AN ARTIFICIAL NEURAL NETWORK FOR PREDICTING THE RESISTANCE OF DEEP GROUND RODS

Operational Safety Dept., National Center for nuclear safety

Determination of effective earth resistivity for use in deep ground rod resistance calculations is an

important contribution to ground design. An Artificial Neural Network (ANN) algorithm for

predicting the resistance of deep driven ground rods has been developed. The ANN has been trained

on the data obtained from field ground rod test results. The algorithm extrapolates known test ground

rod data up to seven times the known depth values. This algorithm can be used to predict the

practical limits for driving deep rods, and thus eliminate the many extra rods and valuable

construction time spent installing deep rods.