Application of Genetic Programming to Flow Routing in Simple and Compound Channels

Abstract

AbstractHydraulic methods can model channel flow with high accuracy using data related to channel geometry and flow regime that render the computational effort burdensome. In contrast, hydrologic methods apply simplifying assumptions in their algorithms for flow routing. This paper implements genetic programming (GP) to calculate hydrographs in simple and compound channels. Predicted hydrographs for the simple and compound channels are compared with those predicted by a Muskingum model and a one-dimensional (1D) coupled characteristic-dissipative-Galerkin (CCDG-1D) procedure. Results show that the differences between predicted hydrographs by GP and modeled hydrographs by the Muskingum and CCDG-1D methods are similar in simple and compound channels. Moreover, GP yields acceptable predicted hydrographs with decreased computational burden. These results indicate that the proposed GP method is effective in the prediction of open-channel flow.