Research

Advanced simulation frameworks are required to capture the different damage modes induced by various types of hazards. For instance, an integrated soil-water-foundation-structure computational platform would enable the analysis of the structure when face the collective impacts from material degradation, seismic loads (main and after shocks), geotechnical failures (e.g., soil liquefaction and landslide), hurricane-induced wave and surge loads, as well as flood-induced scour and erosion, etc. 

Our built environment is supposed to possess a much stronger hazard-resistant capability through embracing more intelligent, cost-effective, and energy-efficient solutions. To achieve this, we need to develop smart systems that have a significant degree of adaptability to enable them to maximize performance, provide coherent functionality, and minimize life-cycle cost. New classes of smart materials such as shape memory materials shall be investigated to minimize the cost of the material and the sensitivity of the properties to thermal loads. The promise of using rocking isolation for earthquake mitigation should be transformed from its current form on theoretical and numerical analyses to experimental investigations and design specifications.

Our infrastructure systems are spatiotemporal in nature. Regional-scale infrastructure fragility, risk, and resilience assessments require the development of high-fidelity computational models to capture the design details for a portfolio of infrastructure components, the temporal evolution of design philosophy, as well as the geospatial interplay of the network.  Interdisciplinary research is featured herein to incorporate (1) structural mechanics and dynamics for numerical analyses; (2) geotechnical engineering for modeling foundations and soil behavior; and (3) statistical and machine learning for model sampling, fragility development, and uncertainty treatment. 

The emerging of the large volume of data from numerical simulations, sensing techniques, computer vision, and experiment database enable the machine-learning techniques to offer new solutions and new dimensions of solutions in dealing with infrastructure resilience problems. Implementation examples include probabilistic hazard analysis, response pattern classification, hazard demand and capacity prediction, rapid post-hazard performance evaluation, dynamic system identification and damage detection, surrogate metamodeling for parametric fragility and risk assessment, as well as smart control for hazard mitigation (i.e., controllers induced by well-trained machine-learning models).