Reliability-based design of tuned-mass dampers with soil−structure interaction

The design of tuned-mass dampers is based on minimising lateral displacement (displacement-based design) and then maximising performance (reliability-based design). The research study reports that genetic algorithms (GAs) were used to solve this optimisation problem. The performance of displacement-based designs was evaluated using Monte Carlo simulations (MCSs) for uncertainties in structural, soil and loading parameters. For lateral loading uncertainty, artificial records were generated by a proposed new method and applied as random vibrations, while soil−structure interaction (SSI) was simulated using a cone model. In addition, for solving the time-consuming problem of combining GAs with MCSs, trained artificial neural networks (ANNs) were used as alternatives. The study showed the importance of considering SSI in the design of tuned-mass dampers. It also showed that trained ANNs could accurately evaluate the performance of a tuned mass damper in a fraction of a second compared with several hours using a MCS.