Publicação
Analytical and numerical thermal modelling of a low power transformer
| Resumo: | In past decades, thermal modelling of electric machines was overlooked and based on empirical experience. Due to modern requirements, using novel materials and innovative configurations, the thermal model became a crucial component in the modern design process. The thermal modelling can be performed by analytical methods, as the thermal resistance network (TRN) or by numeric methods as the finite element analysis (FEA). This work presents a methodological approach to thermal modelling, addressing both approaches and applying them to a study case of a low power shell-type single-phase transformer. The results from both approaches are compared with experimental results, achieving relative errors of 5.70% and 21.24% on the determination of windings' temperature, for the FEA and TRN model respectively, which helps define model improvements. |
|---|---|
| Autores principais: | Mendes, Gabriel Dias |
| Outros Autores: | Ferreira, Ângela P.; Miotto, Ednei |
| Assunto: | Thermal resistance network Finite element analysis Thermal model Electric machine |
| Ano: | 2020 |
| País: | Portugal |
| Tipo de documento: | comunicação em conferência |
| Tipo de acesso: | acesso restrito |
| Instituição associada: | Instituto Politécnico de Bragança |
| Idioma: | inglês |
| Origem: | Biblioteca Digital do IPB |
| Resumo: | In past decades, thermal modelling of electric machines was overlooked and based on empirical experience. Due to modern requirements, using novel materials and innovative configurations, the thermal model became a crucial component in the modern design process. The thermal modelling can be performed by analytical methods, as the thermal resistance network (TRN) or by numeric methods as the finite element analysis (FEA). This work presents a methodological approach to thermal modelling, addressing both approaches and applying them to a study case of a low power shell-type single-phase transformer. The results from both approaches are compared with experimental results, achieving relative errors of 5.70% and 21.24% on the determination of windings' temperature, for the FEA and TRN model respectively, which helps define model improvements. |
|---|