A new study presented by a team of researchers at Institute of Industrial Science at the University of Tokyo have created a Artificial Intelligence computer program for know-how where emitted radioactive material will settle to minimize loss of life in the event of accident in a nuclear power plant. The computer program can predict with accuracy over 30 hours in advance where emitted radioactive material will settle based on forecasts of wind patterns.
The tool aids evacuation plans and other measures to be implemented minimize loss of life in the event of a nuclear power plant accident akin to Fukushima Daiichi Nuclear Power Plant accident in 2011.
Wind Pattern Forecasts used to Predict Radioactive Fallout
The unreliability of the existing atmospheric modeling tools that could not be used for planning immediately after the disaster prompted this study. The study involved creating a system based on a type of artificial intelligence called machine learning. The program can use data of previous weather patterns to estimate the route radioactive material is likely to take.
The new tool was tested using past weather-related data to estimate where radioactive material would be distributed if it were released from a particular point. Following this, in subsequent testing, the tool could predict with at least 85 percent accuracy where the radioactive material would disperse. The accuracy of predictions rose to 95 percent in winter when weather patterns are more predictable.
The fact that the accuracy of the Artificial Intelligence tool remained reliable to make predictions for the future over 30 hours in advance is extremely important for disaster scenarios. This gives time to authorities to arrange evacuation strategy in the event of an accident.