Estimador de variações de tensão de curta duração em sistemas elétricos de potência utilizando estratégias evolutivas. / Estimate short duration voltage variation using evolutionary strategies.

AUTOR(ES)
FONTE

IBICT - Instituto Brasileiro de Informação em Ciência e Tecnologia

DATA DE PUBLICAÇÃO

19/09/2011

RESUMO

A methodology is herein proposed to estimate Short Duration Voltage Variation (SDVV) in electric power systems, caused by electrical faults. To attain this target, values recorded by measurement equipment in specific sites are used. In fact, there are few power quality meters installed in power networks, due to the high cost of such meters. Although these costs have decreased in recent years, the installation of a sufficient number of meters to ensure monitoring the entire network is still unfeasible. This reality poses a challenge to developing techniques that, with a small number of monitoring points, allow the determination of fault locations and estimation of SDVV values in specified buses. As contribution this thesis proposes an algorithm called Evolutionary Strategies (ESE), which integrates the group of evolutionary algorithms. This algorithm can be easily implemented and finds a solution within a wide solution space. The ESE determines the fault location, fault type and fault impedance, that characterize an individual, so that the resulting voltages on monitored buses are as close as possible to the measured ones. An initial population is generated as alternative solutions to the problem. Some of the individuals in the population will be submitted to mutation and recombination operators. Individuals are then selected to the future generation. An iterative process is carried out to determine a solution as close as possible to the desired one. Each individual is evaluated by the objective function, which represents the quadratic error between the measured and calculated values. This calculation is based on short circuit calculation related to the evaluated individual and from information of voltage values gathered from power quality meters. Voltage values in specific network buses can then be determined to monitor their corresponding SDVV values. This allows, for example, determining quality indicators associated to the SDVV, such as the System Average RMS Frequency Index (SARFI), to evaluate sensitive areas, i.e. which are prone to cause SDVVs and to develop plans for preventive maintenance. Two algorithms that calculate the minimum number of meters and their locations have been implemented. The first algorithm aims to ensure monitoring the entire network regarding SDVVs, while the second algorithm ensures the smallest error of SDVV estimation in buses where no meters are installed. This methodology can be applied to meshed or radial networks. It was initially implemented in small networks (IEEE 14 and 30 buses) with the purpose of verifying the ability of algorithm. In sequence the methodology was applied to more complex networks (IEEE 57 and 118 buses). To assess the efficiency of the methodology a comparison with other optimization methodology based on Genetic Algorithms (GA) was carried out.

ASSUNTO(S)

alocação ótima de medidores busca em vizinhança variável electrical faults estimador de variações de tensão de curta duração estratégias evolutivas evolutionary strategies faltas elétricas fault location state estimate short duration voltage variation variable neighborhood algorithm

ACESSO AO ARTIGO

Documentos Relacionados