Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 3
Encontrada(s) 1 página(s)

Repositori Obert UdL
Publication Contribución de congreso (ConferenceOutput)

Improved modeling of thermal energy storage tanks state of charge Using LSTM neural networks and image segmentation

  • Mateu Piñol, Carles
  • Rojas Cala, Edgar Felipe
  • Béjar Torres, Ramón
  • Borri, Emiliano
  • Cabeza, Luisa F.
This study presents a deep learning framework designed to accelerate the modeling of sensible heat thermal energy storage (TES) tanks, addressing the high computational costs associated with traditional computational fluid dynamics (CFD). By integrating long short-term memory (LSTM) neural networks with image segmentation, the authors developed a method to predict the thermal state of charge across various tank configurations characterized by different internal arrangements of solid storage spheres. A dataset comprising over 7 million thermal images from 2D CFD simulations was used to train and compare several architectures, including Convolutional Autoencoders and Inception-based models. The results demonstrate that an Autoencoder-LSTM architecture, particularly when segmentation data is introduced after the recurrent layer, yields the highest accuracy with a Structural Similarity Index (SSIM) of 86.7%. While the approach offers a rapid and viable alternative for evaluating diverse tank geometries, the study also identifies challenges regarding cumulative errors in long-term sequence predictions, suggesting future directions for optimizing AI-driven thermal analysis., This project was funded by the European Union Horizon Europe Research and Innovation Pro-gramme under grant agreement 101103552 (SUSHEAT). Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or CINEA. Neither the European Union nor the granting authority can be held responsible for them. This paper is part of the RYC2023-044196-I, funded by MCIU/AEI/10.13039/501100011033 and FSE+. The authors thank the Generalitat de Catalunya for the quality accreditation granted to their research group (2021 SGR 01615). GREiA is a cer-tified agent TECNIO in the category of technology developers of the Government of Catalonia. This work is partially supported by ICREA under the ICREA Academia programme.
Repositori Obert UdL
Publication Contribución de congreso (ConferenceOutput)

Heat transfer enhancement of thermal energy storage systems using phase change materials: experimental work

  • Megra, Kassahun Tadesse
Growing energy demand and environmental concerns are accelerating the shift to renewable sources. Solar and other renewables need efficient storage due to their intermittent nature. TES systems using phase change materials (PCMs) are promising for storing heat at constant temperatures. This research enhances TES performance by optimizing heat transfer and evaluating PCMs under real conditions. The results aim to support efficient and sustainable integration of renewable energy., This project was funded by the European Union’s Horizon Europe Research and Innovation Programme under grant agreement HYBRIDplus (101084182) and SUNSHINE (101103552). Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or CINEA. Neither the European Union nor the granting authority can be held responsible for them. This paper is part of the RYC2023-044196, funded by MCIN/AEI/10.13039/501100011033 and FSE+, and Agencia Estatal de Investigación (AEI) (PID2021-123511OB-C31 – MCIN/AEI/10.13039/501100011033/FEDER, UE). This work is also partially supported by ICREA under the ICREA Academia programme. The authors would like to thank the Departament de Recerca i Universitats of the Catalan Government for the quality accreditation given to their research group (2021 SGR 01615). GREiA is a certified TECNIO agent in the category of technology developers of the Government of Catalonia. This project is also co-funded by the European Union’s research and innovation programme Horizon Europe under the project NitRecercat2425 (GA 101162003).
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Improved modeling of thermal energy storage tanks state of charge Using LSTM neural networks and image segmentation

  • Mateu Piñol, Carles
  • Rojas Cala, Edgar Felipe
  • Béjar Torres, Ramón
  • Borri, Emiliano
  • Cabeza, Luisa F.
This study presents a deep learning framework designed to accelerate the modeling of sensible heat thermal energy storage (TES) tanks, addressing the high computational costs associated with traditional computational fluid dynamics (CFD). By integrating long short-term memory (LSTM) neural networks with image segmentation, the authors developed a method to predict the thermal state of charge across various tank configurations characterized by different internal arrangements of solid storage spheres. A dataset comprising over 7 million thermal images from 2D CFD simulations was used to train and compare several architectures, including Convolutional Autoencoders and Inception-based models. The results demonstrate that an Autoencoder-LSTM architecture, particularly when segmentation data is introduced after the recurrent layer, yields the highest accuracy with a Structural Similarity Index (SSIM) of 86.7%. While the approach offers a rapid and viable alternative for evaluating diverse tank geometries, the study also identifies challenges regarding cumulative errors in long-term sequence predictions, suggesting future directions for optimizing AI-driven thermal analysis., This project was funded by the European Union Horizon Europe Research and Innovation Pro-gramme under grant agreement 101103552 (SUSHEAT). Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or CINEA. Neither the European Union nor the granting authority can be held responsible for them. This paper is part of the RYC2023-044196-I, funded by MCIU/AEI/10.13039/501100011033 and FSE+. The authors thank the Generalitat de Catalunya for the quality accreditation granted to their research group (2021 SGR 01615). GREiA is a cer-tified agent TECNIO in the category of technology developers of the Government of Catalonia. This work is partially supported by ICREA under the ICREA Academia programme.