blast-loaded elements. The left bottom column reaching its maximum deformation at a time of 13 ms, which is within the approximate blast duration, illustrates this phenomenon. As the structure exits from the blast duration, its global response becomes more signicant. This is evident
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

RESPONSES TO DETONATED BUILDING IN SINGAPORE

231

as the deformations due to the blast loadings propagate up the structure. The higher oors of the structure still experienced deformations of up to 30 mm at a time step of 100 ms. This exhibits the global response of the structure in its free vibration phase. Even at a time step of 200 ms, there were still deformations in the higher oors that were up to 16 mm. The core wall plays an important role in the global response of the structure in this case scenario. At a time of 30 ms, the core wall experienced an obvious horizontal vibration due to the force transferred to it through the beams. As time progressed in the analysis, the motion began to propagate up in a vertical direction. The vibration only reached the top storey at approximately 200 ms during the analysis. The vibration of the core wall became evident mostly in the free vibration phase. The core wall experienced a maximum displacement of 32.9 mm at around 300 ms. This maximum displacement was located at a height close to the 10th story of the building. 5.2. Case 2: Five-m standoff distance The distribution of the structural deformation with time when an explosion with an equivalent weight of 1 ton of TNT placed at a standoff distance of 5 m based on the FEM analysis is illustrated in Figure 8. It was clearly illustrated from the analysis results shown in Figure 8 that the dynamic deformation response of the frame when the explosive charge is placed at a standoff distance of 5 m is more localized in contrast to the rst case and concentrated mostly at the bottom column. The mid-height of the second storey column reached a peak lateral displacement value of 87.2 mm at a time of 13.3 ms. The core wall attained its maximum displacement of 15.1 mm at a time of approximately 100 ms. This maximum displacement of the core wall was located near the location of the blast loading. There was little displacement in the higher stories of the core wall as the structure displayed minimal global response from this case scenario. It is also evident from the analysis that while large residual deformation was present in some structural members, the recorded global structural dynamic response was not as extreme as compared with the rst case scenario. This is because after the end of the blast loading duration, the structural free vibration phase is more obscure in the second case as compared with the rst case. This is caused by the shorter loading duration in the second case scenario as compared with the 10-m standoff case. The second case scenario had a blast loading duration of 2.8 ms, during which, the column on the bottom oor experienced high blast pressure. This short loading duration creates an impulsive force on the column, giving it little time to respond to this pulse loading and not allowing it to spread to stresses out to the rest of the frame. This behaviour of frame structures when subjected to distant blast conditions is developed from the study carried out by Li et al. (2006). This reduced global response under the higher load could have been enhanced further due to the increased local damage to members that prevented the entire mass to be excited. The peak deformation of the column at the bottom occurs almost at the same time step for both case scenarios. This time step is located at the blast duration phase. However, for the second case scenario, the global structural dynamic response in the free vibration phase is much smaller than the 10-m standoff case. This is displayed by comparing the global response of the structure at various time steps. The core wall on the higher stories for this case scenario had a displacement of 15 mm in contrast to a displacement of 31.6 mm at the same location and time step of the previous case scenario. This phenomenon is conrmed further when comparing the displacement at a later time step. At a time step of 200 ms, the core wall displacement at the higher stories had attained 5 mm in this case scenario as compared with a displacement of 16 mm in the previous case at the same location. 6. BLAST DAMAGE ASSESSMENT The performance of exural structural elements is evaluated with a local damage index based on the curvature of the element. The index is dened as DI =
Copyright 2009 John Wiley & Sons, Ltd.

m u

(2)
Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

232

B. LI, T-C. PAN AND A. NAIR

Figure 8. Deformation contour at various time steps of case 1 (ve-m standoff). where, fm, is the maximum curvature demand for the reinforced concrete cross section under blast conditions and fu is the curvature capacity, which can be obtained using the reinforced concrete section analysis software, RESPONSE (Bentz, 2000). Table 2 lists out the various damage levels that correspond with deformation limits of typical reinforced concrete exural components. The damage index of any member should not exceed a maximum value of one.
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

RESPONSES TO DETONATED BUILDING IN SINGAPORE

233

Figure 8. Continued

6.1. Identication of structural components The structural elements of the frame were given each their own identication tag (ID) in order to locate and identify the damage indices corresponding to individual elements. The rightmost column of the frame is labelled as axis A, while the leftmost column is labelled as axis D. In accordance with
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

