A Data-Driven Methodology for Modelling Losses in HTS Power Systems

• Autor(a) principal: Simas, Henrique da Fonseca Borges Soares e
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10362/152026
• topic: AC Losses; Superconductor; Cryogenics; Neural Network; Data-Driven Model; Transformer; Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática

When designing a superconducting device one of the main obstacles is the AC losses. These losses created numerous difficulties, particularly in the design of the cryogenic system: the heat created from these losses must be removed in such a way that the cryogenic temperature is not affected, as to not change the materials state from superconductor to normal. Currently, most simulations of AC losses in superconductors are done using numerical methods, such as the finite element method. This type of simulation requires a significant amount of time and computational power. A data-driven model is proposed in this work to make determining AC losses in a superconducting device easier. A lock-in amplifier method of AC loss measuring is applied to superconducting coils and transformers, as well as a direct V–I method. With these results, an artificial neural network is constructed, trained and optimized in order to accurately predict AC losses in such devices. This approach is meant to determine AC losses quickly and without the requirement for significant computational power by using only a macro description of a device, such as the number of turns in a coil and the core size of the transformer. This work was developed in the ambit of the tLOSS project “Transforming Losses Calculation in High Temperature Superconducting Power Systems” (reference PTDC/EEIEEE/ 32508/2017_LISBOA-01-0145- FEDER-032508).