The disastrous Izmit 1999 earthquake besides causing loss of life, has also produced significant damage to the engineering structures. In this study, the effect of underlying soil to the structural damage is investigated without taking into account structural effects that caused damage to structures. Critical soil data inputs is used to investigate the effect of soil conditions on damage potential of the existing buildings in the cities Kocaeli and Sakarya during 1999 earthquake. Artificial neural network method is used to determine the damage potential of the buildings. Data derived from around 259 borehole logs were used as input dataset to artificial neural network model. Predicted output is a ratio obtained from structural damage which gives an insight on the potential structural damage. Analysis showed that best success is reached with 4 layer backpropagation neural and general regression neural network. The results of this study indicate that the damage potential of the engineering structures is highly correlated to soil conditions.
