ESTIMATION OF PLASTIC FINE ALTERED RIVER BED PERMEABILITY USING ARTIFICIAL NEURAL NETWORKS

The permeability of the soil is one of the most important properties of an unlined earthen canal or river bed. Using fine plastic particles has experimentally proven to reduce soil permeability, but the experimental study of the effect of a variety of types of plastic fines and their percentages in riverbed soil is tedious work to do. Estimation of permeability of riverbed soil by altering it with plastic fines using Artificial Neural Networks (ANNs) may reduce this effort. Particle size distributions (PSDs) have a significant influence on the permeability of bed soils. Being able to predict the permeability of bed soil by knowing the PSDs may provide an easy approach to know the loss of water by percolation. This study has investigated the quantitative relationships between permeability and PSD indices using ANNs. The aim was to build a mathematical model capable of predicting the permeability of bed soil by PSD indices of choice. A model was built using ANNs including PSD indices as input and permeability as output. The model stated that the coefficients of curvature and uniformity (Cc) and (Cu) and effective particle size (D50) may be used to predict the bed permeability. The computational model was able to predict the effect of variation of PSD indices on bed permeability, thus allowing increasing the efficiency of the river bed, to ensure maximum downstream water supply, lesser seepage and percolation and better productivity. The test result has confirmed the efficiency of the developed ANN tool in predicting the bed permeability for different PSD combinations.