Prediction of unsaturated hydraulic conductivity using fuzzy logic and artificial neural network

Abstract

Knowledge of hydraulic properties is necessary for hydrological studies, artificial recharge of the aquifer, watershed management and agriculture system. The major objective of this study was to develop a fuzzy logic and artificial neural network (ANN) based models for estimating the unsaturated hydraulic conductivity of soil (Ku). A mini disk infiltrometer, being handy used for determining infiltration characteristics. In this study mini Disk Infiltrometer (Decagon Devices, Inc.) at a suction head varying from 1 to 6 cm was used for determining Ku of sandy soil. All the measurements have been done on predetermined initial condition of different proportions of rice husk ash and fly ash mixed with sand. For modeling randomely selected 70% data was applied for training and residual 30% for the test. Comparison of results show that the prediction with ANN approach works well with correlation of coefficient value of 0.8662 (root mean square error 4.5607 cm/h).