Models for kinetic energy and momentum correction coefficients for non-prismatic compound channels using regression and gene expression programming

This paper studies the effect of different independent variables related to the characteristics of the compound channels on the kinetic energy and momentum correction coefficients (termed as α and β , respectively) of non-prismatic compound channels. A series of experiments were carried out in a converging compound channel with three converging angles, six width ratios, six relative depths and three relative distances. Meanwhile, flow data sets from three different authors were used for diverging compound channels, which include simple rectangular and trapezoidal cross sections. In the present study, conventional multivariable regression technique and gene expression programming were used for developing different models for calculating kinetic energy and momentum correction coefficients for converging and diverging compound channels, respectively. The independent variables used in the curve fitting process were also tested for their accountability and influence on the proposed models. The developed models were finally established with the validation data set and a natural river system.