Recent advances in the stability assessment of natural and engineered rock slopes

Abstract

 Science’s understanding of the failure mechanisms of large natural and engineered slopes has been improved considerably over the past 15 years. Significant improvements have been realized in innovative methods of data acquisition through field measurement and monitoring, as well as numerical modelling techniques. However, inadequate understanding of complex geology and landslide processes means that any interpretation of landslide data remains mostly subjective. This causes major uncertainty in landslide risk assessment.

Over the past decade, Australia’s Commonwealth Scientific and Industrial Research Organisation (CSIRO; http://www.csiro.au/) has developed novel techniques to facilitate efficient assessment of rock slope stability. These include SirovisionTM, Siromodel, and three CSIRO numerical codes: CSIRO‑SPH, CSIRO‑DEM and CSIRO‑COSFLOW.

SirovisionTM is a geological/geotechnical mapping and analysis system that generates accurate, scaled 3D images of rock faces from stereo photographs of exposed rock surfaces, allowing for rapid rock mass structural mapping. Siromodel is a polyhedral modelling system that reads the SirovisionTM data, generates discrete fracture networks (DFN) and performs polyhedral (rock block) modelling and a first‑pass stability analysis.

CSIRO‑SPH, CSIRO‑DEM and CSIRO‑COSFLOW are all used for detailed stress‑deformation analysis of rock slopes; however, each code has its own problem‑specific advantage. CSIRO‑SPH is suited for large deformation problems, and can simulate large scale fluid flow problems, such as modelling a dam breakage. CSIRO‑DEM is suited for rock breakage process analysis, and assessment of the runout distance of failure debris. CSIRO‑COSFLOW is designed specifically for efficient, accurate stress‑deformation analysis of stability of structures on bedded sedimentary rocks, where failures along the preexisting bedding planes and through the intact rock layers occur simultaneously.