Research

GEHL's Vision

The ever more complicated urban space that it has become today requires proportionally sophisticated and robust protection from natural disasters. One of them is earthquakes, which throughout history have caused loss of human lives and huge damages to infrastructure leading to both immediate and long-term impacts on society, rendering seismic protection of infrastructure essential to improving the quality of human lives. Beyond the purely earthquake-induced damages, various types of subsequent disasters follow an earthquake, such as building and structural collapses, landslides, tsunamis, to name a few. They make earthquake engineering a challenging task that calls for multidisciplinary seismic design involving engineering seismology, structural dynamics, and geotechnical engineering with various research methodologies (experimental, analytical, and numerical). The principle behind GEHL's earthquake engineering research simplifies such complicated problems into three parts: “input motion,” “system identification,” and “output prediction.” To prevent urban areas from strong earthquakes, we need to first decide the input motion (e.g., intensity, frequency, duration, etc.) subjected to the target sites, then identify the system characteristics (e.g., natural period, damping, material property, etc.) to understand the system’s dynamic mechanism, and finally predict the dynamic responses of the system (e.g., structure, foundation, soil, etc.) for effective seismic designs. Our previous, current, and future research have aligned, and will continue to operate within this three-part framework. Our ultimate goal is to apply this framework to other natural disasters as well in assessing the resilience, hazards, and risks of modern urban societies against them.

Soil Liquefaction during Earthquakes

Despite wide uses of stress-based approach in evaluating soil liquefaction, the energy dissipated by liquefiable soil can serve as a fundamental index of liquefiability as it combines the effects of amplitude, duration, frequency, and irregularity of earthquake waves. This research uses an energy-based method to estimate excess pore pressure buildup without pore water pressure transducers. 

Earthern Dams/Levees - Cracking and Concentrated Leak Erosion Mechanisms due to Seismic Fault Rupture

Seismic fault activities shaped valleys, giving rise to river channels where dams are typically located in accordance with these natural formations. The resulting vertical offsets also have the potential to cause CLE problems, particularly through cracks at the dam crest. The progression of CLE, however, hinges on factors such as applied hydraulic stress, erodibility, and the swelling properties of the soil. This study investigates dams situated on active faults with two key aspects: 1) estimating the crack dimensions given an amount of vertical offset by understanding crack propagation mechanisms and 2) examining the potential considering the erodibility and swelling properties of the soil along with the crack dimensions. Triaxial, swell, erosion, and beam bending tests are performed to understand the fundamental soil behavior and their properties. Centrifuge and numerical modeling explore crack propagation and CLE mechanism of levees. 

Research Methodology

Research methodology is all about creating reliable data and processing them effectively and efficiently. Our current research methodologies include experimental, numerical, analytical modeling and use data from earthquake case histories. We have numerous experiences in centrifuge modeling that have produced reliable test data for validating numerical simulations and analytical modeling. The validated numerical models were used to conduct sensitivity analyses, and the validated analytical models provided closed-form solutions and insights on the physical meaning of system dynamic mechanisms. We also utilize earthquake case histories, given the advantage of being able to investigate the actual phenomena of system responses and their consequences. Centrifuge data and data relevant to earthquake case histories are easily accessible through public databases such as DesignSafe, Kik-net, and CESMD, which we plan to continue utilizing to obtain reliable data for future research. Finally, we will incorporate probabilistic approaches into our future research. Upon accumulating big data, we will use a probabilistic approach to reveal the reliability, trend, and distribution of the dataset, to which we will incorporate machine learning to predict output responses per given input parameters. All these methodologies combined will expand our research into risk assessment and comprehensive hazard analysis.