Landslide stability analysis with artificial neural network

In this thesis, accountability of slope safety on landslides was investigated using Artificial Neural Networks (ANN) method. Firstly, 100 synthetic landslide data were created for using ANN program and safety factor of these landslides were calculated with GeoStudio software. 80 of these synthetic landslides in training set and 20 of them were used in testing set. The input parameters of ANN method were the height of slope, the height of water level, the inclination of slope, the unit weight of soil, the cohesion of soil and the friction angle of soil, the output parameter were the factor of the safety of the slopes and safety factors were calculated with ANN. The safety factor of calculated by the classical method and the ANN were a very good correlation between each other. Then, seismic velocity was added each landslide according to NEHRP Soil Classification taking into account the parameters of the ground. Also, in this case safety factors were recalculated with ANN. Compressional and shear wave velocities of Giresun Organized Industry landslides were calculated using seismic refraction and seismic surface wave methods. Soil parameters of Araklı-Yiğitözü landslide determined by sampling. The safety factor of three landslides were evaluated using these parameters with GeoStudio software. Unit weight of soil, cohesion and internal friction angle parameters of 3 landslides were given range according to seismic velocities and calculated safety factor with ANN. The results are analyzed, cohesion is detected the most effective parameter on SF. Finally, SF of analyzed landslides evaluated with ANN. As a result, safety factor of landslides can be calculated fast and easy way with ANN, stability calculation for landslides can be execute using seismic velocities instead of undefined some soil parameters.