1

Seismic Design of Transfer Structures for Hong Kong Conditions

Dr. Ray SUThe University of Hong Kong

Recent Advances in Structural Designin Regions of Low-to-Moderate Seismicity

INTERNATIONAL SYMPOSIUM

28 June 2019, Hong Kong SAR

2

Transfer structure (TS) in modern buildings

Transfer structures are commonly used in low-to-moderate seismicity regions.However, their resilience to extreme loads, such as earthquake attack, is questionable.

Malaysia

Sydney Hong Kong Singapore

3

cracks on walls

Cracks above TS

4

Cracks above TS

5

Bearing walls / columns

Transfer beams / platesColumns / walls

Structural irregularitiesLateral stiffness

Lateral strength

Mass

Seismic vulnerability of transfer structure

Out-of-plane deformation of transfer structure

Soft/weak storey failure

Shear concentration at structural wallsTensile failure in structural slabs

Those failure modes will be examined under the HK design based earthquake (DBE) loads.

Transfer structures are vulnerable under a seismic attack.

6

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 20 40 60 80 100 120 140RSD (mm)

RSA

(g)

Site 0

Site 1

Site 2

Site 3

Earthquake demands for Hong Kong

Su RKL, Tang TO, Lee CL and Tsang HH (2015), CIC Research Journal, 2, p45-54.Su RKL, Lee CL, He CH, Tsang HH and Law CW (2015), HKIE Transactions, 22(3), p179-191.

PGA = 0.11gPSA = 0.8 gPSV = 0.5 m/sPSD = 125 mm

Rock

Site 1

Site 2

Site 3

<10m 10-20m 20-40m >40m

Site Specific Earthquake Spectrum Analysis

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 50 100 150 200 250 300RSD (mm)

RSA

(g)

Site 0

Site 1

Site 2

Site 3

Rare Earthquake (or MCE) Spectra(RP = 2475 years)

PGA = 0.22gPSA = 1.5 gPSV = 0.87 m/sPSD = 250 mm

Rock

Site 1

Site 2

Site 3

Seismic response is significantly affected by the type of sites and RP considered. Rock sites are better as their seismic response is the lowest.

Seismic response is significantly affected by the type of sites and RP considered. Rock sites are better as their seismic response is the lowest.

Design Based Earthquake (DBE) Spectra(RP = 475 years)

7

Shake Table Tests

8

Above the transfer structure Below the transfer structure Above the transfer plate Below the transfer plate

Below the transfer plate Above the transfer plate

y

Podium structure

Case 1

Xu

et al 2000 Case 2

Ye et al 2003

Case 3

Huang et al 2004 Case 4

Li et al 2006

Shake table tests

Xu, P., Wang, C., Hao, R. and Xiao, C. (2000) Building Structures 30(1), p38-42.Ye, Y., Liang, X., Yin, Y., Li, Q., Zhou, Y. and Gao, X. (2003) Structural Engineers 4, p7-12.Huang, X., Jin J., Zhou, F., Yang, Z. and Luo, X. (2004) Earthquake Engineering and Engineering Vibration 24(3), p73-81.Li, C.S., Lam, S.S.E., Zhang, M.Z. and Wong, Y.L. (2006) Journal

of Structural Engineering ASCE 132(11), p1732-1744.

68 stories 33 stories

28 stories 34 stories

9Li et al. 2006Gao et al. 2003

Huang et al. 2004

Shake table tests

Case 2 Case 3 Case 4

Elevation views

10

Su RKL (2008) Electronic Journal of Structural Engineering, 8, p99-109.

Shake table tests

Earthquake intensity

Ye et al. (2003)

Huang et al. (2004)

Li et al. (2006)

Minor 0.02-0.03g 0.035-0.04g 0.02-0.06g Moderate 0.07-0.16g 0.07-0.12g 0.08-0.14g Major 0.12-0.30g 0.16g 0.15-0.34g

Peak ground accelerations (PGA) of the prototypes adopted in shaking table tests

Note: PGA of Hong Kong ≈

0.11 g for DBE

and 0.22 g for MCE

11

Under frequent earthquake attack, all the building models remained elastic, no cracks

were found and

no change of the natural frequencies.

Under occasional earthquake, cracks began

to

occur at the tops of columns below transfer beams and at the base of columns above transfer floor. The natural frequencies dropped

by 10 to 20%.

After rare earthquake, all the models were severely damaged. The natural frequency

of the structures

decreased by 20-46%. The damping ratio

was increased from 2% after frequent earthquakes to 4.5-

7.5% after a rare earthquake.

