A specialist program to evaluate slope safety factor
Joana Figueira Nobrega1; Fernanda Lins Gonçalves Pereira1; Deane Roehl1
1 Tecgraf Institute - PUC-Rio
doi:10.20906/CPS/CILAMCE2017-0192
Resumo
Slope stability is one of the processes evaluated during the construction of dams, embankments and tailings. Aiming the optimization of this evaluation, this study applies theories and methodologies that provide the slope stability safety factor (SF) efficiently. The study is concentrated on the development of a specialist program that evaluates the slope SF with minimum resources and time. An Artificial Neural Network (ANN) is trained using the results of a number of numerical simulations carried out with a nonlinear finite element (FE) program. In order to obtain the SF through this numerical approach, the Strength Reduction Method is used together with the FE program Abaqus and a subroutine, written in Python language, which permits the parametrization of the problem variables. Finally, the MATLAB program is used to construct and train the ANN using the soil strength parameters as input data. A single layer slope stability problem is applied to illustrate the methodology presented. A set of tests varying the number of samples and the architecture of the network is investigated. According to the results, good performance is reached.
Palavras-chave: Slope stability; Finite Element Method; Artificial Neural Networks; Safety Factor