Uncertainty Characterization in the Design of Flow Diversion Structure Profiles Using Genetic Algorithm and Fuzzy Logic

Abstract

Flow diversion head works are constructed across rivers to divert flow into irrigation, navigation, or power generation channels. Hydraulic structures, such as weirs or barrages, are integral parts of these diversion head works. The optimal design of these hydraulic structures is generally obtained by considering the deterministic values of hydrogeological parameters. However, there is a high degree of local soil variability and imprecision in the determination of soil parameters, such as the safe exit gradient. The seepage head also exhibits a high degree of variability, depending on complex hydrological and metrological factors. This work considers the hydrogeological parameters safe exit gradient and seepage head as imprecise or uncertain. An optimization-based methodology is presented to incorporate uncertainty in the safe exit gradient and seepage head in the optimization formulation and obtain the optimum structural dimensions that minimize the total cost. The subsurface flow consideration is embedded in the optimization formulation. The nonlinear optimization formulation (NLOF) solution procedure using a genetic algorithm (GA) is implemented to demonstrate the characterization of uncertainty in design and hence, overall cost from the uncertain safe exit gradient and seepage head. The limited evaluation shows the potential applicability of the proposed methodology.