Su RKL (2008) Electronic Journal of Structural Engineering, 8, p99-109.

Shake table test results

12

After rare earthquake, serious damage

was found in

the peripheral shear walls

above the transfer floor (in cases 1 and 3).

Tension failure

was found on

the end shear walls above the transfer plate (in case 4).

Floor slabs

and beam-wall joints

were also cracked (in cases 2 and 4). A weak floor formed at the floor above the transfer structure (in case 3).

Most of the damage was caused by shear concentration effects.

All the buildings survived without collapse

Shake table test results

Case 4, after Li et al. 2006

Case 3, after Huang et al. 2004

13

Soft Story Investigation

14

Section A-A

Transfer Plate

Soft storey analysis

•Su RKL, Chandler AM, Li JH and Lam NTK (2002), Structural Engineering and Mechanics, 14(3), p287-306. •Li JH, PhD thesis, Seismic Drift Assessment of Buildings in HK with particular application to transfer structures, 2004

Seismic analysis was conducted for a 35-storey RC residential building with a transfer plate located at the 6th

floor which was

designed to the old HK code 2004. Structural systemShear walls above TPColumn frame below TP

Basic informationHeight: 112.5 mPlan: 18 m x 56 mTP level: at 6th

floor (20 m above ground)Plate thickness: 2.5 mMega column diameter: 2.5 mConcrete grade:

fcu,k

= 45 MPa

below 25/Ffor columns and walls

Steel grade: fy,k

= 460 MPaLive load = 2.5 kPa

Roof drift ratio under wind loads = 1/677 < 1/500

Drift ratio below TP = 1/800 < 1/700

Str. layout at TP level

Typical floor layout above TP

pinned

fixed

15

0 20 40 60-1.8

-0.9

0.0

0.9

1.8

Gre

ound

Acc

eler

atio

n (m

/sec2 )

Period (sec)

SOIL01

Plastic hinges

p yu

Transfer Plate

Potential Soft-Storey Failure Mechanism

5 simulated time-history records at soil sites are considered

Input seismic loads

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 20 40 60 80 100 120 140

RSD (mm)

RSA

(g)

Site 0

Site 1

Site 2

Site 3

Li(2004)

Similar to HK design base earthquake

Average acceleration response spectrum

1.1 s0.44 s

16

Maximum displacement profiles

0

10

20

30

40

0 40 80 120 160Displacement (mm)

Stor

ey N

o.

SOIL01

SOIL02

SOIL03

SOIL04

SOIL05

ENVELOPETP level

53 mm 92 mm

The mega-columns designed to the old HK code can survive under DBE.

However the seismic demands would be double under rare earthquake attack. The strength of the mega-columns, if designed as force controlled members following PBSD, would be insufficient.

Higher concrete grade could be used to reduce the ALR and improve the axial strength and lateral deformability of the columns.

Drif

t cap

acity

17

Soft storey analysis

Seismic analysis was conducted for a 7-storey RC building with a transfer beam located at the 1st

floor

Structural systemBeam-column frame above TBColumn frame below TB

Basic informationHeight: 28.9 mWidth: 23.7 mTB level: at 1st

floor (7 m above ground)TB size: 1.7 x 2.1 m (dp)Base column size: 0.675x1.6 mConcrete grade:

fcu,k

= 40 MPaSteel grade:

fy,k

= 460 MPaLive load = 2.5 kPa

Gravity load case controlled the RC design2050 2050

2890

0

8050

4350

5@3 3

00

Transfer Beam (TB)

General Notes:(A) Concrete grade (B) Beam Schedule (c) Column ScheduleAll floors 40D B1, B2 1700 x 2100dp Base Col. 675 x 1600 B3, B4 600 x 650dp Exterior Col. 600 x 900 Interior Col. 600 x 900

1 2 3 4

9400 4900 9400

B1 B2

B3 B4

B4B3

B3 B4

B3

B3

B4

B4

B3 B4

1/F

2/F

3/F

4/F

5/F

6/F

7/F

X

Z

•Li JH, Su RKL and Chandler AM (2003), Engineering Structures, 25(12), p1537-1549. •Li JH, PhD thesis, 2004

