Influence of near-field ground motions and their equivalent pulses on nonlinear seismic response of intake-outlet towers and predicting based on artificial neural networks

Near-fault earthquakes with forward-directivity effects contain high energy and large amplitude velocity pulses. The nearby structures can be severely damaged by such pulses. This paper studies nonlinear seismic response of reinforced concrete intake-outlet towers subjected to forward-directivity near-field earthquake ground motions and their equivalent pulses. It also compares the responses under non-forward-directivity as well as far-field earthquakes. For this purpose, 72 forward-directivity near-field earthquakes, 120 non-forward-directivity near-field records, and 132 far-field records are selected to obtain statistically-scattered and reliable results. The tower-water-foundation interaction is considered in all analyses. The displacement, dissipated energy, tensile and compressive damage of the tower are determined employing detailed finite element simulation, and the obtained responses are thoroughly discussed. Additionally, an artificial neural network model is proposed to predict displacement and tensile damage of the intake tower. The results reveal that the mean seismic demand parameters of the tower under the forward-directivity and non-forward-directivity earthquakes are 3.8 and 2.7 times of those from the far-field earthquake, respectively.