An evolutionary computational approach for the identification of distribution networks models

In this paper, we present a novel methodology for generating low-voltage distribution network models. The methodology is based on leveraging the existing knowledge of the network topology and a comprehensive catalog of the conductors that are installed in each segment of the grid. By using genetic algorithms and data obtained from the smart-meters in the network, the proposed method can produce highly accurate network models. The effectiveness of this methodology has been confirmed by extensive simulation studies achieving errors of less than 0.4 V in the estimation of nodal voltages in scenarios without measurements noise and on the order of the standard deviation of the error considered in the measurements when disturbances are added to the problem. • Introduce a methodology based on genetic algorithms for obtaining low voltage network models. • Build the model from the data available to the distribution companies, trying to bring the problem closer to reality. • Obtain a network model that relates the input and output data of the network with enough accuracy.