Research Interests

Reliability-based approaches are usually adopted to explicitly consider uncertainties associated with rock properties while analyzing the stability and design of rock structures. Different methods such as First/Second-Order Reliability Methods (FORM/SORM), Point Estimate Methods (PEMs), Monte-Carlo Simulations (MCSs) etc. are being increasingly employed recently. These traditional methods require site-specific statistical parameters of rock properties such as mean and standard deviation (SD), probability density functions (PDFs), correlation function and correlation length (scale of fluctuation) for the analysis. Estimation of these parameters requires extensive laboratory and in-situ measurements. This is generally not possible in rock engineering practice due to difficulties like sample disturbances, huge costs, difficulty in site preparation etc. Under such complex situations, the accuracy of the estimated parameters becomes highly questionable, resulting in inaccuracies in the reliability estimates of rock slopes and tunnels. This thesis aims to counter this issue by developing advanced reliability approaches based on resampling methods (i.e., bootstrap and jackknife), Bayesian inference, multi-model inference, copula theory, Bayesian compressive sampling etc.