Prediction of compression index using artificial neural network

Abstract

Over the decades, a number of empirical correlations have been proposed to relate the Compression Index of normally consolidated soils to other soil parameters, such as the natural water content, liquid limit, plasticity index and void ratio. In this article too it has been attempted to establish a correlation between compression index and physical properties for the clayey soils of Mazandaran region. Due to the multiple effects of various parameters, Artificial Neural Network (ANN) has been adapted for predicting the compression index from more simply determined index properties. In order to develop the ANN model, four hundred consolidation tests for soils sampled at 125 construction sites in the province of Mazandaran, in the north of Iran were collected and 90% of these were used to train the prediction model and the other 10% were used to test it. A comparison was carried out between the experimentally measured compression indexes with the predictions. Furthermore, the predictions of a number of previously proposed empirical correlations were obtained using the available data and it has been shown that an improvement of 1 - 4% with respect to the other correlations has been achieved. Key words: Compression index, consolidation, settlement, artificial neural network, regression analysis.