Using self-organizing map in identification of load-loss state

Abstract

This paper presents a self-organizing map (SOM) based method for power system load-loss state classification. This classifier maps vectors of an N-dimensional space to a 2-dimensional net in a nonlinear way while preserving the topological order of the input vectors. Input features to SOM are real and reactive power at each load bus and available real power generation at each generation bus. After the training of the SOM, the generalization capability of the SOM can cope with various operating conditions which have not been encountered during the training phase and hence give a correct classification result. The effectiveness of the proposed method has been demonstrated on a 9-bus test system. This proposed method is useful for power system operation, power system reliability assessment and state screening.