Technological innovations such as the use of Artificial Intelligence also referred to as AI, Internet of Things plus the use of big data among others are some of the measures that are continuing to transform the accommodation sector alongside high rates in the level of urbanization. However, climate change has real problems to this growth; natural hazards such as; earthquakes are a real menace to stability of structures.
Among them is the soil liquefaction hazard which occurs whereby, saturated soil which has lost its strength and stiffness due to stress which could be as a result of an earthquake shake up behaves like a liquid. It also significantly poses a big problem in meeting the demand for supporting buildings and other constructional projects. And hence, mitigating soil liquefaction is vital for the advancement of cities.
A group of researchers from the Shibaura Institute of Technology in Japan has proposed the AI predictive model that makes geological risk maps on soil liquefaction more detailed. Performed under the supervision of Professor Shinya Inazumi with the help of Ms. Arisa Katsuumi and Ms. Yuxin Cong, their research was made available on July, 17, 2024, in the Smart Cities journal.
As Prof. Inazumi states, “We conducted this research to fulfill the need for enhanced earthquake resilience in cities of the world, especially in built-up seismically active regions that are urbanizing fast.” The current geotechnical earthquake risk assessment and urban planning have various downsides. Standard methods of predicting soil liquefication proof too weak in matters concerning data combinatory and analysis time; hence, there is a risk gap left unmanaged.
The study’s investigators used enhanced artificial neural networks and assembled an analytical component that incorporated geotechnical and geographical data. To the best of my knowledge, this model was effectively used in improving the urban planning and infrastructures in Yokohama, Japan that has highly susceptible loose sedimentary ground especially in reclaimed areas and is very much prone to earthquakes as well.
Artificial neural networks and the gradient-boosting decision tree was used in the model to enhance the accuracy of risk predictions of soil liquefaction. The researchers realised high accuracy in the predictions on the classification of soils and the N-values essential in ascertaining liquefaction susceptibility of soils. To verify the effectiveness of the provided model, they relied on a large amount of geotechnical survey data.
From Prof. Inazumi’s analysis, the practical use of the study is to come up with hazard maps to guide urban planner or engineers when designing infrastructural projects especially in regions that are most prone to such occurrences such as soil liquefaction.
The present paper demonstrates that the AI incorporation in geotechnical engineering has achieved major fleetness in the likelihood of soil liquefaction. It is such innovative approaches that can play a critical role in the improvement of the resilience and of sustainable development of cities.