Serviços Personalizados
Journal
Artigo
Indicadores
- Citado por SciELO
- Acessos
Links relacionados
- Citado por Google
- Similares em SciELO
- Similares em Google
Compartilhar
Tecnura
versão impressa ISSN 0123-921X
Resumo
LOPEZ SEPULVEDA, Gloria Patricia; SEPULVEDA LONDONO, Christian David e MOLINA CABRERA, Alexander. Voltage stability margin estimation in a power system using artificial neural networks. Tecnura [online]. 2013, vol.17, n.37, pp.22-32. ISSN 0123-921X.
Abstract Power systems have become more sensitive to voltage stability problems due to load growth and the increasing utilization of alternative energy. Voltage stability appears not to be properly evaluated by methods of static analysis, such as load flux, because these methods use simple models to describe the system's components. Moreover, dynamic models for voltage stability analysis require high computational cost. Considering this, there is a need of developing more efficient and accurate methodologies. This article shows a model based on Artificial Neural Networks (ANN). ANN is a technique of artificial intelligence inspired in the biological neural networks of the human brain. An ANN is capable of learning through experience and takes the main characteristics of a dataset. Additionally, such networks may offer correct answers for inputs that present variations due to noise effects or distortions in the environment. ANNs provide a switching behavior as a function of the environment, that is, they get an input dataset and transform it into a consistent output set. This represents the rationale behind using ANNs for the development of this work; in other words, ANNs allow the prediction of a voltage-stability margin in power systems using information about their state variables. The training datasets are obtained from the load flux method, which delivers the initial conditions (i.e. the network input variables). Results are obtained through an estimating NN based methodology and compared to the results obtained when using estimating margin deterministic methods in a 4-node tests system.
Palavras-chave : voltage stability; load flow; energy barrier; neural networks.