Response parameter is pseudo-acceleration for 1, 5 10 and 20% damping.
Use 3 site classes:
Vs,30> 760m∕s. S1= S2= 0.
360 ≤ Vs,30≤ 760m∕s. S2= 1, S1= 0.
Vs,30< 360m∕s. S1= 1, S2= 0.
Use 3 fault mechanisms:
Strike-slip. F1= F2= 0.
Normal. F1= 1, F2= 0.
Reverse/reverse-oblique. F2= 1, F1= 0.
Classify using rake angles. Find reverse/reverse/oblique motions similar based on residuals so combine into
Derive model for use in study on number of scaled and matched accelerograms required for inelastic
dynamic analyses of structures.
Use PEER NGA West database (Chiou et al., 2008). Only use data with known Mw, rjb and site class.
Exclude records from Taiwan to reduce database and prevent the 1813 accelerograms from the 1999
Chi-Chi sequence dominating results. Exclude records where either horizontal component was filtered with
high-pass cut-off > 0.33Hz.
Most records from strike-slip earthquakes, followed by reverse and lastly normal, for which limited data
available and over magnitude range 5.6 ≤ Mw≤ 6.9.
Note that model is not for general use and hence coefficients not reported in journal article.