Sei sulla pagina 1di 25

21/03/2013

THE INSTITUTION OF ENGINEERS, MALAYSIA

2-DAY SYMPOSIUM AND WORKSHOP ON


EARTHQUAKE ENGINEERING IN MALAYSIA AND ASIA PACIFIC REGION
10 11 April 2013

Recommended Earthquake Loading


Model for Peninsular Malaysia
T.W. Looi

on behalf of WG1 and research team

M.C. Hee, H.H. Tsang, N.T.K. Lam

Contents
1.

Looking back at Earthquake Symposium/Workshop (Dec 2011)

2.

DISTANT SEISMIC HAZARD MODELLING

3.

LOCAL SEISMIC HAZARD MODELLING

4.

THE UNIFIED EARTHQUAKE LOADING MODEL FOR MALAYSIA

5.

Discussion

21/03/2013

1. Looking back at
Earthquake
Symposium/Workshop
(Dec 2011)
3

Endorsement in 2011 workshop

1. Hybrid method

2. Incorporate site natural period as an additional parameter


for site classification
4

21/03/2013

Disclaimer

The following presentation is based on the paper written by


the authors and published in IEM Jurutera Magazine April
2013 edition
Recommended Earthquake Loading Model for Peninsular Malaysia

All references should be made in accordance to the paper.

This presentation considers ground motions on bedrock only

2. DISTANT SEISMIC
HAZARD MODELLING

21/03/2013

Far field earthquake sources


1. Sunda Arc subduction fault
source off-shore of Sumatra
(530 730 km)
26.12.2004
Acheh (M9.1)
28.03.2005
Nias (M8.7)
2. Sumatran strike-slip fault source
(300 400 km)

Recommended distant earthquake model

Pappin and co-workers [15, 16] conducted Probabilistic Seismic Hazard


Assessment (PSHA) and produced Uniform Hazard Spectrum (UHS)

Attenuation model: Megawati 07 (large magnitude events); Atkinson &


Boore 06 (local earthquakes generated from a stable crustal structure)

21/03/2013

Recommended distant earthquake model

the UHS requires modifications due to subsequent improvements in


the accuracies of the regional specific attenuation relationships. Eg.
Megawati (2007) [39] has been updated to Megawati (2010) [19]

To achieve a more robust UHS, utilise Megawati (2010) [19] and the
latest development of CAM [32] with a logic tree weighting factor
of 0.5

CAM:

Megawati 2010:
refer to Table IV in Ref.[19]

global trend to benchmark design seismic hazard level to a RP of


2500 years as opposed to 500 years, in order to achieve a higher
level of protection for civil engineering assets

Recommended distant earthquake model

Step 1:

Three earthquake scenarios, 1) M9.3 R530, 2) M9.4 R650 and 3) M9.5 R730
were first identified by calibration using the (original) attenuation model of
Megawati (2007) [39]

10

21/03/2013

Recommended distant earthquake model

Step 2:

For each of the calibrated earthquake scenarios their respective response


spectra were then recalculated using the updated attenuation model of
Megawati (2010) [19] along with CAM [32], based on equal weightings.
The modified UHS at the reference periods were taken as the geometric
mean (GM) of results associated with the three calibrated scenarios.

(2010)

+
11

Using
GM

Recommended distant earthquake model

Modified UHS with period dependent correction factor

12

21/03/2013

A comparison with recorded data

conservative to envelope the distant earthquake

13

3. LOCAL SEISMIC
HAZARD MODELLING

14

21/03/2013

Local earthquake activities

In 2007, about 24 events of


M0.3 to 4.2 earthquakes in
Bukit Tinggi area, Pahang (30km
radius from KL center)

Other occurrences eg.


Terengganu due to Kenyir
reservoir (80s), Jerantut Pahang,
Manjung Perak and Kuala Pilah in
2009

15

Scenario based modelling and


Recommended local earthquake model

Uncertainties associated with local earthquake sources and the scarcity of


recorded data (Since 1979, the Malaysian Meteorological Department (MMD)
has installed 19 seismological stations in the Peninsula [25] )

It is not appropriate nor reliable to conduct probabilistic assessment (PSHA)


for earthquake event with long recurrence interval based on only 33 years
records.

Use Deterministic Seismic Hazard Assessment (DSHA) with suitable M-R


combinations

M-R combination of M4.2 R15 may well be considered


to be the critical earthquake scenario ?

inappropriate to do so simply because a


magnitude event from the
identified fault source cannot be completely ruled out!

larger

16

21/03/2013

Small area (eg. Peninsular


Malaysia) may appear to have
very low level of seismic
activity - only because of the
smallness of the country.

A longer period of exposure


would have displayed the
actual level of the underlying
hazard.
(see next slides)

17

18

21/03/2013

19

20

10

21/03/2013

21

22

11

21/03/2013

Epicentres of Historical Earthquakes 1856 1985


23

(modified from Dowrick, Gibson and McCue 1995)

Scenario based modelling and


Recommended local earthquake model

The apparently low level of recorded seismic activity in


the history does not mean that the underlying seismic
activity is actually low!

