Clustering of voltage control areas in power system using shuffled frog-leaping algorithm

Determination of critical voltage control areas (VCAs) is a very important task in both voltage stability assessment and control. However, it is impossible to detect VCAs in real time during appearance of emergency cases for large-scale power systems. Therefore, it is a reasonable solution to employ an artificial intelligent system (AIS) to detect VCAs and to identify the prone buses for monitoring and control purposes as quickly as possible. The training data must contain the simulation results and the historical data collected from a wide range of various emergency cases. Using this database, a clustering process which provides finite clusters of all possible VCAs and a classification function which affiliates each emergency or stress case to its own cluster of VCAs are the main stages to prepare AIS for automatic VCA identification. In this paper a novel data clustering method based on shuffled frog-leaping algorithm (SFLA) is presented for the first task. The results are finite numbers of clustered groups of VCAs with a representative vector of participation factors (PF) for each group. SFLA combines the benefits of the genetic-based memetic algorithms as well as social behavior-based particle swarm optimization methods. In present study the application of SFLA in data clustering is also compared with the most popular analytic algorithm of clustering, K-means, and also with genetic algorithm-based data clustering to demonstrate the validity of proposed clustering method. Numerical results are also presented for IEEE 14-bus test system and an artificial database. The comparative results show the effectiveness of proposed algorithm.