Prediction of lateral outflow over triangular labyrinth side weirs under subcritical conditions using soft computing approaches

This paper presents the results of laboratory model testing of triangular labyrinth side weirs located on the straight open channel flume. The discharge capacity of triangular labyrinth side weirs is estimated by using two different artificial neural network (ANN) techniques, that is, the radial basis neural network (RBNN) and generalized regression neural network (GRNN), and gene-expression programming (GEP), which is an extension to genetic programming. 2500 laboratory test results are used for determining discharge coefficient of triangular labyrinth side weirs. The performance of the ANN and GEP models is compared with multi-linear and nonlinear regression models. Comparison results indicated that the neural computing and gene-expression programming techniques could be employed successfully in modeling discharge coefficient from the available experimental data.