234

B. LI, T-C. PAN AND A. NAIR

Table 2. Blast damage levels based on curvature (Clough and Penzien, 1993). Damage indices, DI
0 00.2 0.20.4 0.40.6 0.61.0 >1.0

Consequence
Elastic behavior, no damage Slight/minor damage Slight to moderate damage Severe damage Incipient or partial collapse Complete collapse or destruction

30

20

Left zone

Right zone

10

C B

Figure 9. Identication tags for the frame used in the analysis.

this scheme of labelling, the core wall is labelled within axis C and B. The frame is then accordingly divided into two zones, the right and left zones. The illustration in Figure 9 shows the respective zones labelled in the frame. The beams are numbered as CDi or ABi, where, CDi represents the beam located at the ith storey level in the left zone of the frame and ABi represents a beam located at the ith storey but in the right zone. The columns are numbered as Dij or Aij, where the characters A and D display the axial position of the column and the subscripts i and j represent the corresponding storey levels that the column element connects. Lastly, a group of beams labelled as ABij or CDij represents all beams in the respective zone from the ith to jth storey. 6.2. Damage assessment of case scenario 1 (10-m standoff distance) The curvature of selected beams from their right and left cross-sections are shown in Figure 10. Most of the structural damage was concentrated on the left zone of the frame coinciding with the location of the charge. Beam CD2 was calculated to have damage indices of 0.67 and 1.0 at its left and right ends, respectively, with the aid of Equation (2). This indicates that the right end of beam CD2 was destroyed by the blast. In addition, the left end of the beam was in a state of incipient collapse. It was evident from the analysis that the beams located on stories one to six of the left zone of the frame sustained a variety of damages. Beam CD3 was almost at complete failure due to the blast. The extent of damage induced on beams with the progression of stories was gradually reducing. No signicant damage was recorded for beams located above the sixth storey. However, there was some
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

RESPONSES TO DETONATED BUILDING IN SINGAPORE

235

Beam CD-2 (left end section)


0.08

0.14

Beam CD-2 (right end section)


0.12

0.07

0.10

curvature (rad/m)

0.06 0.05 0.04 0.03 0.02 0.01 0.00 0.00 0.05 0.10 0.15 0.20 0.25 0.30
curvature demand curvature for initial damage curvature for complete failure

0.08 0.06 0.04 0.02 0.00 -0.02 0.00


curvature demand curvature for initial damage curvature for complete failure

curvature (rad/m)

0.05

0.10

0.15

0.20

0.25

0.30

Time (s)

Time (s)

DI = 0.67 (Incipient Collapse)


0.12

DI = 1.0 (Complete Failure)


Beam CD-4 (left end section)
0.11 0.10 0.09

Beam CD-3 (left end section)

0.10

curvature (rad/m)

0.08

0.08

0.06

curvature (rad/m)

curvature demand curvature for initial damage curvature for complete failure

0.07 0.06 0.05 0.04 0.03 0.02 0.01

curvature demand curvature for initial damage curvature for complete failure

0.04

0.02

0.00
0.00

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Time (s)

-0.01 0.00

0.05

0.10

0.15

0.20

0.25

0.30

Time (s)

DI = 1.0 (Complete Failure)


0.008

DI = 0.41 (Severe Damage)


Beam CD-30 (left end section)

Beam CD-5 (left end section)


0.10

0.006
0.08

curvature (rad/m)

curvature (rad/m)

0.004

0.06

curvature demand curvature for initial damage curvature for complete failure

curvature demand curvature for initial damage

0.002

0.04

0.02

0.000

0.00 0.00 0.05 0.10 0.15 0.20 0.25 0.30

-0.002 0.00

0.05

0.10

0.15

0.20

0.25

0.30

Time (s)

Time (s)

DI = 0.26 (Moderate Damage)

DI = 0 (No Damage)

Figure 10. Damage indices of selected beams of case scenario 1 (10-m standoff).

damage at the beam located at the left zones top storey. This could be due to the constraint of the beam at the top storey being much lesser as compared with the beams at the other oors and thus causing the vibration at the top storey to be much greater and eventually resulting in it accumulating some damage. Column C12, being the closest to the detonation center, is the most affected by the blast pressure and thus resulting in this particular column accumulating the most damage. Damage indices of selected upper and bottom sections of columns in the frame are provided in Figure 11. At approximately 10 ms,
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

