Nonlinear Mechanical Analysis of Rock Slopes

  geological outcrop in open cast mining Copyright: © LIH




Nonlinear mechanical analysis of rock slopes

Funding: DAAD

Project term: 2011 to 2012


With the economical and social development, more and more huge-scale civil engineering projects are developed. Many geotechnical problems are becoming more and more complicated and influenced by many factors. For example, the shiplock slope of the Three Gorges Project, the main powerhouse chamber of Xiluodu hydropower station and the deep anchor ingot foundation pit of Yunyang Bridge are examples for technically and scientifically challenging large projects. Technologies to solve the complicated geotechnical problems arising from such projects are very important for secure and economic construction projects world wide. At present, there are two main attempts for solving geotechnical problems, namely conventional methods and newly developed thoughts like artificial intelligence methods or medical imaging-based methods. The conventional methods are well established and are very frequently and extensively employed by geotechnical scientists and engineers. However, there often exist considerable differences between the measured/observed responses of geotechnical systems and those computed by means of the conventional methods. This is mainly due to the variable and heterogeneous nature of the geotechnical materials (rocks and soils) and the simplifying assumptions and idealisations that are adopted in the conventional models and theories to describe the behaviour of such materials. The shortcomings of the conventional methods require new approaches. It is an inexorable trend to integrate conventional scientific theory with artificial intelligence science to study the geotechnical problems. However, up to now, no one has employed this integration to analyse the stability of rock slopes successfully. Rock slopes can be treated as complicated systems and their constitutive equation could be obtained by adopting intelligent methods, for instance by the coupling of neural networks and genetic algorithm. Then, on the basis of the intelligent constitutive equation, FEM and DEM could be adopted to compute stress field, displacement field and other physical quantities.