Machine Learning Model Enhances Urban Resilience Against Liquefaction in Earthquake-Prone Areas

In earthquake-prone regions, liquefaction poses a significant threat to infrastructure as urban areas expand. Researchers from Shibaura Institute of Technology have developed machine learning models that predict soil behavior during seismic events, utilizing artificial neural networks and ensemble techniques to enhance the accuracy of underground layer mapping. This innovation aids civil engineers in identifying suitable construction sites and assists disaster management authorities in risk assessment, promoting safer urban development and improved resilience against earthquakes.

As urban populations increase, the risk posed by natural disasters amplifies, compelling city planners and disaster management officials to devise effective strategies. In earthquake-sensitive countries, particularly Japan, liquefaction represents a significant threat to infrastructure. During an earthquake, liquefaction occurs when seismic activity causes loosely packed, water-saturated soils to lose their strength, resulting in ground behavior akin to that of a liquid. This phenomenon has catastrophic consequences, such as building subsistence, foundation fractures, and damage to essential utilities, including water supply lines. Historically, liquefaction has been a common occurrence in major seismic events, incurring extensive damage. For instance, the devastating Tōhoku earthquake of 2011 led to liquefaction that impacted approximately 1,000 homes, while the 6.2 magnitude earthquake in Christchurch resulted in liquefaction that destroyed 80% of its water and sewage infrastructure. More recently, the Noto earthquake in 2024 caused significant liquefaction, affecting around 6,700 residences. To mitigate the impacts of liquefaction, Professor Shinya Inazumi and his student Yuxin Cong from Shibaura Institute of Technology in Japan have pioneered machine learning models aimed at predicting soil response during seismic events. Utilizing geological data, their approach generates detailed three-dimensional maps of subterranean soil layers, discerning regions of stability from those susceptible to liquefaction. This innovative technique provides a wider and more precise understanding of soil behavior compared to traditional manual testing methods. In a recent publication in Smart Cities dated October 8, 2024, the researchers employed artificial neural networks (ANNs) and ensemble learning methods to assess the depth of foundational layers—a vital determinant of soil stability and liquefaction susceptibility. “This study establishes a high-precision prediction method for unknown points and areas, demonstrating the significant potential of machine learning in geotechnical engineering,” stated Professor Inazumi. “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.” The identification of areas with deep, stable bearing layers enhances the selection of construction sites that can better withstand liquefaction during seismic activity. The researchers meticulously gathered data regarding the bearing layer depth from 433 points within Setagaya-ku, Tokyo, employing standard penetration tests alongside mini-ram sounding tests. Furthermore, additional site-specific information such as geographic coordinates and elevation was documented to enhance the model’s precision. Using this data, the ANN was trained to predict the bearing layer depth at 10 specified locations, with actual measurements utilized to verify predictive accuracy. A method known as bagging (bootstrap aggregation) was implemented to further enhance prediction reliability, yielding a remarkable 20% increase in accuracy. The resultant predictions facilitated the creation of a contour map showcasing the bearing layer depth within a 1 km radius around four selected sites in Setagaya Ward. This visual representation serves as an essential tool for civil engineers to identify locations with favorable soil conditions for construction, while also assisting disaster management authorities in identifying areas with heightened vulnerability to liquefaction, thus improving risk assessment and preparedness measures. The researchers foresee their methodology as instrumental in facilitating smart city development, underscoring the necessity for data-driven approaches in urban planning and infrastructure enhancement. “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,” Prof. Inazumi emphasized. Future enhancements to their model aim to incorporate additional ground conditions and develop specialized models tailored for coastal and inland areas, acknowledging the impact of groundwater levels—a critical factor influencing liquefaction dynamics.

The phenomenon of soil liquefaction has long been recognized as a significant hazard in areas prone to earthquakes, particularly in densely populated urban environments where infrastructure can be severely compromised. As cities grow, the need to ensure the resilience of structures against such natural disasters has become increasingly urgent. Machine learning technologies present innovative solutions to enhance predictive capabilities regarding soil behavior during seismic events. The research conducted by Professor Inazumi and Yuxin Cong is emblematic of the intersection between advanced technology and geotechnical engineering, providing vital insights for urban planning and disaster mitigation efforts in earthquake-prone regions.

In summary, the study conducted by Professor Shinya Inazumi and Yuxin Cong represents a critical advancement in the field of earthquake resilience, utilizing machine learning to predict liquefaction risks effectively. Their innovative modeling approach not only improves understanding of soil stability but also aids in the development of safer urban infrastructures. By enhancing predictions of ground behavior during seismic incidents, this research fosters advancements in urban planning, contributing to the establishment of robust, adaptable, and smarter cities. The ongoing efforts to refine these models will ensure that urban areas are better prepared to withstand the impacts of natural disasters, ultimately enhancing community safety and resilience.

Original Source: techxplore.com