236

B. LI, T-C. PAN AND A. NAIR

0.05

Column D-1-2 (bottom section)

0.05

Column D-1-2 (top section)

0.04

0.04
0.03

curvature (rad/m)

curvature (rad/m)

0.02

curvature demand curvature for initial damage curvature for complete failure

0.03

curvature demand curvature for initial damage curvature for complete failure

0.01

0.02

0.00

0.01
-0.01

0.00
-0.02 0.00 0.05 0.10 0.15 0.20 0.25 0.30

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Time (s)

Time (s)

DI = 0.69 (Incipient Collapse)


0.05

DI = 0.14 (Slight Damage)


0.05

Column D-2-3 (bottom section)

Column D-2-3 (top section)

0.04

0.04

curvature (rad/m)

0.03

curvature demand curvature for initial damage curvature for complete failure

curvature (rad/m)

0.03

curvature demand curvature for initial damage curvature for complete failure

0.02

0.02

0.01

0.01

0.00 0.00 0.05 0.10 0.15 0.20 0.25 0.30

Time (s)

0.00 0.00

0.05

0.10

0.15

0.20

0.25

0.30

Time (s)

DI = 0

DI = 0

Figure 11. Damage indices of selected columns of case scenario 1 (10-m standoff).

column C12 attained its maximum curvature of 0.037 rad/m. With the progression of time in the analysis, the curvature reached its negative peak at a time step of 25 ms. Despite these high curvatures, only slight damage appeared in the upper section of this column. This could be due to the rapid reduction of the blast pressure. It was also observed from the analysis that almost no obvious damage occurs to column C23. Through the exural damage assessment carried out on the columns, it can be concluded that only column C12 was destroyed severely while other columns in the frame sustained no signicant damage. Figure 12 generalizes the damage of the whole frame when subjected to the blast loading specied in case scenario 1. It can be seen that the second and third storey beams have completely failed while column C12 and the beams in stories four and ve sustained severe damage. In addition, the beam at the top storey was slightly damaged due to the blast loadings. 6.3. Damage assessment of case scenario 2 (ve-m standoff distance) The curvature of selected beams from their right and left cross-sections are shown in Figure 13. The beams on the right zone of the building have no signicant damage in this case scenario. These beams have their maximum curvatures less than the corresponding curvatures for the initial damage. This could be due to the impulsive force not being transferred to the right zone of the frame because of the shorter duration of the blast loading in this case scenario. The loaded height by the relatively high
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

RESPONSES TO DETONATED BUILDING IN SINGAPORE

237

Slight damage

Severe damage Failure


Figure 12. Damage assessment of the frame subjected to the blast loading in case scenario 1 (10-m standoff).

pressure from the detonation center was about 10 m. This coincides to about the same point as the fth storey of the frame. As such, there is no signicant damage on the beams of the frame located beyond the fth storey onwards. Complete failure occurs to beams CD2 and CD3 as both their damage indices were at a maximum value of one. The exural damage becomes less severe for beam CD4. This is because the local response of the structure dominates when loaded by this case scenario while the global vibration is not as prominent. Damage indices of selected upper and bottom sections of columns in the frame subjected to the loading specied in case scenario 2 are provided in Figure 14. The columns located above the fourth storey level experience almost no exural damage when subjected to the blast loading. Incipient collapse occurs to column D12, which is severely loaded by the blast pressure directly. Severe damage occurs to column D23. As the blast forces are decreasing along the height of the building, the damage to columns also obviously diminishes along the same order. Figure 15 generalizes the damage subjected to the frame from the analysis carried out and illustrates the partial collapse, severe damage and complete failure that occurred to individual elements. In comparison with the damage assessment summary with case scenario 1 as illustrated in Figure 12, it can be seen that the damage for the second case scenario is more localized and the frame in its top storey beam sustains no damage.

