Author(s): Anna Rita Scorzini, Anna Rita Scorzini

Linked Author(s):

Keywords: Discharge coefficient; De Marchi; Side weirs; Inclined; Neural networks;

Abstract: The use of De Marchi’s approach, solving the 1D dynamic equation of spatially varied steady flow with non-uniform discharge, is commonly accepted in the design of side weirs. A key issue in applying De Marchi’s formula is the assessment of the discharge coefficient CM. However, no explicit equation for the experimental determination of CM can be derived for inclined side weirs, due to the longitudinal change of crest’s height. In this context, the present study analyzes the feasibility of using alternative methods for the estimation of the discharge coefficient (i.e. Dominguez’s, Schimdt’s or other approaches), which may be suitable to be used in De Marchi’s equation also for inclined side weirs. However, this solution necessarily yields an additional error in the estimation of CM, due to the different modelling assumptions underlying these other methods. Therefore, in this study, the magnitude of this error is first quantified using a 1D numerical model for different tested hydraulic conditions and geometric configurations of the side weir, including: Froude number (only subcritical flows), channel and friction slope, crest angle, water depth/weir height and weir length/channel width ratios. Results indicate that error factors (i.e. observed/predicted ratio) in the assessment of De Marchi’s coefficient can range from 0.57 to 15.60 for inclined lateral weirs, depending on the selected modelling approach. As a second step, a Multilayer Perceptron Neural network is applied to derive transfer functions from the discharge coefficients calculated using the different methods to the corresponding values of CM to be used in De Marchi’s weir equation.

DOI: https://doi.org/10.3850/38WC092019-1029

Year: 2019