Radial Basis Function Regression (RBFR), ARRBFR models for Estimation of Particle Froude Number in Sewer Pipes Under Deposited Conditions

Abstract

The amount of water that can flow through a channel is affected by sediment deposition in water drainage. Because of this, the self-cleaning mechanism is used a lot in sewer systems in cities. The particle Froude number (Fr) is an essential factor in the self-cleaning of sewer systems. This study looks at how well different machine learning (ML) models, both standalone and ensemble, can estimate the Froude number of particles in the condition of non-deposition with the deposited bed (NDB). In this study, wide ranges of the volumetric sediment concentration $(C_{v})$, the particles dimensionless grain size $(D_{gr})$, the median size of sediment (d), the hydraulic radius (R), and the pipe friction factor (k) were taken. Radial Basis Function Regression (RBFR) was used for standalone methods, while Additive Regression (AR) was used for the ensemble method. The correlation coefficient (CC), the Nash-Sutcliffe efficiency (NSE), the mean absolute error (MAE), and the mean square error (MSE) are model evaluation criteria that were used to analyze the proposed models. AR-RBFR is the best model for estimating the Fr(CC=0.954, NSE=0.911, MAE=0.527, and MSE=0.627), followed by RBFR and empirical equations.