Abstract
The strength reduction method is often used to predict the stability of soil slopes with complex soil properties and failure mechanisms. However, it requires a considerable computational effort. In this paper, we make use of a convolutional neural network to reduce the computational cost. The factor of safety of 600 slopes with different inclination and soil properties is first calculated with the strength reduction method. A convolutional neural network is then trained and validated. We demonstrate the performance of our approach and show how to augment the dataset to further enhance its capability and prevent overfitting.
Funding Information
  • Otto Pregl Foundation for Geotechnical Fundamental Research
  • University of Natural Resources and Life Sciences Vienna

This publication has 51 references indexed in Scilit: