PSI - Issue 44

Gianluca Salamida et al. / Procedia Structural Integrity 44 (2023) 139–146 Gianluca Salamida et al. / Structural Integrity Procedia 00 (2022) 000–000

144

6

multi-stripe analysis was used in order to evaluating the displacement demand as a function of an IM. For each IM stripe, a set of 30 accelerometric recordings was used to obtain the maximum displacement of the 27 considered SDoF systems. For each stripe, acceleration recordings were selected from PEER Strong Ground Motion Databases in order to minimise scale factors and to ensure good variability of ground motion characteristics, also limiting the number of records associated with the same seismic event. Non-impulsive recordings with a maximum usable period greater than 3 s were considered. 5.1. Efficiency of different IMs Different IMs were considered in order to assess its efficiency in the description of the structural response. In particular, the following IMs were studied in this work: peak ground acceleration, PGA ; peak ground velocity, PGV ; pseudo-acceleration at the SDoF system periods, PSA(T 1 ) ; average pseudo-acceleration over the period ranges between 0.3 s and 0.6 s, 0.5 s and 1.0 s and T 1 and 2T 1 (where T 1 is the SDoF system period), PSA avgT0.3-0.6 , PSA avgT0.5 1 and PSA avgT-2T ; Housner Seismic Intensity, SI H . Considering 14 IM stripes in the analysis (in terms of PSA avgT0.3-0.6 values), 11340 IM-displacement value pairs were obtained. An example of results for PGA and PSA avgT0.5-1 is shown in Fig. 3, in which also regression lines are reported. To make the comparison possible, IM values were normalised, in order to have, for the logarithm of each IM, a zero mean value and a unit standard deviation.

Fig. 3. Results in terms of normalised IMs and displacement logarithm for PGA and PSAavgT0.5-1; µ and σ represent the mean value of ln(IM) and its standard deviation.

A comparison between the efficiency of the different IMs studied was made on the slope of regression lines and on the standard deviation. Results of this comparison are shown in Table 3. Although PGA is the most widely used IM, as expected, it showed the highest dispersion in terms of maximum displacement. PGV showed a better correlation with displacement than PGA , but the dispersion of results is on average greater than spectral IMs. On average, the best correlation between IM and displacement of SDoF systems was found with PSA avgT0.5-1 and PSA avgT-2T . Housner's intensity also demonstrated good performance. Although the elastic period of the SDoF systems is in the range of 0.3 0.6 s, the increase in period due to damage during the analysis makes it more correlated with pseudo average accelerations at higher period intervals ( PSA avgT0.5-1 ).

Table 3. Regression slopes, a 1 , and regression standard deviation, σ , for different IMs. IM PGA PGV PSA(T 1 ) PSA avgT0.3-0.6 PSA avgT0.5-1

PSA avgT-2T

SI H

a 1

σ

a 1

σ

a 1

σ

a 1

σ

a 1

σ

a 1

σ

a 1

σ

X masses X mode1 Y masses Y mode1

1.48 1.44 1.31 1.28

0.66 0.66 0.66 0.68

1.50 1.47 1.38 1.36

0.62 0.59 0.49 0.48

1.50 1.47 1.34 1.32

0.63 0.61 0.59 0.59

1.51 1.47 1.34 1.31

0.61 0.59 0.59 0.60

1.52 1.49 1.39 1.36

0.57 0.54 0.47 0.58

1.52 1.49 1.39 1.36

0.57 0.55 0.48 0.48

1.50 1.47 1.39 1.37

0.62 0.58 0.46 0.46