Recommendation: Scenario based modelling

24

12

21/03/2013

Scenario based modelling and


Recommended local earthquake model
From the global perspective, reference PGA values for RP
of 2500 years have been compiled from the literature for
a number of major cities around the world

seismic zones
Low:
eg. London (lower), Melbourne (mid), Hong Kong (upper) <0.25g
Moderate:
eg. Wenchuan (Sichuan), Christchurch (New Zealand) 0.25g-0.50g
High:
eg. Taiwan, Tokyo, Los Angeles - >0.50g

a)

b)

c)
25

Scenario based modelling and


Recommended local earthquake model

Lumantarna et al. [22]


collated Response Spectral Displacement (RSD) of 8 GMPE models
eg. 2 New Generation Attenuation (NGA) models (Abraham and Silva (2008), Campbell and
Bozorgnia (2008)) [34] (references should be made to Lumantarna et al.), CAM [30, 31, 32]
- Most likely intraplate earthquake M5.5-M6.9 (from PEER database)

Peak Displacement Demand (PDD) values with M-R combinations are listed in Table 1
considering the reference distances all notional PGA are <0.25g (low seismic zone)

26

13

21/03/2013

Scenario based modelling and


Recommended local earthquake model

The short period of Modified UHS is somewhat too low


0.23g

0.13g 0.14g

3 sub-categories
under low
seismic zone

0.07g

Too low compared to


international benchmark
27

Scenario based modelling and


Recommended local earthquake model

reasonable to adopt the mid classification sub-category


and the corresponding projected scenario of M6 R30
(notional PGA 0.13g on rock).

This level of ground motions can be taken as the basis for


defining the design local hazards for the metropolitan
area surrounding the capital city of Kuala Lumpur and
other major cities.

28

14

21/03/2013

4. THE UNIFIED
EARTHQUAKE
LOADING MODEL FOR
MALAYSIA
29

THE UNIFIED EARTHQUAKE LOADING


MODEL FOR MALAYSIA

Unifying local (T<1s) and distant earthquake (T>2s) RSA model into single hybrid RSD
model

DISTANT

LOCAL

30

15

21/03/2013

THE UNIFIED EARTHQUAKE LOADING


MODEL FOR MALAYSIA

The RSD in the transition zone (1-2s) of this proposed


hybrid model features a straight line bridging the two
parts of the displacement response spectrum

Transition
linear line

31

The distance effects

The general framework of the hybrid model can be


extrapolated for use in different cities across Peninsular
Malaysia by making use of the "path component of the
seismological model in Eq 1, which is principally a function
of distance R

Table 2: Distance effect of Path Component Attenuation


City

Kuala
Lumpur

Penang

Klang,
Melaka

Johor Bahru

Kuantan and
east-coast

R (km)

600

400

500

800

1000

Distance Factor (DF)

1.0

2.373

1.525

0.447

0.206

SD(2) (mm)

20

47

30

32

*Depending on SD(1) value,


note that SD(2) SD(TD=1)

16

21/03/2013

The distance effects

For codification purposes, displacement spectral ordinates


SD(T) are as defined by Equation 5-8, along with the
parameters summarised in Table 3

For short period, it


is consistent with
that in EC8, up to
2s for KL, and up
to 1s for other
parts of Peninsular.

Table 3: Values of the parameters describing the design response spectra


Location

for T > TD, the


model is used for
accounting for
distant hazard,
which is not
covered in EC8.

SD(TD)

SD(2)

TC

TD

Kuala Lumpur

20

20

0.2

1.0

Others

10

20 (600/R)2.4

0.2

1.0

33

*Suggestion

The distance effects

Example: Penang (R=400km to Sumatera), SD(TD =1.0s) = 10mm

Use Eq 3 and 4

DF = (600/400)^(2.4) = 2.646
SD(2) = 20 x 2.646 = 52.92mm
Use Eq 8
Compute SD(T>2)
SD(T=3) = 52.92 + 10x(3-2) = 62.92mm
(Compute the rest of SD(T>2), see tabulated SD(T>2) in next slide)
34

17

21/03/2013

The distance effects


tabulated SD(T>2)

T (s)
2
2.2
2.4
2.6
2.8
3
3.2
3.4
3.6
3.8
4
4.2
4.4
4.6
4.8
5

SD (mm)
52.92
54.92
56.92
58.92
60.92
62.92
64.92
66.92
68.92
70.92
72.92
74.92
76.92
78.92
80.92
82.92

35

The distance effects


At transition zone,
Use Eq 7
T = 1.2
SD(TD =1.0s) = 10 + (52.92 10) x (1.2-1.0) = 18.58mm
Compute the rest of SD(TDT 2)

T (s)
1
1.2
1.4
1.6
1.8
2

SD (mm)
10.00
18.58
27.17
35.75
44.34
52.92

36

18

21/03/2013

The distance effects


Using the same principle based on Eq 5 and 6:

Compute the short period RSD


T (s)

SD (mm)
0.001
0.1
0.2
0.4
0.6
0.8

0
0.5
2
4
6
8

37

The distance effects

RSD model

Consistent with EC8,


up to 2s for KL, and up
to 1s for other parts of
Peninsular.

