Identification of topology errors with use of unbalance indices and neural networks

Abstract

The paper deals with identification of topology errors, i.e. identification of incorrect modelling of the physical connections of a power system, which is one of the most important tasks of the real-time modeling of a power system. The paper presents the method, based on utilization of introduced-byauthors so-called unbalance indices and artificial neural networks. Values of the unbalance indices create characteristic sets for different cases of modeling of branches and nodes of the considered power system. In the paper, analyses of the mentioned sets is made. The analysis allows to point out these cases of topology errors, which can be distinguished by examination of values of the unbalance indices. The analysis gives bases for determination of the cases, for which the considered method is efficient. The presented considerations are illustrated by the case study. At the end, the features of the method are summed up.

Enhancement of network monitoring & security analysis using phasor measurement units

“…Available processors in power system are only capable of obtaining the topology of the local system, utilizing the local measurements [106], [107]. It is well documented that knowledge of internal system about the connectivity information of the external system in an inter-connected network is very important for the system operation [6].…”

“…[95], [98], and [99] use rule-base or artificial neural network method. [97], [100], [101], and [107] proposed method based on least absolute value state estimation. [102], and [103] use geometrical based hypothesis testing for topological error identification.…”