Nonlinear Seismic Response of an Existing Concrete Arch Dam Subjected to Deconvolved Earthquake Input

In the present paper, the effects of deconvolved earthquake input on the linear and nonlinear seismic response of an existing arch dam in a 3D space are investigated. nonlinearities originate from the opening/slipping of the vertical contraction joints within the dam body. The reservoir–structure interaction is taken into account by the finite element method with the appropriate boundary conditions. The reservoir was assumed to be compressible. The Shahid Abbaspour arch dam was selected for the case study. Finally, the viscous condition at the far-end boundary of the foundation is used to model the radiation effect. A quasi elastic damping model is utilized. The stiffness and mass proportional damping, equivalent to 10% of the critical damping based on the 2Hz and 6Hz frequencies of the dam foundation system, is applied to the structure. Three components of the 1994 Northridge earthquake as maximum credible earthquake are selected as the free field ground motions. The analysis is carried out in two steps. First a deconvolution analysis is performed to adjust the amplitude and frequency contents of an earthquake ground motion applied at the base of the foundation to achieve the desired output ground acceleration at the dam-foundation interface at the different points. Then the calibrated base acceleration history is applied to the foundation base of the dam-reservoir-foundation-system to perform the seismic analysis. Based on the results, spectra of the response at the dam-foundation interface at different points match very closely with the spectra of the horizontal free field ground motions. However, the existing deconvolution procedure does not produce appropriate results for high frequency ground motion records. To overcome such limitation, a modified procedure has been used for vertical earthquake which has led to better convergence. In existing procedure, a correction factor for each frequency is computed using the ratio of the Fourier amplitudes of the reproduced ground acceleration at the dam-foundation interface and free-field ground acceleration signals in a given iteration. The acceleration signal applied at the base of the foundation model is modified using the correction factor for each frequency. In modified procedure, Instead of adjusting the Fourier amplitudes, the response spectra at different frequency are adjusted. It is worth mentioning that the main novelty of the present investigation, is that it takes into account the effects of deconvolved earthquake input in addition to both the joints nonlinearity. According to the analyses, modeling vertical contraction joints leads to a decrease in the maximum value of stensile stress levels through the dam body by 6%. The extreme values of joints opening/sliding experienced by the contact elements located on the upstream face along the crest are 6.3mm and 18.1mm, respectively. The maximum values for joints sliding occurred in vicinity of the abutments. Also, maximum values of joint opening/sliding along the height of the dam body experienced by the contact elements located between the central cantilever and the adjacent ones on the upstream face occurred in crest of the dam body. However, to achieve more realistic results, other factors such as the spatial variation in ground motion, should be considered.