Analysis of Artificial Neural Network in Predicting the Maximum Amount of Stress on Clay Slopes Stabilized With Cement, Lime and Fibers by Laboratory Methods and Numerical Modeling

In this research, soil stabilization has been done by an optimal combination of fibers and traditional stabilizers, as well as using the results obtained in the analysis of the stability of the roofs and finally simulating the results in the environment of genetic programming. A numerical model was obtained by simulating slopes in the OptumG2 software environment. Numerical simulation modeling environment was used by two famous artificial neural networks, the feed forward and genetic programming neural networks. For slopes with an angle of 75 degrees, the maximum vertical stress applied on the foundations is equal to 8, 285, 499, 808 and 1516 kilopascals for soils with a shear resistance of 25 kilopascals (or unstabilized soil), 100, 200, 300 and it was 400 kilopascals. It is worth mentioning that the relationship between the undrained bond strength and the slope reliability factor (for a constant condition of the slope geometry and the location of the foundation relative to the crest of the slope) is a linear relationship. Metal fibers showed the greatest effect on the strength of the fixed samples. Cement stabilizers showed more results than lime stabilizers. The increase in undrained soil resistance can greatly affect the stability of these slopes and also the maximum stress that can be applied to the crest of these slopes. In predicting the maximum amount of stress on the slopes, the artificial neural network analysis was a good representative of this issue.