Advancements in Machine Learning: Enhancing Urban Resilience Against Liquefaction in Earthquake-Resistant Infrastructure

Researchers at the Shibaura Institute of Technology have developed machine learning models to predict soil behavior during earthquakes, specifically addressing liquefaction risks in urban areas. Their study utilizes artificial neural networks to create detailed maps indicating stable and vulnerable ground locations, significantly aiding in infrastructure planning and disaster management. This advancement aims to contribute to smarter and safer urban development in earthquake-prone regions.

As urban development progresses, the menace of natural disasters intensifies for city planners and disaster management officials, particularly in seismically active regions such as Japan. A significant threat to infrastructure in these areas is the phenomenon of liquefaction, wherein intense seismic shaking causes saturated, loose soils to lose their strength, effectively behaving like a fluid. This can lead to severe consequences, including subsiding buildings, foundation cracking, and the destruction of utilities such as water lines and roads. Historically, liquefaction has been a major contributor to damage during earthquakes, with documentation from notable events: the 2011 Tōhoku earthquake led to liquefaction damage impacting 1,000 homes, while a 6.2 magnitude earthquake in Christchurch demolished 80% of its water and sewage systems due to similar reasons. In 2024, the Noto earthquake exhibited widespread liquefaction affecting approximately 6,700 residences. In response to this critical issue, Professor Shinya Inazumi and his student Yuxin Cong from the Shibaura Institute of Technology in Japan have undertaken the development of machine learning models aimed at predicting soil reactions to seismic activities. Their approach employs geological data to formulate extensive 3D soil layer maps, effectively identifying stable sites versus areas vulnerable to liquefaction. This innovative methodology surpasses manual soil testing techniques, which often lack coverage and detail. Their research, recently published on October 8, 2024, in the journal “Smart Cities,” integrates artificial neural networks (ANNs) alongside ensemble learning methods to accurately assess the depth of underlying bearing layers—a paramount factor determining the stability of soil and its susceptibility to liquefaction. “This study establishes a high-precision prediction method for unknown points and areas, demonstrating the significant potential of machine learning in geotechnical engineering. These improved prediction models facilitate safer and more efficient infrastructure planning, which is critical for earthquake-prone regions, ultimately contributing to the development of safer and smarter cities,” states Prof. Inazumi. By predicting areas with deeper and more stable bearing layers, the researchers assist in locating grounds that could effectively support construction, especially pertinent during liquefaction events. Among their methodologies, they gathered bearing depth data from 433 points within Setagaya-ku, Tokyo, through standard penetration tests and mini-ram sounding assessments. Besides recording the depth of the bearing layers, they also noted crucial information such as longitude, latitude, and elevation for each site. This dataset enabled the training of an ANN to forecast bearing layer depth at ten different sites, verified through actual measurements to ascertain prediction accuracy. The researchers enhanced their model’s precision by implementing a bagging technique, resulting in a significant 20% increase in accuracy. With the predicted findings, they established a contour map showcasing the bearing layer depths within a 1-kilometer radius around four selected locations in Setagaya Ward. This map serves as a vital resource for civil engineers, empowering them to identify optimal construction locations characterized by stable soil conditions and assisting disaster management professionals in recognizing areas at heightened risk for liquefaction, thus improving risk assessment and mitigation frameworks. The researchers foresee their technique as a crucial component in fostering smart city development, emphasizing the need for data-driven decision-making in urban planning and infrastructure projects. “This study provides a foundation for safer, more efficient, and cost-effective urban development. By integrating advanced AI models into geotechnical analysis, smart cities can better mitigate liquefaction risks and strengthen overall urban resilience,” concludes Prof. Inazumi. Looking ahead, the researchers aim to refine their models further by incorporating additional ground condition parameters and tailoring their methodologies for both coastal and non-coastal environments, while also considering the influence of groundwater—an essential element in liquefaction context.

The increasing urbanization poses an amplified risk from natural disasters, particularly in regions vulnerable to earthquakes. Liquefaction is a serious environmental concern in these contexts, manifesting when seismic activity destabilizes saturated, loose soils. Such incidents can yield catastrophic outcomes, including infrastructure collapse and extensive damage to utilities. Both historical records and recent events underscore the necessity for enhanced predictive methodologies to assess and mitigate liquefaction risks effectively.

The innovative application of machine learning models to predict liquefaction hazards marks a significant advance in geotechnical engineering, offering cities a new avenue for enhanced disaster resilience. By leveraging artificial neural networks and ensemble learning, researchers have produced effective tools for identifying stable building sites, thus contributing to safer urban planning. Future enhancements of these models will further refine risk assessments and underpin the development of smarter, more resilient urban environments in earthquake-prone regions.

Original Source: techxplore.com