Design Optimization of I-Wall Levee System Supported by Sand Foundation

Failures of I-wall levee systems were found to be the major cause of the extensive flooding after the Hurricane Katrina and the events thereafter. These events suggest that the current design procedures must be revised to reduce the overall probability of failure of the system by incorporating the uncertainties in the input parameters into the system. In this paper, a framework to perform design optimization of a typical I-wall levee made of clay resting on a sand foundation is proposed. The uncertainties in the system such as undrained shear strength of levee fill, friction angle of sand foundation, and flood water level behind I-wall were considered as random variables in the design optimization procedure. The design variables used were the penetration depth of I-wall from the levee crown, width of levee crown, and landside slope of levee. Several design cases were simulated and finite element analysis was performed for selected cases to estimate the overall stability of the system considering the possible variations in the random variables. A response surface was developed based on the FEM simulation results to represent the factor of safety as a function of random and design variables and implemented in optimization procedure. In the design optimization, total cost of I-wall levee system was minimized to reach cost-efficient designs and at the same time the probability of failure was evaluated and minimized. Finally, the optimization yielded a set of preferred designs known as Pareto front from which the optimal design can be selected.