Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 1
Encontrada(s) 1 página(s)

O2, repositorio institucional de la UOC
Publication Contribución de congreso (ConferenceOutput)

Physics-Informed Neural Networks for Enhanced Thermal Regulation in a Spacecraft

  • de Curtò y Díaz, J.
  • de Zarzà i Cubero, I.
Advancements in Artificial Intelligence (AI), particularly in Machine Learning (ML), are increasingly being utilized to address complex challenges in spacecraft technologies. One of the fundamental areas of focus is the thermal regulation system, which is crucial for the optimal functioning of the spacecraft. This paper introduces an innovative approach that employs Physics-Informed Neural Networks (PINNs) to model and predict temperature distributions within spacecraft. Unlike traditional neural networks, PINNs incorporate known physical laws, in this case, the heat equation, to ensure that the model’s predictions are both accurate and physically consistent. We present a detailed methodology for developing a PINN that can accurately predict temperature variations over time, using sparse and noisy temperature data from a spacecraft. The proposed model is trained using a combination of data-driven and physics-informed loss functions, ensuring that the predictions adhere to the underlying physics of heat transfer. The results demonstrate the efficacy of PINNs in improving the accuracy of thermal predictions, which can significantly enhance the performance of the spacecraft’s thermal regulation system. This work paves the way for further integration of physics-informed ML models in various aspects of spacecraft design and operation, offering a robust framework for tackling complex physical phenomena in space exploration.