7. POST-BLAST ANALYSIS OF DAMAGED FRAME The previous sections have discussed the dynamic response of the modelled frame structure in the transient blast loading conditions. The damage assessment on the frame indicates that some of the elements in the frame are severely damaged while some others are in a state of complete failure. The next question of interest would be to determine if this damaged state of the frame would lead to its progressive collapse. Post-blast analysis was carried out on the damaged frame structure to evaluate its capacity to prevent progressive collapse following the blast damage. To simplify the process of this evaluation, the following assumptions were taken:
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

238

B. LI, T-C. PAN AND A. NAIR

0.12

Beam CD-2 (left end section)

0.16 0.14 0.12

Beam CD-2 (right end section)

0.10

0.08

curvature (rad/m)

0.06
curvature demand curvature for initial damage curvature for complete failure

curvature (rad/m)

0.10 0.08 0.06 0.04 0.02


curvature demand curvature for initial damage curvature for complete failure

0.04

0.02

0.00

0.00 -0.02 0.00

-0.02 0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.05

0.10

0.15

0.20

0.25

0.30

Time (s)

Time (s)

DI = 1.0 (Complete Failure)


0.16
0.08

DI = 1.0 (Complete Failure)


Beam CD-3 (right end section)

0.14 0.12

Beam CD-3 (left end section)


0.06
curvature demand curvature for initial damage curvature for complete failure

curvature (rad/m)

0.10 0.08 0.06 0.04 0.02


-0.02
curvature demand curvature for initial damage curvature for complete failure

curvature (rad/m)

0.04

0.02

0.00

0.00
0.00 0.05 0.10 0.15 0.20 0.25 0.30

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Time (s)

Time (S)

DI = 1.0 (Complete Failure)

DI = 0.8 (Incipient Collapse)

Figure 13. Damage indices of selected beams of case scenario 2 (ve-m standoff).

Any element will progressively lose its strength as the level of damage increases up until the point of its failure. Any structural element can be substituted by statically applying its internal force to the supported cross section of the remaining structure in the opposite direction. An elements collapse can be simulated by dynamically applying an additional force to the supported cross section of the remaining structure, with a magnitude equal to that elements internal force.

7.1. Progressive collapse assessment of case scenario 1 (10-m standoff distance) The analysis was further simplied by assuming that elements that have failed completely due to the blast loads were not included in this assessment and were not taken to contribute to the frame resisting the service loading. In addition, the strength of structural components whose damage indices are less than 1.0 were taken to decrease their indices linearly as the loading was increased. These abovementioned assumptions were incorporated into case scenario 1, by removing the beams at the lower two storeys of the frame. These beams had damage indices equal to 1.0. In addition, the strength of the blast loaded bottom column was decreased by 70%.
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

RESPONSES TO DETONATED BUILDING IN SINGAPORE

239

Beam CD-4 (left end section)


0.08
0.10

Beam CD-4 (right end section)

0.07 0.06

0.08

curvature (rad/m)

curvature (rad/m)

0.06

curvature demand curvature for initial damage curvature for complete failure

0.05 0.04 0.03 0.02


curvature demand curvature for initial damage curvature for complete failure

0.04

0.02

0.01
0.00 0.00 0.05 0.10 0.15 0.20 0.25 0.30

0.00 0.00 0.05 0.10 0.15 0.20 0.25 0.30

Time (s)

Time (s)

DI = 0 (No Damage)
0.008 0.006 0.004

DI = 0 (No Damage)
0.008 0.006 0.004

Curvature demands of beams cd-6 ~ cd-30


curvature for initial damage

Curvature demands of beams ab-2 ~ ab-30


curvature for initial damage

curvature (rad/m)

0.002 0.000 -0.002 -0.004 -0.006 -0.008 0.00


curvature for initial damage

curvature (rad/m)

0.002 0.000 -0.002 -0.004 -0.006 -0.008 0.00

curvature for initial damage

0.05

0.10

0.15

0.20

0.25

0.30

0.05

0.10

0.15

0.20

0.25

0.30

Time (s)

Time (s)

DI = 0 (No Damage)

DI = 0 (No Damage)

Figure 13. Continued.

Nonlinear static analysis was carried out with aid from the ABAQUS (Hibbit, Karlsson & Sorensen, Inc., 2001) software to determine the stress distribution and deformation of the damaged building and is illustrated in Figure 16. The analysis results showed that the maximum compression stress at the bottom of the core wall reached a value of 20.2 MPa. In contrast, the concrete compression stress at the bottom of the damaged column reached a compression stress of 8.5 MPa. This value is below the remaining ultimate strength of the column. As such, it can be concluded that the damaged structure can withstand its service load if no further exural damage induced by the service loads appear in the element. The analysis provided that the additional lateral displacement due to the blast loading of the bottom column is nearly equal to zero (0.7 mm). Therefore, it can be concluded that after the impact from the blast loading, the frame can sustain the gravity loading without further damage accumulation. 7.2. Progressive collapse assessment of case scenario 2 (ve-m standoff distance) The equivalent frame system utilized for the progressive collapse assessment of case scenario 2, based on the assumptions made in the previous subsection is provided in Figure 17. The computation procedure utilized to carry out the dynamic simulation of strength losses of the removed columns is as follows:
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