Fixed

Fixed

18

Collapse Mechanism from POASoft storey analysis

1 2

21/F

2/F

3/F

4/F

5/F

6/F

7/F

Base

1 2

21/F

2/F

3/F

4/F

5/F

6/F

7/F

Base

TB

reach ultimate capacity (beyond Point C in Figure 7-3)

immediately prior to collapse(just before or at Point C in Figure 7-3)

in the range of the effective yield capacity and the ultimate capacity (in the segment of labels ‘B’ and ‘C’in Figure 7-3)

reach ultimate capacity (beyond Point C in Figure 7-3)

immediately prior to collapse(just before or at Point C in Figure 7-3)

in the range of the effective yield capacity and the ultimate capacity (in the segment of labels ‘B’ and ‘C’in Figure 7-3)

)

)

)

0

-15.0 -7.5 0.0 7.5 15.0-10

-5

0

5

10-15.0 -7.5 0.0 7.5 15.0

-10

-5

0

5

10

idealized moment-curvature relationships [14]

ED

CB

A

Under Ultimate Load Under Working Load

Mom

ent (

MN

.m)

Curvature (m/1000)(rad/1000 m)

Moment curvature of base column

21 mmDrift capacity

19

Displacement profiles from incremental dynamic analysisSoft storey analysis

The base columns could barely satisfy the seismic demand under DBE.

In general, low rise buildings are more critical than high rise buildings under seismic loads.

Intensity Ratio = 0.5 1.0 1.5

0

5

10

15

20

25

30

0 50 100 150 200 250 300 350

Displacement (mm)

Hei

ght (

m)

1/F

2/F

3/F

4/F

5/F

6/F

7/F

POA

Drif

t cap

acity

21 m

m

(≈DBE)

TB level19 mmDrift demand

20

Shear Concentration Effect

21

Columns

Shear walls

Transfer Structure

Core wallColumns

Shear walls

Transfer Structure

Core wallColumns

Shear walls

Transfer Structure

Core wall

Shear force increasing

Shear force decreasing

θe1 θc θe2

TT

CC

T- TensionC- Compression

Shear concentration effects

•Su RKL and Cheng MH (2009), The Structural Design of Tall and Special Buildings 18(6), p657-671.•Su RKL (2008), Electronic Journal of Structural Engineering, 8, p99-109.

Local out-of-plane deformations of transfer structure under lateral loads

Very high tension / compression force in slabs

and shear force in walls.

Substantial reduction

of the shear span (Ls

=M/V) of walls.

Decrease in

deformability / ductility of walls.

Center Wall

Exterior Wall

Transfer Beam

Column Center Wall

Coupling Beam Enlarged

22

a b c b a

27@

3000

=810

00

1350

0

Center Wall

Exterior Wall

Transfer Beam

Column Center Wall

Coupling Beam

0

5

10

15

20

25

30

-0.0005 0.0005 0.0015 0.0025

Inter-storey drift

Floo

r lev

el

Core wallExterior wallColumn

Transfer level

Inter-storey drift

Shear concentration effectsShear force

Shear force (kN)

0

5

10

15

20

25

30

0 500 1000 1500 2000

Core wallExterior wallColumn

Transfer level

The difference in wall rotations above the transfer structure causes shear concentration in walls.

The difference in wall rotations above the transfer structure causes shear concentration in walls.

23

Shear concentration effects

Shear span-to-depth ratio

Dis

tribu

tion

(%)

The drift capacity of squat walls (SDL

1.5) controls the seismic performance of the structure.

SDR of a big majority of walls (>70%) are less than 1.5.

Shear span-to-depth ratios (SDR) of structural walls above transfer floor for 3 tall buildings are examined.

Fstrut

WALL

VM

dw

Shear span =LW

= M/VWall depth = dw

TS

SDR = Lw

/dw

SDR = Lw

/dw

h ≈

Lw

24

Stiffening

the transfer beams can only moderate the shear concentration effect

as the local rotation of TB cannot be

eliminated.

0

1

2

3

4

5

6

1400 1600 1800 2000 2200 2400

Transfer beam depth (mm)

SCF

Model A

Model B

Model C

Model D

Transfer beam

Dep

th

Effect of depth of transfer beam

•Su RKL and Cheng MH (2009), The Structural Design of Tall and Special Buildings 18(6), p657-671.

Under Seismic LoadsShear concentration factor (SCF) is defined as the shear stress at the wall concerned

to the average shear stress of all the walls

above the transfer level.

External wall

Ext. wall

Int. wall

25

Placing the transfer structure at a high level can remarkably increase the shear concentration effect. It is because both the global rotation of superstructure

and local rotations of transfer structure

increase with the level of the transfer structure.

For seismic design, the transfer level should be limited to a lower storey, e.g. the bottom level of transfer structure should be less than 20 m above ground .

