On the accuracy of the N2 inelastic spectrum for timber structures

Timber buildings in earthquake prone regions must be designed to withstand the seismic actions. The N2 method, already employed in the Eurocode 8 to check the seismic resistance of a building employing its capacity curve, uses inelastic spectra calculated on the basis of the system ductility. Such a procedure was extensively used for the design of steel and reinforced concrete structures. Timber structures, however, have a different and more complex hysteretic behaviour characterized by significant pinching, strength degradation and softening. Consequently, the current version of the N2 proposed spectrum may be inadequate for the design of timber systems. This paper investigates the accuracy of the N2 spectrum for timber structures by calculating the ‘rigorous’ inelastic spectra for natural seismic records selected to match, with their average, a chosen design elastic spectrum. A purposely developed software has been used to obtain the inelastic spectra for a Single Degree of Freedom (SDOF) system characterized by a slip-type hysteretic relationship with pinching typical of timber structures. Two different sets of records have been considered: the former is consistent, on average, with a given design spectrum from Eurocode 8, the latter is taken from a strong motion database by selecting the records having soil class A and PGA within a chosen range. Non-linear dynamic analyses have been carried out by varying the level of ductility and the natural vibration period of the SDOF systems. The effects of hardening or softening of the system behaviour have also been analysed. Two analysis procedures have been used with the aim to have a further confirmation of the achieved results. The comparisons between the rigorous and the approximated N2 spectrum demonstrate that, in general, the N2 method always give fairly good results in estimating the inelastic spectra even for timber structures.