240

B. LI, T-C. PAN AND A. NAIR

0.05

Column D-1-2 (bottom section)

0.05

Column D-1-2 (top section)

0.04

0.04
0.03
curvature demand curvature for initial damage curvature for complete failure

curvature (rad/m)

0.02

curvature (rad/m)

0.03

curvature demand curvature for initial damage curvature for complete failure

0.01

0.02

0.00

0.01

-0.01

0.00
-0.02 0.00 0.05 0.10 0.15 0.20 0.25 0.30

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Time (s)

Time (s)

DI = 0.93 (Partial Collapse)


0.05

DI = 0.65 (Partial Collapse)


0.05

Column D-2-3 (bottom section)

Column D-2-3 (top section)

0.04

0.04

curvature (rad/m)

0.03
curvature demand curvature for initial damage curvature for complete failure

curvature (rad/m)

0.03

curvature demand curvature for initial damage curvature for complete failure

0.02

0.02

0.01

0.01

0.00 0.00 0.05 0.10 0.15 0.20 0.25 0.30

0.00 0.00 0.05 0.10 0.15 0.20 0.25 0.30

Time (s)

Time (s)

DI = 0.33 (Moderate Damage)

DI = 0.42 (Severe Damage)

Figure 14. Damage indices of selected columns of case scenario 2 (ve-m standoff).

The static force, Ps under the service loading is computed by nonlinear static analysis with the ABAQUS (Hibbit, Karlsson & Sorensen, Inc., 2001). Ps is then applied upwards to the remaining bottom column of the equivalent frame system initially. This force was determined to be 8155 kN for the left zone of the frame at column D23 and 8735 kN for the right zone of the frame at column A12. A dynamic loading, Pd as dened in Figure 18 is applied downwards to the structure. This loading reaches its peak at a time of t1 and subsequently remains constant. The peak value is equal to the static internal load, Ps. From the inelastic transient nite element analysis, time t1, corresponding to the failure of the removed elements is determined. For case scenario 2, t1 is approximately 13 ms as indicated in the damage assessment section. The results from the post-blast analysis carried out on case scenario 2 based on the abovementioned procedures are shown in Figure 19. The results indicate that with the progression of time, the vertical downward deformation of the cantilever part of the frame gradually increases. The vertical downward displacement at the left tip of the frame reaches 240 mm at a time step of 300 ms. Plastic hinges formed at both ends of all the left beams and they eventually contributed to the progressive collapse mechanism. The frame is expected to experience much larger deformations if the analysis was allowed
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

RESPONSES TO DETONATED BUILDING IN SINGAPORE

241

Partial collapse zones

Failure components

Figure 15. Damage assessment of the frame subjected to the blast loading in case scenario 2 (ve-m standoff).

to proceed further. Thus, it can be deduced that, for case scenario 2, the possibility of progressive collapse occurring in the frame is highly likely, as the blast-loaded columns in the bottom have lost their strengths.

8. CONCLUSION This paper outlined the dynamic structural behaviour of a modelled frame structure subjected to two case scenarios of an equivalent of 1 ton of TNT charge placed at a standoff distance of 10 and 5 m, respectively. This was followed by a damage assessment of the individual elements of the modelled frame and a post-blast behaviour study of the damaged frame structure when subjected to service loadings. The obvious difference between the structural responses between the two case scenarios would have to be the perspective at which the damage occurred for the frame structure. Case scenario 1 (10-m standoff) resulted in the frame adopting a more global response. In contrast, case scenario 2 (ve-m standoff) resulted in the frame adopting a more local response. This difference is mainly due to the blast loading distribution. In case scenario 2, the action range of the blast forces covers about 10 storeys (approximately 35 m). However, the scope of the blast loading for case scenario 2 was limited to only the rst storey (approximately 8 m). This was mainly induced by the difference in the blast loading duration between the two case scenarios. Case scenario 2 had comparatively shorter blast duration than case scenario 1. This resulted in the system in case scenario 2 not having sufcient time for the frame to respond before the blast loading was reduced to zero. Damage assessments carried out demonstrated that in case scenario 1 (10-m standoff), only column D12 was severely damaged. The second and third storey beams were also nearly in complete failure state and the fourth and fth storey beams were severely damaged. However, about three columns located in the rst, second and third storeys experienced signicant damage for case scenario 2 (ve-m standoff). In addition, the second and third storey beams were also close to failure. The differences in response between the two case scenarios are mainly due to global and local perspective taken by the frame structure when subjected to the blast loading.
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

