Tools and Strategies for Dealing with Uncertainty in Geotechnics

Abstract

Working with uncertainty is an essential aspect of engineering—the larger the uncertainty and the closer to critical, the greater the need for evaluating its effect(s) on the results. To characterize the uncertainties in soil and/or rock properties, the engineer needs to combine, in addition to actual data, knowledge about the quality of the data, knowledge on the geology and, most importantly, engineering judgment. Once the uncertainty in input parameters and model(s) for solving a particular problem are quantified, the engineer has a variety of tools at his disposal to evaluate the uncertainty in the output. The most common practical tools are Monte Carlo simulation techniques, first-order, second moment (FOSM) approach, first- and second-order reliability methods (FORM and SORM), and event tree analysis. Each has its advantages and shortcomings. The more complicated methods often provide more useful information about the possible outcomes of a problem. These methods are described and their applications are demonstrated through example problems. Many geotechnical problems involve several possible failure modes, which may or may not be correlated. These problems should be treated as systems. Component reliability vs. system reliability are discussed, and example calculations are presented.