3 min read
What if we could accurately predict and prepare for the impact of natural disasters on a country’s infrastructure? Just imagine how much less vulnerable society would be in the face of earthquakes, hurricanes and other catastrophic events.
Artificial intelligence (AI) is making this vision a reality.
Traditionally infrastructure was built to withstand natural disasters, but predicting their precise impact has been difficult, if not impossible. Variables like local topography, weather conditions and intensity, and infrastructure design make it challenging to foresee where vulnerabilities lie. NTT is building a trusted, secure AI solution for disaster prediction that solves these problems by leveraging vast datasets and advanced predictive modeling. Using machine learning, this is currently being developed and tested in Japan, with the goal of benefitting people, society and the planet into the future.
How AI technology predicts and protects against disasters
NTT’s AI system provides disaster mitigation and early recovery by integrating historical data from across Japan, detailing past infrastructure damage, which then allows it to learn disaster patterns and anticipate risks. By combining proprietary data with publicly available environmental and meteorological data in a fair and open manner, the AI constructs predictive models.
There are three elements involved in this AI system, all of which include technologies that can accurately predict damage to individual facilities throughout Japan without any special field surveys. This includes:
- Landslide disaster prediction for utility poles: Using data such as rainfall intensity, elevation, soil strength and proximity to rivers, AI predicts which utility poles are most vulnerable to landslide disasters, achieving 98 percent accuracy.
- River flooding and bridge-attached pipelines: Analyzing river dynamics like water level fluctuations and width, AI identifies bridge-attached pipelines at risk, predicting damage with 90 percent accuracy.
- Earthquake impact on underground pipelines: By considering factors like earthquake magnitude and pipeline type, AI forecasts damage to underground pipelines with an accuracy rate of 87 percent.
The advantages of AI-driven disaster preparation
Using this predictive data, at-risk facilities and infrastructure can be prepared in advance to ensure that damage is minimal and recovery can happen quickly in the event of heavy rain or earthquakes. For example, pipelines that are at high risk of damage from anticipated seismic motion can be reinforced, and materials necessary for restoration can be in place in advance of anticipated heavy rainfall.
In addition, the applications of this technology extend far beyond utility poles and pipelines. The predictive models for utility poles can be adapted to power and signal poles, for instance, while the river flooding model can inform the protection of entire bridges. Similarly, the earthquake prediction model for underground pipelines can be applied to safeguard water and gas distribution networks.
Building a safer, more resilient infrastructure through AI
Looking ahead, NTT’s goal is to expand this technology to ensure all infrastructure can be made more robust and resilient. By predicting disaster impacts in advance, the aim is not only to minimize damage but also to expedite recovery efforts, ensuring communities can bounce back swiftly from earthquakes, heavy rain and other disasters.
As climate change intensifies the frequency and severity of natural disasters, innovative technologies like NTT’s AI-driven predictive models are more crucial than ever. Using the power of data and AI, we are working to ensure infrastructure is not just resilient but predictive and proactive. This technology has the potential to redefine disaster preparedness, creating a safer future for all.
