Data: 04/11/2024 à 07/11/2024
Local: Florianópolis-SC
Mais informações: https://www.abrhidro.org.br/iebhe
Design Flood from Annual and Partial Series - the case of HPP 14 de Julho, South Brazil
Código
I-EBHE0199
Autores
Ailton João Gonçalves Moreira, Júlia Mella Massing, João Marcos Carvalho, ANAXIMANDRO STECKLING MULLER, LIU MING
Tema
WG 1.10: Hydrologic Design - Solutions & Communication
Resumo
In the context of climate change, concerns about extreme floods are increasing worldwide. This issue has gained more attention in Brazil, particularly in the state of Rio Grande do Sul in the south of the country, due to recent flood events in the Taquari Basin. These events affected the Hydroelectric Power Plant (HPP) 14 de Julho, causing a partial failure of the dam spillway. With the increasing frequency of extreme rainfall events associated with high return periods, concern grows about the quantity (time-series length) and quality of basic data, crucial for sizing spillways during dam safety studies. Considering the quantity of the data, discussions were carried out considering the use of annual and/or partial historical flow series. The focus of these discussions addressed two key questions: (i) the appropriate choice between the use of annual series or partial series, and (ii) determining the adequate discharge threshold flow and the number of probability exceedances to select floods for the partial series method. Partial series are less commonly used in projects due to their inherent complexity. However, they seem to be more sensitive to detect critical events of greater frequency. Unlike annual series, they consider the largest events in the historical series above the threshold flow. This helps to consider a greater frequency of certain critical events in statistical adjustments for extreme flows. To compare both methods, historical data from the National Electrical System Operator's (ONS) daily average reconstituted natural flows were used. A visual analysis of the daily discharge values was conducted by plotting the data, which spans over 93 years starting from 1931 and includes recent events up to the last hydrological year. For annual analysis, the annual maximum daily average flow was determined, resulting in a dataset of 93-length data. To define partial series, the threshold flow was set at half the average of the annual maximum daily flows resulting in 1,170 m3/s. A probability of exceedance of 2 values was used within the recommended interval values to ensure independence between the events chosen. The resulting partial flow series consists of 182-length data. The Yue and Wang (2004) Mann-Kendall modification test was used to assess the stationarity of the resulting series, indicating no trend for both annual (p < 0.174) and partial (p < 0.059) maximum daily flows. Frequency analysis (FA) was conducted on both series as well, using generalized extreme values (GEV) and considering a return period of 10,000 years. The 95% and 5% confidence intervals were estimated using Monte Carlo and Markov Chain Monte Carlo (MCMC) simulations. The maximum designed discharge of the HPP from the annual series was found to be 30,892 m3/s, while from the partial series, it was 23,922 m3/s. The results did not demonstrate a significant difference when considering the confidence intervals. The study presented an effective method for analyzing the maximum designed discharge, using both annual and partial series. This contributes to the hydraulic sizing and the reconstruction of the HPP 14 de Julho?s spillway.