Sediment removal efficiency computation in vortex settling chamber using artificial neural networks

The flow mechanism in the vortex settling chamber is so complicated that it is difficult to establish a general regression model to provide accurate estimation for sediment removal efficiency. Hence in this study an alternative approach of artificial neural network is proposed to determine the sediment removal efficiency of the vortex settling chamber. Experimental data collected in present investigation and the laboratory and field data collected from literature having wide range of hydraulic and geometrical variables are used to train, test and validate the network. A network architecture complete with trained values of connection weight and bias and requiring input of ungrouped parameters pertaining to Qi, Qu, Zh, hp, DT, B, du, d50, and ωo is recommended in order to predict the removal efficiency of vortex settling chamber. On the basis of the sensitivity analysis, it is observed that QR is the most significant parameter after Re. Neural network predictions have been compared with various existing regression models. Predictions based on original raw data (Qi, Qu, Zh, hp, DT, B, du, d50, ωo) were better than those based on grouped dimensionless forms of the data (QR, ZR, BR, DR, Re).