Use of geological data to determine soil resistance with decicson tree classification algorithm

Grounding, which can be defined as connection of electrical devices to grounding system with a conductor for any possible electric leakage danger, is a quite simple but extremely significant protection system. In order for this system to enable expected protection, electrical resistance of the device to be grounded from the body to earth should be low enough. This electrical resistance basically consists of resistance of grounding conductors and ground transresistance of grounding conductors. Since resistance of grounding conductors is negligible, main significant resistance in the system is grounding transresistance of grounding conductors. In evaluating ground transresistance, geological structure of ground and variable ground spesific resistance are of great importance. For this reason, if possible, necessary measurements should be made to determine required ground spesific resistance. In our country, however, designs are made considering ground spesific resistance as 100Ω without making any measurements in most practices. In Turkey, following Kocaeli earthquake in 1999 use of geological data was made obligatory during construction of buildings having power plant. In our study, without making measurement we applied data mining in order to determine ground spesific resistance most accurately using geological data. Geological data- ground spesific resistance relation was modelled with decision tree classification algorithms in Weka software environment and ground spesific resistances were estimated using this model. Estimation results yielded significant outputs. We expect this study to be pioneer for more widespread practices enhancing field diversity.