AI Helps Solar Scientists Solve Problems

AI helps solar scientists with a hard to solve problem.

The sun affects us profoundly.

Without the sun, we perish.

But more commonly, when the sun acts up, we are affected. Those coronal mass ejections cause electromagnetic flare-ups that can disrupt communication, affect the moods of people and animals, disrupt power grids, air travel, and more.

After Quebec Canada lost power due to solar flares in 1989, NASA launched a solar satellite, the SDO, or Solar Data Observatory, to measure UV rays.

The intent was to attempt to predict when flares would occur, in the hopes that preparation or technology could avoid the possible $2Trillion that a direct solar flare could cost in damages.

The SDO satellite has three major instrument components:

Atmospheric Imaging Assembly (AIA) – captures images of the solar atmosphere in multiple wavelengths (up to 10) for every 10 seconds in IMAX resolution (x10 times the precision of HD images). In other words, this measures what is happening in the sun’s atmosphere.

EUV Variability Experiment (EVE) – measures the solar extreme ultraviolet radiation (EUV) to understand the influence on earth’s (and near-earth space’s) climate changes.

Helioseismic and Magnetic Imager (HMI) – studies the oscillations and the magnetic field at the solar surface, or photosphere.

Andy Thurai, Emerging Technology Strategist

Sadly, one of the components of the SDO satellite broke down in 2014, and prevented the daily data collection of 1TB to be performed.

AI Helps Solar Scientists Solve The Problem

Luckily, AI was at a point where something could be done.

With 5 years of data, there was enough historical information to run training sets, and see if an AI program could be created that fit the data accurately enough to be useful.

the scientists wanted to analyze the superior images of the sun generated by AIA and predict the EUV radiation measurements….

Impressively, the engineers took to the task using only common software and hardware tools: Jupyter notebook, PyTorch, NVIDIA GPUs, IBM Watson AI, and the cloud to host this all. Each one of the tools was chosen for a specific reason. Jupyter notebook is the easiest way for engineers to collaborate. NVIDIA is the best GPU available today.

Andy thurai, ETS, ORACLE CORP

The training results achieved a match for 97.5% of the data, enough to be useful to NASA. Now, the Atmospheric Imaging Assembly images can be processed by AI and generate AIA data for NASA useful enough to calculate current solar outputs. The next step is “can we now predict future EUV spectra accurately?

Scientists are looking to see if this technique for filling in missing data can be used in other areas, like IOT sensor malfunctions.

Read Andy’s full article here, and learn more about Oracle GPU computing here.

Related: See Spark and Tensorflow develop deep learning.

Be First to Comment

    Leave a Reply

    Your email address will not be published. Required fields are marked *