High-Level Project Summary
The Carrington event was the most intense geomagnetic storm in history. It peaked on September 1 to 2, 1859, during solar cycle 10. Our goal is to develop a neural network for the FC instrument of the DSCOVR spacecraft to accurately track the variation of the maximum solar wind speed so that DSCOVR can provide early warnings to prevent disasters such as the Carrington event. In fact, the objective is to determine for each current vs voltage spectrum the solar wind ion density, thermal velocity and velocity vector, these parameters will help to determine the peak tracking.
Link to Final Project
Link to Project "Demo"
Detailed Project Description
First, we collected the magnetic field data Wind and DSCOVR, BW(t) and BD(t) for a time period suitable for training. Next, we derived DTW of the two databases, dynamic time warping (DTW) is an algorithm to measure the similarity between two sequences that can vary over time. Subsequently, we collated the database of Wind ion parameters and then selected the time series from the Wind ion parameters (density, n(D(t)); temperature, w(D(t)); velocity, v(D(t))) for periods where Pτ( BW(D(t)), BD(t) ), a correlation measure, is close to 1. Finally, we will use this data to train our ML model in order to predict the solar wind peaks
Space Agency Data
In our project, we used different datset from the nasa open source website :
- Magnitude field database for both DSCOVR and Wind
- Wind spacecraft ion parameters (ion density, thermal velocity, velocity vector)
- The DSCOVR spectra dataset
Hackathon Journey
During the Hackathon, we learned a lot of skills that helped us develop many competencies, such as teamwork, determination to achieve results and time management. In this sense, we developed our skills in big data analysis and preprocessing. On top of that, we discovered new mathematical tools such as DTW and other computational tools in the Tenserflow library. The SPace apps challenge is an interesting experience to discover the world of space and create a new generation capable of inventing the world of tomorrow.
References
Notional DSCOVR Faraday Cup Instrument “Calibration” and Data Analysis Procedure: https://www.spaceappschallenge.org/space-apps-challenge-2022-example-resource-save-the-earth-from-another-carrington-event/
Dynamic Time Warping : https://towardsdatascience.com/dynamic-time-warping-3933f25fcdd
Data : https://cdaweb.gsfc.nasa.gov/
Implementing Artificial Neural Network training process in Python : https://towardsdatascience.com/dynamic-time-warping-3933f25fcdd
Tags
#Carrington