Date of Award:

5-2010

Document Type:

Thesis

Degree Name:

Master of Science (MS)

Department:

Civil and Environmental Engineering

Advisor/Chair:

Dr. John Rice

Abstract

A new method for assessing the potential for unsatisfactory levee performance due to underseepage is presented. Specifically, the method assesses the potential for the initiation of piping (the internal backward erosion of the foundation or embankment caused by seepage). Current assessment methods consist of deterministic seepage analyses and simplified reliability methods. Deterministic methods produce either a maximum hydraulic exit gradient or a Factor of Safety against piping but they do not account for high levels of uncertainty in soil properties and subsurface geometry that are inherent to many levee analyses. The most common simplified reliability approaches that are currently being used to analyze levees with regard to underseepage apply the First Order Second Moment (FOSM) Taylor Series method using the US Army Corps of Engineers "Blanket Theory" equation as the performance function. These methods take into account the uncertainty of the soil properties but are limited to simplified subsurface geometries and often do not model the actual mechanism responsible for levee failure due to underseepage piping. The proposed new method uses a Monte Carlo simulation to calculate the probability of unsatisfactory levee profile performance and can take into account complexities in subsurface geometry that cannot be assessed using the simplified reliability methods. The relationships between uncertainty of the soil parameters, the subsurface geometry, and the Factor of Safety against piping are defined through parametric variation analyses of a finite-element seepage model. The results of the parametric analyses are used to develop a series of equations that define the relationship between the various input parameters and the factor of safety. Using these equations, probability density functions for the various input parameters, and the computer program @Risk which interfaces with Excel, a Monte Carlo analysis is performed to calculate the probability of unsatisfactory performance which represents the probability of initiating erosion given a river flood level. The results of the analysis represent a single node of the event tree. In order to assess failure potential, other points in the event tree will need to be assessed with calculations or judgment since it is only the first phase that is currently considered to be in the process of piping. The new method is demonstrated using actual data of levee profiles from the Natomas Basin in Sacramento, California as a case study. The case study highlights the benefits of reliability-based analyses over the Factor of Safety and demonstrates the importance of subsurface geometry in reliability calculations.

Share

COinS