Seismic risk analysis using Bayesian belief networks - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 197

Table 7.10 Building performance modifi ers from the walk down survey

Building performance modifi ers Field observation

Structural system (BT) Steel frame with concrete shear walls (S4)

Site seismic hazard (SSH) 0.769 g

Vertical irregularity (VI) No

Plan irregularity (PI) Yes

Deterioration of structure (DS) No

Year of construction (YC) 1950

Increase in
demand

Decrease in
resistance

Building
vulnerability

Structural damage
states

Increase in demand Decrease in capacity Building vulnerability Structural damage state

0 to 1
1 to 2
2 to 3
3 to 4
4 to 5

20.0
20.0
20.0
20.0
20.0

0 to 1
1 to 2
2 to 3
3 to 4
4 to 5

0 +
84.6
8.86
0 +
6.57

VL
L
M
H
VH

.002
35.6
9.26
15.2
40.0

1 to 2
2 to 3
3 to 4
4 to 5
5 to 6

0 +
8.04
17.6
23.5
50.8

2.5 ± 1.4 1.79 ± 0.83 3.1 ± 1.4 4.67 ± 1

7.15 Results of BBN intermediate nodes.

• Decrease in capacity, DC: (DC VL , DC L , DC M , DC H , and DC VH ) = (0, 0.84,

0.09, 0, 0.07).

• Building vulnerability, BV: (BV VL , BV L , BV M , BV H , and BV VH ) = (0, 0.84,

0.09, 0, 0.07).

• Building damage, BD: (BD VL , BD L , BD M , BD H , and BD VH ) = (0, 0.08,

0.18, 0.23, 0.51).

It is interesting to note that the probabilities for ID have an equal mass for
all ID values (ID VL , ID L , ID M , ID H , and ID VH ). Since the intermediate nodes
are calculated from the fi eld observational data, this might be the case
where there is a lack of information. In such a case, equal mass is assigned.
For the interval-valued damage states, 1-2, 2-3, 3-4, 4-5, 5-6, the correspond-
ing mean values are multiplied by the probabilities, to obtain the mean
damage state of 4.67 (Fig. 7.15). The actual damage state from the North-
ridge earthquake was moderate. The example highlights that, to get the CPT
from the fi eld observation, there should be suffi cient datasets. Thus training
the CPT with incomplete information is not desirable, and expert knowl-
edge should be solicited.

7.6 Conclusions and future trends

Consideration of seismic hazard, vulnerability assessment, and consequence
of failure is subject to different levels of uncertainty and model complexity.