RESEARCH 002

Research Archive

Active research combining scientific machine learning with solar observations from the Aditya-L1 mission.

FEATURED RESEARCH

STATUS: ACTIVE

Physics-Informed Neural Network for Explainable Solar Flare Nowcasting and Forecasting Using SoLEXS and HEL1OS Data from Aditya-L1

MISSIONAditya-L1 / Solar Physics
DATA SOURCESoLEXS + HEL1OS
OBJECTIVESolar flare nowcasting and forecasting
RESULTIn progress

METHOD

  • Physics-Informed Neural Networks
  • Time-Series Forecasting
  • Explainable AI
  • Uncertainty Quantification

PIPELINE

Data flow

  1. 01ADITYA-L1
  2. 02SoLEXS + HEL1OS
  3. 03Data Processing
  4. 04Feature Engineering
  5. 05Physics-Informed AI
  6. 06Explainable AI
  7. 07Forecast + Uncertainty
  8. 08Solar Flare Risk Assessment

INSTRUMENT 01

SoLEXS

Solar Low Energy X-ray Spectrometer

Measures the soft X-ray spectrum of the Sun, tracking how coronal plasma heats and cools across a flare.

INSTRUMENT 02

HEL1OS

High Energy L1 Orbiting X-ray Spectrometer

Observes higher-energy, hard X-ray emission produced during the impulsive phase of a flare.

RATIONALE

Why soft and hard X-rays are read together

Soft X-rays describe the thermal response of the solar atmosphere: how much plasma has been heated, and how that heat decays. On their own they mostly tell you what has already happened.

Hard X-rays trace accelerated, non-thermal electrons and typically peak earlier. Combining both bands gives an earlier and physically richer signal for nowcasting, and lets a physics-informed model reason about energy release rather than fitting a single light curve.