This paper is concerned withthe optimal operation of reservoirs for electric generation. It reports several tests carried out on a simple system under special conditions highlighting some important characteristics of their optimal behavior. The tests were carried out in a progressive way in order to make clear the influence of several factors including water head, discount rate, inflow seasonality and system design. The main conclusions have been verified on an actual large scale system, which demonstrated the conservative nature of the optimal operation of reservoirs and the different roles they play according to their relative position in the cascade. The conclusions address the issue of adequately modelling the long term hydrothermal scheduling problem. The results suggest that the assumption of parallel operation of reservoirs, implicit in composite reservoir models and some simulation heuristics, could be a modelling oversimplification.
This paper presents an application of a deterministic optimization algorithm in the hydrothermal scheduling of the large scale Brazilian south-southeast interconnected system, composed of 51 hydro and 12 thermal plants, corresponding to 45 GW of installed capacity. The application considers the system operational conditions according to the 1986 Operational Plan coordinated by ELETROBRAS -the Brazilian electric holding company.The employed algorithm is based on a network flow approach especially developed for hydrothermal scheduling. For the south-southeast interconnected system the problem formulation suggests a primal decomposition optimization approach. Two applications have been performed: (1) in a study context the system has been evaluated for the most critical stream flow ever recorded; (2) in an operational context the system has been simulated through an adaptive operational planning approach. The results show that the employed algorithm can be a useful tOQl for large scale hydrothermal systems, giving insight into operational characteristics and/or adding to the guidance of real system operation.
-A type of recurrent neural network has been proposed by H. Jaeger. This model, called Echo State Network (ESN), possesses a highly interconnected and recurrent topology of nonlinear processing elements, which constitutes a "reservoir of rich dynamics" and contains information about the history of input or/and output patterns. The interesting property of ESN is that only the memoryless readout is trained, whereas the recurrent topology has fixed connection weights. This reduces the complexity of recurrent neural network training to simple linear regression while preserving a recurrent topology. In this paper, the ESN is used to forecast hydropower plant reservoir water inflow, which is a fundamental information to the hydrothermal power system operation planning. A database of average monthly water inflows of Furnas plant, one of the Brazilian hydropower plants, was used as source of training and test data. The performance of the ESN is compared with SONARX network, RBF network and ANFIS model. The results show that the Echo State Network provides pretty good results for one-step ahead water inflow forecasting, providing a valuable information for the system operator.
Este trabalho apresenta um estudo comparativo entre ferramentas de análise, aplicável à Qualidade da Energia Elétrica (QEE), enfatizando-se a Transformada de Fourier com Janela (TFJ), a Transformada Wavelet (TW) e Redes Neurais Artificiais (RNAs). Das ferramentas apontadas, a TFJ e a TW, mostram-se aplicáveis à detecção, localização e classificação de distúrbios agregados às formas de ondas de tensão em um sistema de distribuição, com o intuito de prover um diagnóstico preciso das situações enfrentadas. Como será evidenciado, além da detecção, localização e classificação pelas técnicas citadas, os distúrbios também podem ser classificados segundo sua natureza, utilizando-se métodos alternativos, como pela aplicação de RNAs. Os testes efetuados mostraram que as ferramentas mencionadas possuem uma grande potencialidade quanto às suas aplicações na avaliação da QEE. Neste contexto, serão apontadas algumas peculiaridades e características inerentes a cada ferramenta. This work presents a comparative study amongst tools for the analysis of Power Quality, emphasizing the Windowed Fourier Transform (WFT), the Wavelet Transform (WT) as well as Artificial Neural Networks (ANN). From the tools mentioned, the WFT and WT are applicable to the detection, location and classification of abnormalities related to voltage waveforms in a distribution system for the diagnosis of the present situation. As it will be shown, the classification of the phenomena can also be performed using alternative methods, as the ANN. Tests show that the mentioned tools have great potentiality to be applied to the evaluation of Power Quality. Some peculiarities of each tool will be emphasized
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.