To prevent the collapse of dangerous rock masses on steep rocky reservoir banks which can cause casualties and property loss, it is essential to design and conduct practical experiments to quantify the evolution processes of the reservoir banks and control such dangerous rock masses. Using the Jianchuandong Dangerous Rock Mass project as a case study, this book generalizes the mechanical model of the project to show how improved equipment can be used to measure the mechanical state transition under the continuous action of axial pressure. It details a series of experiments to study the evolution of a severely steep rocky reservoir bank, which comprehensively consider the influence of hydraulic coupling, dry-wet cycles, axial pressure, and time-dependent effects. The results support a new method for determining the stability of dangerous rock masses on reservoir banks. Combines engineering principles, real data, experimental methods and results Provides a complete research method for investigating hydrogeology failure processes The book suits practitioners in hydropower engineering, engineering geology, and disaster protection.

Bayesian data analysis and modelling linked with machine learning offers a new tool for handling geotechnical data. This book presents recent advancements made by the author in the area of probabilistic geotechnical site characterization. Two types of correlation play central roles in geotechnical site characterization: cross-correlation among soil properties and spatial-correlation in the underground space. The book starts with the introduction of Bayesian notion of probability “degree of belief”, showing that well-known probability axioms can be obtained by Boolean logic and the definition of plausibility function without the use of the notion “relative frequency”. It then reviews probability theories and useful probability models for cross-correlation and spatial correlation. Methods for Bayesian parameter estimation and prediction are also presented, and the use of these methods demonstrated with geotechnical site characterization examples. Bayesian Machine Learning in Geotechnical Site Characterization suits consulting engineers and graduate students in the area.

Uncertainty, Modeling, and Decision Making in Geotechnics shows how uncertainty quantification and numerical modeling can complement each other to enhance decision-making in geotechnical practice, filling a critical gap in guiding practitioners to address uncertainties directly. The book helps practitioners acquire a working knowledge of geotechnical risk and reliability methods and guides them to use these methods wisely in conjunction with data and numerical modeling. In particular, it provides guidance on the selection of realistic statistics and a cost-effective, accessible method to address different design objectives, and for different problem settings, and illustrates the value of this to decision-making using realistic examples. Bringing together statistical characterization, reliability analysis, reliability-based design, probabilistic inverse analysis, and physical insights drawn from case studies, this reference guide from an international team of experts offers an excellent resource for state-of-the-practice uncertainty-informed geotechnical design for specialist practitioners and the research community.