Abstract
In this paper, in order to determine the optimal radial power system structure rapidly I propose to use a chaotic neural network model with the constrained noise approach which can obtain the global minima. When a radial power system has a number of connected feeders, the combinatorial number for possible system structures becomes too huge. Determination of the optimal system structure from a great number of combinations becomes a combinatorial optimization problem, and it has been difficult to solve this type of problem quickly with conventional algorithm so far. However, with the chaotic neural network model, it has become possible to carry out the optimization efficiently. Furthermore, in order to operate the radial power system more securely, I employ a network switching scheme so as to obtain a more practical system structure considering cases of fault occurrence at each substation. The technique is demonstrated with an actual 21 substation system. The proposed method is compared with a Hopfield neural network model approach.
