Adaptive Technique to Solve Multi-objective Feeder Reconfiguration Problem in Real Time Context

This paper presents an innovative method to solve the reconfiguration problem in a distribution network. The main motivation of this work is to take advantage of the power flow analysis repetition when reconfiguration leads the network to a previous configuration due to cyclical loading pattern. The developed methodology combines an optimization technique with fuzzy theory to gain efficiency without losing robustness. In this methodology, the power flow is estimated by well-trained neo-fuzzy neuron network to achieve computing time reduction in the evaluation of individuals during evolutionary algorithm runs. It is noteworthy that the proposed methodology is scalable and its benefits increase as larger feeders are dealt. The effectiveness of the proposed method is demonstrated through examples. The overall performance achieved in the experiments has proved that it is also proper to real time context.