0

1

2

3

4

5

6

7

8

9

3 6 9 12

Vertical position of transfer beam (floor level)

SCF

Model A

Model B

Model C

Model D

heig

ht

Transfer beam

Effect of vertical positioning of TB

•Su RKL and Cheng MH (2009), The Structural Design of Tall and Special Buildings 18(6), p657-671.

Under Seismic Loads

26

Gravity loads can cause out-of-plane deformation of transfer structures and shear concentration.

Effect of gravity loads

transfer structure

•Tang TO and Su RKL (2015), Proceedings of the ICE Structures & Buildings, 168(1), p40-55.

For example,σG = 11 MPa, w ≈

0.3×11 MPa = 3.3 MPa.

A parametric study by Tang and Su (2015) found that for improperly designed transfer structures, the induced shear stress can go up to around 30% of the average vertical stress above the transfer structure.

To mitigate the excessive deflection of the transfer structures under gravity loads, the maximum sagging deformation of transfer structure under working gravity loads is recommended to be limited to L/1000, where L is the clear span of the transfer structure.

Such shear forces are self-balance at each floor.

27

Deformation limitsInter-storey Drift Ratio (IDR)It is defined as the relative horizontal displacement of two adjacent floors to the floor height ratio. It has been widely used for controlling damage to structural and non-structural components.

transfer structure

Shear walls above transfer structure

IDR is small but internal forces are high.

θ

xi-1hi

xi

Sway deformation of frame under lateral load

IDRi

= (xi+1 -

xi

)/hi

IDRi

= (xi+1 -

xi

)/hi

θ

θf

However, it is not applicable for the cases where there is a significant rigid body rotation θf

.

Upper zone of a tall building

θ

≈

θf

, IDR is high but internal forces are small.

For example:

28

Deformation limitsDistortional Inter-storey Drift Ratio (DIDR)It is obtained by eliminating the floor rotation

(θf

) from the IDR, is an appropriate measure of the shear deformation

of a structural wall. This ratio is

particularly suitable for quantifying local distortions and deformations induced by gravity and seismic loads.

θd

= IDR -

θf

θd

= IDR -

θf

IDR = θ DIDR = θd

θf

DIDR = θd

True distortional deformation after eliminating the floor rotation

29

Comparison of DR demand and capacity

A B C

DIDR demands

are all less than 0.5% which is much less than the DR capacity

of around 0.8% (assuming an ALR of 0.3).

Thus those shear walls could survive under the DBE.

THA was conducted to obtained the seismic drift demands of tall residential buildings with TS under DBE.

L/1560 L/3370 L/4080Sagging of TS:<L/1000

DIDR Demand

0.8%, ALR = 0.3

30

Soft storey failure mode is more critical for low-rise than high rise buildings.

Local deformation of transfer structure can cause abrupt increase in force demands above the transfer floor and substantial reduction of shear span of structural walls.

Drift capacity of squat walls is around 0.8% (taking the ALR = 0.3).

To mitigate the excessive deflection of the transfer structures under gravity loads, the maximum sagging deformation of transfer

structure under gravity loads should be limited (e.g.1/1000 of the clear span).

To control the seismic induced deformation demands, transfer structure should be placed at a lower storey (e.g. ≤

20 m above

ground).

DIDR is better than IDR for control damage.

To increase the drift capacity and shear strength of structural walls, higher grade concrete (e.g. C60) can be used.

Rock sites are favorable as they induce less seismic response.

Conclusions

31

Bearing walls / columns

Transfer beams / platesColumns / walls

Structural irregularitiesLateral stiffness

Lateral strength

Mass

Out-of-plane deformation of transfer structure

Soft/weak storey failure

Shear concentration at structural wallsTensile failure in structural slabs

Thank YOU

for your kind attention

Ray Su Email: klsu@hku.hk

AcknowledgementI would like to express my gratitude to Mr

Lijie

Chen and Ms

Shuyi

Liu for computing the SDR and DIDR results.

32

180851

4507508512751

1451

..

..Af.

NAf..

N

ccuccu

P = N/1.4551850 .f.f k,cuectedexp,c

Working load:Expected cylinder strength:

cc AfPALR

Axial load ratio:

Control of axial load ratio (ALR)Since the implementation of the new concrete code in 2013, the axial compression stress in shear walls has being controlled. It is very effective in increasing the drift capacity of structural walls.

The drift ratio capacity

of RC squat walls

designed according to the current concrete design code should be more than 0.8% (assuming an ALR of 0.3). Higher grade concrete (e.g. C60) is recommended to be used above

TS to

increase the shear strength and drift capacity of walls