TC = 0.2

TD = 1.0

38

19

21/03/2013

The distance effects


RSA model (use Eq 9)

TC = 0.2
39

TD = 1.0

A comparison with recorded data

conservative enough for civil protection with a 2500 year RP

40

20

21/03/2013

A comparison with EC8

Take a large magnitude distant earthquake scenario of


M9.3 R530 (notional PGA 0.095 m/s) to scale the RSA
model EC8 Type 1 (for M>5.5)

Similarly, take a local earthquake scenario of M6 R30


(notional PGA1.6 m/s) to scale the RSA model of EC8
Type 2 (for M<5.5)

Rock site conditions

41

A comparison with EC8

Type 2 model spectra are comparable to the respective


(scenario specific) simulated RS

42

21

21/03/2013

A comparison with EC8

Could have been understated by both EC models in the


longer period range depending on the location of the city

43

The 1.5% notional load

Nominal horizontal design load of 1.5% gravity load (Cl.


3.1.4.2 BS8110(1):1997) as a simplified format of
providing coverage for the seismic design requirement - is
proven to be not adequate!

1.5% notional load

44

22

21/03/2013

5. Discussion

45

1. Standardize SD(1s) for whole Peninsular


Malaysia?

Table 3: Values of the parameters describing the design response spectra


Location

SD(TD)

SD(2)

TC

TD

Kuala Lumpur

20

20

0.2

1.0

Others

20

20 (600/R)2.4

0.2

1.0

Table 2: Distance effect of Path Component Attenuation


City

Kuala
Lumpur

Penang

Klang,
Melaka

Johor Bahru

Kuantan and
east-coast

R (km)

600

400

500

800

1000

Distance Factor (DF)

1.0

2.373

1.525

0.447

0.206

SD(2) (mm)

20

47

30

46

*SD(2) SD(1)
R600km (SD=20mm) is the
benchmark, also as minimum!

23

21/03/2013

1. Standardize SD(1s) for whole Peninsular


RSD model
Malaysia? (contd)

The authors inclined to


propose having uniform hazard
across the whole Malaysia in
the short period range more
straightforward and unified!

The recommended model


presented earlier was
intended, as a more
complicated case, as
alternative.

SD(TD) = 20mm

47

TC = 0.2

TD = 1.0

1. Standardize SD(1s) for whole Peninsular


Malaysia? (contd)

RSA model
Disappearing of
hump with uniform
hazard for T<1s

Uniform hazard
for Peninsula T<1s

48

24

21/03/2013

References

51

Key References
* Numbering is according to the paper Recommended Earthquake Loading Model for Peninsular
Malaysia, IEM Jurutera Magazine April 2013 edition

[8] Eurocode 8 - Design of structures for earthquake resistance. EN 1998-1 Part 1: General rules, seismic actions
and rules for buildings.

[15] JW Pappin (Dec 2012). Arup Hong Kong lecture note to IEM.

[16] JW Pappin, PHI Yim, CHR Koo (October 2011). An approach for seismic design in Malaysia following the
principles of Eurocode 8. IEM Jurutera Megazine. 22-28.

[19] K Megawati, TC Pan (Oct 2010). Ground-motion attenuation relationship for the Sumatran megathrust
earthquakes. Earthquake Engineering and Structural Dynamics. 39: 827-845.

[22] Lumantarna, E; Wilson, J; Lam, N (2012). Bi-linear displacement response spectrum model for engineering
applications in low and moderate seismicity regions. Soil Dynamics and Earthquake Engineering. Vol. 43:

[28] Minerals and Geoscience Department (MMD) Malaysia (2011). Geological study of the seismic activities in
the bukit tinggi area.

[31] NTK Lam, Jwilson, G Hutchinson (2000). Generation Of Synthetic Earthquake Accelerograms Using
Seismological Modelling: A Review. Journal of Earthquake Engineering. Vol. 4, No. 3: 321-354.

[32] NTK Lam, T Balendra, JL Wilson, S Venkatesan (2009). Seismic load estimates of distant subduction
earthquakes affecting Singapore. Engineering Structures. 31: 1230-1240.

[39] TC Pan, K Megawati, CL Lim (2007). Seismic shaking in Singapore due to past Sumatran earthquakes. Journal
of Earthquake and Tsunami. Vol 1. No 1: 49-70.

* Listed here are key references, other references should be referred in accordance to the paper.
52

25

Potrebbero piacerti anche