Faculty of Civil Engineering, University of Tabriz, Tabriz, Iran
Hydraulic jump is a phenomenon which is used to dissipate the kinetic energy of the flow and prevent scour below overflow spillways, chutes and sluices. This paper applies adaptive neuro-fuzzy inference system (ANFIS) as a Meta model approach to estimate hydraulic jump characteristics in channels with different bed conditions (i.e. channels with different shapes and appurtenances). In hydraulic jump characteristics modeling, different input combinations were developed and tested using 1700 experimental data. The obtained results indicated that the applied method has high capability in modeling hydraulic jump characteristics. It was observed that the developed models for expanding channel with a block performed more successful than other channels. For rectangular channels, it was found that the basin with rough bed led to better predictions compared to the basin with a step. In the prediction of jump length, the superior performance was obtained for the model with input combinations of Froude number and the relative height of jump. From the sensitivity analysis, it was induced that, Fr1 (upstream Froude number) is the most significant parameter in modeling process. Also comparison between ANFIS and semi-empirical equations indicated the great performance of the ANFIS.