Seismic risk assessment of water supply systems - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 663

pipeline repairs and locations of repairs after the 1994 Northridge earth-
quake in Los Angeles and incorporated in a large GIS database, consisting
of over 12 000 km of pipelines in the greater Los Angeles area and more
than 240 strong motion records. Records from over 240 strong motion sta-
tions throughout the earthquake-affected area were analyzed with respect
to various seismic parameters (Toprak 1998). The spatial distributions of
different seismic parameters were estimated by interpolation and superim-
posed on the pipeline network and spatially distributed database of pipeline
damage. Where possible, pipeline repairs in zones of documented PGD
were screened from the repair rates so that the resulting statistics would
refl ect principally the effects of seismic waves or TGD. Using a GIS soft-
ware, the repair rate was calculated for areas infl uenced by specifi c seismic
parameters. Correlations then were developed through regression proce-
dures to obtain the most statistically signifi cant relationships among repair
rate and values of different seismic parameters. It was found that the most
statistically relevant parameter for correlation with repair rate is PGV.
Figure 24.3 shows the linear regressions developed by Toprak (1998) and
Jeon (2002) and utilized to calculate the repair rates from PGVs for cast
iron and ductile iron pipelines. Interested readers are referred to Jeon

10.000

Cast iron
Ductile iron

Fit Equation (cast iron):
Ln(

Y )=1.21Ln( X ) - 6.81

2 = 0.85

r

1.000

0.100

0.010

Fit equation (ductile iron):
Ln( Y )=1.84Ln( X ) - 9.40

2 = 0.74

r

0.001

10 100 1000

PGV (cm/s)

24.3 Regressions of repair rate vs. PGV for cast iron and ductile iron
pipelines (after Jeon 2002).