Artificial Neural Network Modelling for Estimation of Suction Capacity

The objective of this study was to estimate the suction capacity in clays by using neural networks on the basis of some simple soil properties. For this purpose, a total of 168 data sets obtained from water suction tests were used in the Artificial Neural Networks (ANNs). These tests were carried out on seven different clay samples by using odometer test equipment. One hundred twenty six data sets were used in the training stage and remaining 42 data sets were used in the testing stage. The best model was determined by using a trial and error approach in which the ANN models were trained with different hidden layer nodes and with different combinations of network parameters. The neural network predictions were compared with the water suction test results. It is seen that, there is a good agreement between the prediction results and the results of experiments. The results of this study demonstrate that the neural network model can serve as an alternative predictive tool for determination of water suction capacity of clay soils.