Preparing for Natural Disasters with AI

Predicting infrastructure damage from natural disasters isn't easy, because of the unpredictability of environmental conditions and the varying intensity of disasters. It currently uses labor-intensive field surveys and basic statistical methods, which can often be inaccurate and makes it difficult to account for complex environmental factors, which then leads to slow responses, more damage, misallocation of resources, and critical infrastructure being put in harm's way.

NTT has developed an AI solution to improve infrastructure resilience and make early disaster recovery easier.

The solution uses machine learning to predict damage to areas during disasters, using publicly available data on topography, weather, and other relevant factors. Unlike traditional methods requiring specific field surveys, it can forecast damage regardless of installation location, using data such as rainfall, elevation, ground strength, and distance from rivers.

How is this possible? It's all thanks to NTT Group's huge amounts of data on telecommunications infrastructure damage from the disasters of the past. By analyzing and understanding patterns from the data, the AI is able to make much more accurate predictions than before, which then allows for proactive responses to mitigate the effects of natural disasters.

For example, the technology can predict damage to utility poles caused by landslides with an accuracy of 98%, using only rainfall, elevation, and ground strength data. It can also predict damage to bridge-attached pipelines due to river flooding with 90% accuracy, by analyzing water level fluctuations and river width. What's more, it can forecast damage to underground pipelines from earthquakes with 87% accuracy, based on seismic activity data. As with many of NTT's breakthroughs, the secret is a simple one: collect large amounts of data, analyze it carefully, then make intelligent inferences about what the data means.

Being able to predict areas that might suffer damage during natural disasters isn't just an academic exercise. It means that places using the technology can manage preemptive reinforcement of vulnerable infrastructure, such as strengthening pipelines before anticipated seismic activity or preparing restoration materials ahead of heavy rainfall. And the technology's usefulness goes beyond telecommunications infrastructure, potentially benefiting power lines, signal poles, bridge structures, and underground utilities like water and gas pipes.

By integrating its AI into disaster preparation strategies, NTT's concept testing shows that it can greatly reduce infrastructure damage and speed up recovery efforts. Although it has been designed and tested in Japan, the technology could be useful all over the world. The state of California could benefit from its earthquake damage predictions. Southeast Asia suffers regular flood and typhoon damage. Meanwhile, the landslides regularly seen in Southern Europe may be mitigated by knowing the severity of the damage they cause.

NTT's clever and human-focused use of AI promises to strengthen disaster mitigation and make quicker recovery possible, safeguarding essential services and infrastructure in the face of natural calamities.

For further information, please see this link:
https://group.ntt/en/newsrelease/2024/04/25/240425a.html

If you have any questions on the content of this article, please contact:
NTT Information Network Laboratory Group
Public Relations
[email protected]

NTT-Innovating the Future

Public link

Please use the above public link if you want to share this noodl on another website.