Predict solar storms This is important, since measures can be taken before an interruption in electricity and telecommunications to cause problems. There are already some methods to make such forecasts. However, they are not too effective. Or they have not been so far because they were Artificial intelligence (AI) developed for this purpose by a team of scientists from New York University Abu -Dabi She is very promising about it.
Capable of predicting sunny storms with 4 days in advanceThe field besides, it does it with 45 % accuracy than current methods. Even with an accuracy of 20 % more than other models based on artificial intelligence.
In the future, perhaps the process can be additionally improved, and AI with greater accuracy. However, with the help of these results, you can already make very useful forecasts to make appropriate decisions. For example, you can Avoid the loss of the satellite Or warn the population of possible areas where the electric network is affected.
What are solar storms?
The sun is a star consisting of many elements. First of all, there is hydrogen, as well as others, such as oxygen, carbon, neon or iron. They are all in the form of plasma, The state of the material, similar to gas, but with loaded particles. These charged particles were found In the movement around the sunSometimes with excellent activities, and in other cases much calmer. This activity is periodically Cycles are about 11 years old. In each solar cycle, the activity of the star king increases until it reaches even the maximum point, after which it will decrease.
When solar activity is very high, the exile of charged particles can be obtained in the form Mass or magic ejaccationThe latter is much more intense than the former, and are usually also accompanied by the release Great plasma cloudThe field as a whole, all these processes make up solar storms, which in some cases They can affect the Earth. If the sudden expulsion of loaded particles goes in the direction of our planet, this may affect the proper functioning of satellites, as well as Electric and telecommunication networks.
How can solar storms be predicted through AI?
To get AI that can predict solar storms, these scientists have trained an algorithm with two data types. On the one side, High resolution ultraviolet images Tedled by the observatory of NASA solar dynamics. And on the other hand, Historical records On solar storms that reached the Earth at some point.
Thus, AI discovered templates that connect two data sets, and determined which parameters of the Sun usually lead to storms on our planet. This has been done, having only one data type, the other can be predicted by the association. This is what the AI algorithms usually do.
At the moment, the results were very good. Solar storms can be predicted with great accuracy, Until 4 days before they happen. This is very useful, because we can usually have an idea that they will occur in the peaks of the cycle of solar activity; But, in addition, it is difficult to make forecasts. If this algorithm is widely used, it can be easier to get ahead of the rage of the sun. Nothing can be done to avoid this, but at least it takes action.
Source: Hiper Textual
