EVALUATING THE SEISMIC EFFECTS ON A CYLINDRICAL LATTICE SHELL USING A SMALL NUMBER OF RECORDS

The time-history analysis procedure and the response spectrum method are used to estimate the seismic responses. The Chinese design codes recommend a set of criteria to select and use three or seven ground motion records to calculate seismic design demands if the time-history analysis is used. It is unknown which one of the two options provide statistically more consistent estimations of the seismic design demands if the peak ground acceleration (PGA) or spectral acceleration (SA) is used to scale the records. The results in this study on a cylindrical lattice shell indicate that the preferred intensity measure for scaling depends on the response of interest. To reduce the variability in the maximum structural member forces, support reaction or base shear caused by record-to-record variability, the use of PGA for scaling is preferable to the use of SA. The preference is reversed if the maximum displacement is of interest. The results also show that the use of the response spectrum method is inadequate for this structural type if the maximum structural member forces or base shear are of interest, partly because the vibration frequencies are very closely spaced and the seismic load effects are affected by higher vibration modes. Statistics of the maximum structural member forces, support reaction and base shear are obtained by using three and seven record components. The results are used to make suggestions on the number of record components and the preferred intensity measure to evaluate the seismic design demands for structural members and supports.