242

B. LI, T-C. PAN AND A. NAIR

Figure 16. Stress distribution and deformation of the damaged building under the service loadings.

Post-blast analysis conducted on the frame structures by removing completely destroyed elements from the model indicated that the frame subjected to case scenario 2 (ve-m standoff) is likely to progressively collapse. In contrast, the frame subjected to the blast loading conditions of case scenario 1 (10 m standoff) is likely to be able to resist the service loads requirement and not lead to a progressive collapse.
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

RESPONSES TO DETONATED BUILDING IN SINGAPORE

243

Figure 17. Equivalent frame system utilized to simulate blast damaged case scenario 2.

Pd P1

t1

Figure 18. A dynamic loading to simulate the strength loss of elements removed.

The results have also shown that while the moment resistance capacity appears to be sufcient under the blast event at a 10 m standoff distance, the shear resistance of the non-ductile section that has been detailed according to current practices has been shown to be decient. The problem of shear resistance is most signicant in the columns and beam column joints of building structures located in Singapore. Therefore, it is desirable to introduce ductile detailing as a potential improvement to the BS 8110 design code, which does not seem to provide sufcient shear resistant capacity and ductility to the structures subject to the blast events that occurs at a standoff distance of 10 m. It is also desirable to develop procedures for strengthening and retrotting that can enhance the shear resistance and ductility of columns, beams and joints. These procedures can rst be applied to strengthen and retrot perimeter columns that are exposed to the possibility of terrorist bomb placements, especially at locations with vehicle access.
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

244

B. LI, T-C. PAN AND A. NAIR

Figure 19. Stress distribution of the damaged building under the service loadings.

Copyright 2009 John Wiley & Sons, Ltd.

Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

RESPONSES TO DETONATED BUILDING IN SINGAPORE

245

Figure 19. Continued

Copyright 2009 John Wiley & Sons, Ltd.

Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

246

B. LI, T-C. PAN AND A. NAIR

REFERENCES
Bentz EC. 2000. Sectional analysis of reinforced concrete members. PhD Thesis, University of Toronto, Canada. BS 8110. 1997. Structural Use of Concrete. Part 1, Code of Practice for Design and Construction. British Standards Institution: London. Clough RW, Penzien J. 1993. Dynamics of Structures (2nd edn). McGraw-Hill, Inc.: New York. Corley WG, Mlakar PF, Sozen MA, Thornton CH. 1998. The Oklahoma City Bombing: Summary and recommendation for multi-hazard mitigation. Journal of Performance of Constructed Facilities ACSE 12(3): 110112. Hibbit, Karlsson & Sorensen, Inc. 2001. ABAQUS/Standard Users Manual, Version 6.2. HKS, Inc.: Pawtucket, RI. Hyde DW. 1991. ConWep: Conventional Weapons Effects. USAEWES/SS-R. Li B, Rong HC, Pan TC. 2006. Drift-controlled design of reinforced concrete frame structures under distant blast conditions Part II: Implementation and evaluation. International Journal of Impact Engineering 34: 755770. Malvar LJ, Crawford JE. 1998. Dynamic increase factors for steel reinforcing bars. Twenty-Eight DDESB Seminar, Orlando, FL. May GC, Smith PD. 1995. Blast Effect on Building: Design of Buildings to Optimize Resistance to Blast Loading, ASCE. Thomas Telford Ltd.: London. Rong HC. 2005. Performance based blast resistant design of reinforced concrete frame structures under distant explosions. PhD Thesis, Nanyang Technological University, Singapore. Rong HC, Li B. 2007. Probabilistic response evaluation for RC exural members subjected to blast loadings. Structural Safety 29(2): 146163.

Copyright 2009 John Wiley & Sons, Ltd.

Struct. Design Tall Spec. Build. 20, 223246 (2011) DOI: 10.1002/tal

You might also like