Application of BP neural network in the prediction of consolidation coefficient

Abstract

The application of artificial neural network (ANN) in the discipline of geotechnical engineering is discussed in this paper. A multi-layer error back-propagation (BP) feed-forward neural network model was proposed to predict an important geotechnical parameter, namely the consolidation coefficient. The conventional methods for predicting consolidation coefficient is briefly introduced. Based on the results of laboratory consolidation tests, the BP model was trained and used to determine the consolidation coefficient. The predicted values were compared to those determined by graphical methods. It is proved that the BP neural network approach yielded similar results compared with other methods.