Application of re-sampling stochastic framework for rock slopes support design with limited investigation data: slope case studies along an Indian highway.

Rock slope projects are often available with insufficient investigation data for rock properties due to involvement of high costs, manual efforts, geological complexities, etc., in testing. Support estimation performed by traditional approaches for slope failure mitigation could be highly inaccurate due to inaccuracies in the estimated statistics of properties. This article describes a computationally efficient re-sampling stochastic framework to overcome this issue by coupling the Advanced Re-Sampling Reliability Approach (ARRA) with deterministic and Target Reliability Approach (TRA) to estimate the required support for rock-slides mitigation with limited data. The proposed methodology was demonstrated for the support design to mitigate two massive rockslides along a rockslide-prone highway, i.e., Rishikesh-Badrinath National Highway (NH-58) in India. It was concluded that the limited data invokes uncertainties in the statistical parameters (mean and standard deviation) and distributions of properties. Support estimated using traditional approaches with the inaccurate statistics of input properties can result in inaccurate support estimates in the presence of insufficient data. Proposed methodology couples ARRA with traditional methods and evaluates the final support design by quantifying the uncertainty in the reliability index induced by the statistical uncertainties of input properties statistics. This improves the overall accuracy, efficiency, and the designer's confidence in the estimated support. [ABSTRACT FROM AUTHOR]