Students develop AI that calculates when and where lightning strikes – and the technology can predict a bolt 30 MINUTES before it happens with 80% accuracy
- System uses meteorological data and artificial intelligence to predict lightning
- Looks at atmospheric pressure, air temperature, humidity and wind speed
- Can predict lightning 30 minutes before and within an 18 mile radius
Lightning is so far considered to be & # 39; the most unpredictable phenomenon in nature & # 39 ;.
By combining meteorological data and artificial intelligence, students have determined where and when lightning strikes within 10 to 30 minutes after the event – and with an accuracy of 80 percent
The technology is designed to work as an early warning system to prevent the effects of lightning strikes on critical infrastructure, sensitive equipment and outdoor facilities.
The system, developed by students of École polytechnique fédérale de Lausanne (EPFL School of Technology), can predict when and where the lighting will strike within 10 to 30 minutes within an 18-mile radius.
Scroll down for video
Lightning is so far considered to be & # 39; the most unpredictable phenomenon in nature & # 39 ;. By combining meteorological data and artificial intelligence, students have determined when and where the lighting will strike in 10 to 30 minutes
Amirhossein Mostajabi, the Ph.D. student who invented the technique, said: & # 39; Current systems are slow and very complex, and they require expensive external data obtained via radar or satellite. & # 39;
& # 39; Our method uses data that can be obtained at any weather station. This means that we can cover remote areas that are outside the radar and satellite range and where communication networks are not available. & # 39;
The method of the EPFL researchers uses a machine-learning algorithm that is trained to recognize circumstances that lead to lightning.
It analyzes atmospheric pressure, air temperature, relative humidity and wind speed.
The method of the EPFL researchers uses a machine-learning algorithm that is trained to recognize circumstances that lead to lightning. It analyzes atmospheric pressure, air temperature, relative humidity and wind speed
& # 39; The dataset used in the ML model consists of data used as predictors, namely available meteorological data (air pressure, air temperature, relative humidity and wind speed) and data on lightning activity as a reaction & # 39; article in the magazine Nature.
& # 39; Lightning location system data is used to first train the ML model and then validate the accuracy of lightning warnings it generates, as well as the competing baselines. & # 39;
The team discovered that the four parameters were correlated with recordings of lightning detection and location systems, and with the help of that method the algorithm could learn under what circumstances lightning occurs.
Once trained, the system made predictions that proved correct almost 80% of the time.
This is the first time that a system based on simple meteorological data has been able to predict lightning strikes through real-time calculations.
The method offers a simple way to predict a complex phenomenon.
WHY DO LIGHTNING LIGHT?
Lightning occurs when strong upward drafts in the air generate static electricity in large and dense rain showers.
Parts of the cloud are positively charged and others negatively charged.
When this charge separation is large enough, a violent discharge of electricity happens – also called lightning.
Such a discharge starts with a small area with ionized air that is hot enough to conduct electricity.
This small area grows into a forked lightning channel that can be several kilometers long.
The channel has a negative tip that drives loads to the ground and a positive tip that collects loads from the cloud.
These charges go from the positive end of the channel to the negative end of another during a lightning flash, causing the load to be released.
& # 39; Unlike some lightning warning systems based on data from lightning detection networks, the ML model now provides a tool for building a lightning scheme without prior or prior lightning data as a precursor to the imminent threat. In other words, it does not depend on the initial detection of the lightning to generate the alerts, & # 39 ;, the team shared in the newspaper.
& # 39; Instead, it uses lightning system data to label the archived data and thus train the model with historical data. & # 39;
& # 39; After training, the model does not need such data to predict the risk of lightning in future time windows. & # 39;
. (TagsToTranslate) Dailymail (t) sciencetech