Analysis and Modeling of Combined Discharge through Box Culverts and over Contracted Broad-crested Weirs

An experimental investigation is conducted on simultaneous flow through culvert and over broad-crested weir. The flow at the culvert outlet is considered as submerged flow. Both the culvert and the broad-crested weir (BCW) widths are smaller than the channel width. The characteristics of the simultaneous flow through both structures and the effects of flow and geometrical parameters of the structure were presented in the first part of this paper. In this part, discharge prediction models are developed using (a) multiple linear regression, (b) basic discharge equations of BCW and culvert and (c) artificial neural networks. The prediction and performance of each of the three models are compared using four criteria, namely, (i) mean relative absolute error, MRAE, (ii) root mean square error RMSE, (iii) coefficient of correlation, R, and standard error of estimate, SEE. Analysis of results indicated that the ANNs are powerful prediction tool compared to other models followed by the model based on the basic discharge equations of BCW and culvert. Sensitivity analysis is also conducted to reflect the relative importance of the involved variables.