Application of chaotic simulation and self-organizing neural net to power system voltage stability monitoring

This paper introduces a chaotic neural net model to calculate the multiple load flow solutions, especially the lower voltage solution for power system voltage stability monitoring. Chaos is now understood to be an inherent feature of many nonlinear systems. Unlike Lyapunov dynamics, the proposed neural net aimed at dealing with global optimization problems, is based on the chaotic dynamics regime which allows neural networks to be temporarily unstable, keeping stability due to convergent dynamics. Therefore, by converting the load flow problem into an energy-minimum problem and taking advantage of 'chaotic itinerary', multiple load flow solutions can be obtained. Numerical calculations have been undertaken in this paper, where a number of fractual structures of orbit and Poincare maps plotted with varying phases were provided to certify chaos occurrence, and a practical power system was also used to show the efficiency and effectiveness of the proposed approach.<>