RBF-BASED EFFECTIVE LENGTH OF GROUNDING VERTICAL RODS UNDER HIGH-VALUED IMPULSE CURRENTS

In this paper, using neural networks based on radial basis functions (RBF), closed-form solution for effective length of grounding vertical rod is extracted in such a way that for the first time the ionization of soil is considered. In creating the model, training data are computed from multi-conductor transmission line (MTL). As a results, firstly in the proposed model, in despite of previous models ignoring ionization, ionization is included. Secondly, the results are in excellent agreement with the MTL. The achieved results for effective length show that considering ionization results in decreasing the effective length with respect to situation where it is ignored. It is well known that this is financially of importance.