Conference Paper

Proceedings of BSA Conference 2024: Sixth Conference of IBPSA-Italy

Impact of Different Radiation Decomposition Models and ERA5 Dataset on Building Energy Simulation Results: A Case Study in Brazil

Matheus K. Bracht, Matheus S. Geraldi, Ana Paula Melo, Roberto Lamberts

DOI: https://doi.org/10.13124/9788860462022_56
Abstract: This study compares three different radiation decomposition models (Erbs, DIRINT, and DISC) for estimating direct normal radiation to the data from the reanalysis dataset ERA5. It also evaluates the impact of considering these different models and datasets on building energy simulation outcomes for three locations in Brazil (Brasília, Salvador, and São Paulo). As the simulation study case, we analyzed a typical residential building in the Brazilian context. This building model was analyzed in three different cases (Brazilian standard building characteristics reference, low solar absorptance values, and considering a 0.80 m overhang). Regarding radiation datasets, the Erbs model exhibited the lowest RMSE for direct and diffuse radiation compared to the monthly values provided by the Brazilian Solar Atlas. By analyzing the RMSE values, we demonstrated that ERA5 overestimated direct normal radiation while significantly underestimating diffuse values compared to the Solar Atlas. Concerning simulation results, we observed differences of up to 14% higher cooling load values when comparing the results using ERA5 data with the DISC model ones. However, maximum operative temperatures did not show such significant differences, with a maximum deviation of 1%. Also, the three cases tested demonstrated the sensitivity of the building simulation to the different radiation datasets. These results are important for advancing the understanding of the impacts of using reanalysis datasets, which is becoming an increasingly common approach.
Pages: 461 - 468
Paper: