Computational and Visual Tools for Teaching Probabilistic Geotechnical Design

Abstract

College students in geotechnical design classes can use various computational tools to learn and apply probabilistic design methods, including spreadsheets, Mathcad ® , and structured programming. Also, learning is significantly enhanced by using visual graphical displays of probability density functions (pdf) and how their shapes are influenced by user assumptions. Three stochastic methods for generating an output safety factor pdf are illustrated in an instructional setting: Monte Carlo simulation, the point estimation method, and Fourier convolution of pdf's. Geotechnical design examples are presented for a geogrid-reinforced retaining wall and for several slope stability evaluations. Students can observe that small changes in the deterministic (mean) safety factor as a result of groundwater drainage may be difficult to interpret (because the overall percentage reduction in the safety factor may be quite small), while the probability of sliding for the same set of conditions may be reduced considerably (say, 80 to 90 percent). By using real-world examples, students can use hands-on computing with useful visual components to become familiar with probabilistic applications and to experience new perspectives on the engineering design process. ABSTRACT: College students in geotechnical design classes can use various computational tools to learn and apply probabilistic design methods, including spreadsheets, Mathcad ® , and structured programming. Also, learning is significantly enhanced by using visual graphical displays of probability density functions (pdf) and how their shapes are influenced by user assumptions. Three stochastic methods for generating an output safety factor pdf are illustrated in an instructional setting: Monte Carlo simulation, the point estimation method, and Fourier convolution of pdf's. Geotechnical design examples are presented for a geogrid-reinforced retaining wall and for several slope stability evaluations. Students can observe that small changes in the deterministic (mean) safety factor as a result of groundwater drainage may be difficult to interpret (because the overall percentage reduction in the safety factor may be quite small), while the probability of sliding for the same set of conditions may be reduced considerably (say, 80 to 90 percent). By using real-world examples, students can use hands-on computing with useful visual components to become familiar with probabilistic applications and to experience new perspectives on the engineering design process.