City Research Online

Abstract

Bridges are one of the most important structures within a transportation network that serve as the crucial lifelines for delivering emergency supports in aftermath of catastrophic seismic exposure. The urge of engineering community to transition towards sustainable infrastructures highlights the critical role of structural optimisation in design of bridges. This underscores the need for developing a robust optimisation approach to facilitate the decision-making process on finding safe design solutions that are not only economic but also reliable in terms of seismic performance. In the present study, for the first time, a robust design optimisation (RDO) framework is proposed for performance-based seismic design (PBSD) of reinforced concrete (RC) bridges with ductile piers. The RDO framework includes a solution strategy seeking for the optimal design solutions that are cost-effective and insensitive to variation in the structural properties and the record-to-record variability of ground motion records used for nonlinear response-history analyses (NLRHAs). The proposed RDO framework is applied for seismic design optimisation of RC piers with circular cross-section adopted from a bridge case-study. The results provide insightful information about the optimal cost-effective solutions that are slightly more expensive but considerably more robust in comparison to the most economical solution.

However, it is found that the optimal solutions obtained from the RDO method may be sensitive to the arbitrary nature of ground motion selection. Within the seismic provisions of modern design codes, specific seismological criteria are outlined to guide the process of ground motion selection. Yet, variability in the choice of ground motion suites is inevitable, despite adhering to the seismological criteria. This variability has never been addressed in the previous studies within the context of PBSD and structural design optimisation. In the present research, this uncertainty is treated as a random noise parameter and is incorporated into the design optimisation problem. For this purpose, a novel reliability-based robust optimisation (RRDO) framework is proposed to explore optimal design solutions that are insensitive to variation of the selected ground motion suite. The application of the proposed RRDO framework is demonstrated using the bridge case-study, and the resulted optimal solutions are compared with those given by the RDO approach. It is observed that the RDO approach may lead to optimal solutions that are considerably more sensitive to variability of the selected ground motion suite when comparing to the RRDO. The findings of this comparison highlight the importance of the RRDO approach.

The proposed robust optimisation methods are associated with significant computational time and efforts. To facilitate the application of these methods, a metamodel-assisted optimisation framework is developed in which surrogate seismic demand models play a pivotal role by creating predictive relationships between the uncertain input design variables and dispersion in structural seismic response. These models entail specific features that distinguish them from the seismic demand models developed in previous studies. These models are capable in explicitly predicting the dispersion of structural response due to random input parameters. In addition, they are consistent with the code-based recommendations of seismic design. The effectiveness of the metamodel-assisted robust optimisation framework in identifying optimal designs is demonstrated using an RC bridge case-study with monolithic deck-to-pier connection.

The robust optimisation of structures is not only limited to initial design and construction but its importance becomes more pronounced within the structure’s service life, particularly when subjected to seismic exposures. A few studies in the past have focused on the variability in the seismic life-cycle cost of bridges. In addition, conventional optimisation frameworks do not take into account the sensitivity of seismic life-cycle cost against the variation in the stochastic selection of ground motion suites. Hence, an innovative robust seismic life-cycle cost optimisation approach is proposed in this study. In addition, a solution strategy is developed to explore the design search space for optimal solutions that are least sensitive against the variation of stochastic process of ground motion selection. For this purpose, seismic vulnerability analyses are performed at different damage states using surrogate seismic demand models. The application of the proposed robust life-cycle cost optimisation method is demonstrated on the adopted bridge case-study. The results indicate that the optimal designs given by the conventional optimisation methods may miss the solutions that are robust with respect to the variation in the suite of ground motions used in the vulnerability analyses.

Publication Type:	Thesis (Doctoral)
Subjects:	Q Science > Q Science (General) T Technology > T Technology (General) T Technology > TA Engineering (General). Civil engineering (General)
Departments:	School of Science & Technology > School of Science & Technology Doctoral Theses Doctoral Theses School of Science & Technology > Engineering

Export

Downloads