Comparative Study of Effective Hydraulic Parameters on hydraulic jump characteristics in Channels with Compound Sections using Kernel Based SVM Approach

Hydraulic jump is the most common method of dissipating water’s kinetic energy in downstream of spillways, shoots and valves. So far several relations have been developed to estimate hydraulic jump characteristics, however, the results of these equations are not general and acceptable due to the uncertainty of the function. In this study, hydraulic jump character-istics such as sequence depth ratio and hydraulic jump were estimated in compound channels (rectangular and trapezoidal channels) with rough beds using Support Vector Machine (SVM). Different models were developed and the influence rate of input parameters in each channel was investigated. Comparison of the obtained results of support vector machine showed the high efficiency of this method in estimation of hydraulic jump characteristics. It was observed that model with input parameters of Fr₁( Fraud number), w/z (ratio of rough elements space to height of them) led to most accurate results and rough elements properties were effective in hydraulic jump characteristics estimation. The best result for test series was obtained for the sequence depth ratio with the values of R=0.979, DC=0.975 and RMSE=0.046 and for and the hydraulic jump length with the values of R=0.935, DC=0.858 and RMSE=0.072 in trapezoidal channel. Also, the results of sensitivity analysis indicated that Fraud number is the most significant parameter in estimation of hydraulic jump charac-teristics.