Cellular computational generalized neuron network with cooperative PSO for power systems

摘要

To enhance the reliability of the power system, newer technologies are being incorporated day by day. Predictions of different states can significantly increase the reliability of the system. To predict the frequency, a cellular computational generalized neuron network (CCGNN) that is trained with particle swarm optimization (PSO) is proposed recently. However, with the size of the system, the dimension of PSO grows that in turn, increases the complexity of the training. To solve the problem, a special version of PSO named cooperative PSO (CPSO) is applied for training the CCGNN in this paper. Through simulation on a 68-dimensional problem of a two-area four-machine system, it is shown that the CPSO performs significantly better than the canonical one.