Banca de DEFESA: AUGUSTO DE RUBIM COSTA GURGEL

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
STUDENT : AUGUSTO DE RUBIM COSTA GURGEL
DATE: 26/02/2024
TIME: 14:00
LOCAL: Videoconferência
TITLE:

DENSIDADE DE POTÊNCIA EÓLICA A PARTIR DE COMBINAÇÃO DE MULTIMODELOS CLIMÁTICOS REGIONAIS NO NORDESTE DO BRASIL PARA O PASSADO RECENTE E FUTURO 

KEY WORDS:

CORDEX-CORE. ECWMF-ERA5 reanalysis. Wind Power Density. Regression by Principal Components. Machine Learning.

PAGES: 135
BIG AREA: Ciências Exatas e da Terra
AREA: Geociências
SUBÁREA: Meteorologia
SPECIALTY: Meteorologia Aplicada
SUMMARY:

The significant increase in energy demand and the consequent increase in greenhouse gas emissions have meant that clean energy sources have become more widely explored across the globe. In Brazil, in particular, the development of wind and solar energy was inspired, mainly in the Northeast Brazil region (NEB). Thus, the general objective of this research was to combine high-resolution multi-models (ensembles) belonging to the CORDEX project using Average Means, Convex Combination, Principal Component Regression and Machine Learning techniques in order to reduce uncertainties and project more accurately the wind speed and wind power density for areas of the NEB under two climate scenarios for the near (2041-2060) and far (2080-2099) future. The thesis is organized in the format of articles. In the first article, it was evaluated how wind speed was related to other climatic variables. To this end, Principal Component Analysis (PCA) was used to correlate wind speed with air temperature, relative air humidity and precipitation from the Regional Climate Model (RCM) RegCM4.7_HadGEM2-ES for the Rio Grande do Norte, in the recent past from 1986 to 2005. As a result, CP1 presented 75.74% of data variability and eigenvalue above 1, representing high wind speed intensity, high temperature, low relative air humidity and precipitation during the months between August and December. Wind speed had a strong negative correlation with precipitation and relative humidity, -0.91 and -0.94, respectively. In the second article, the wind power density was calculated for areas of the NEB in the recent past from 1986 to 2005. Firstly, the ECMWF-ERA5 reanalysis was validated with the data observed from Xavier. From the analysis of the results, a moderate Pearson correlation coefficient was obtained, between 0.4 and 0.7, and bias between -1m/s and 1m/s (except for the austral winter) throughout the NEB. Then, the RCM RegCM4.7, RCA4 and Remo2009 were validated with the reanalysis, and overestimation values above 1m/s were observed in several areas of the NEB in the dry period (July to December). Due to the high climate variability of the NEB and its large territorial extension, in addition to the high resolution of the RCM, four areas were chosen for study. Based on the number of wind farms and different climatic aspects, the following were chosen: the north of Ceará (N-CE), the north of Rio Grande do Norte (N-RN), the Borborema plateau (Borborema) and the center of Bahia (CBA). In this way, the RCM were validated individually in comparison to the reanalysis of each area mentioned above. The RCM with the best statistical indices in each area were chosen to calculate the wind power density. N-RN and N-CE were the areas thar presented the highest wind power density in the recent past. Finally, in the third article, robust techniques were applied to the set of multi-models (ensembles) based on the arithmetic mean (a nonrobust technique, but widely used in the literature), convex combination, Principal Component Regression (PCR) and Machine Learning, for the same areas chosen in the previous article and the same period (recent past). The arithmetic mean and convex combination techniques showed a high underestimation of wind speed for the first half of the year, between 1.5 and 2m/s (except the C-BA area), and a high overestimation for the second half of the year, with values between 1,5 and 2m/s, in all areas of the NEB. The other techniques managed to represent well (close medians – values below 0,5m/s) the ECMWFERA5 reanalysis in all areas studied in the NEB. From the analysis of the Taylor’s Diagram and Willmott’s statistical agreement indices and the standard deviation ratio, the technique that best represented the reanalysis in all NEB study areas was PCR. Thus, using this technique, wind power density was calculated, both for the recent past and for the near and far future, using RCP2.6 and 8.5. A decrease in wind power density was observed for all areas when comparing futures with the recent past with the maximum decrease of 3.86% in N-RN for the far future by RCP8.5. It was therefore considered that there is stability in the values of wind power density for near and far future. Therefore, the N-RN area presented the highest wind power density in the NEB, followed by N-CE, Borborema and C-BA, both for the recent past and for the near and far future.

COMMITTEE MEMBERS:
Presidente - 1914304 - KELLEN CARLA LIMA
Externo à Instituição - DIEGO SILVEIRA COSTA NASCIMENTO - IFRN
Externo à Instituição - DOMINGO CASSAIN SALES - FUNCEME
Externo à Instituição - MARCOS SAMUEL MATIAS RIBEIRO - UFRA
Externa à Instituição - MAYTÊ DUARTE LEAL COUTINHO - INMET