Research on Prediction of Soil Suction in Expansive Soil

Soil-water characteristic curves of expansive clay are usually measured in the laboratory, but soil suction in the field is extremely difficult and time consuming. In this paper, the method of artificial neural network (ANN) is adopted to predict soil suction in the field by using measured water contents. This is done by training the network using laboratory measured soil-water characteristics. Prediction soil suctions using the ANN with some limited in-situ measured water contents are compared with actual suction measurements in the field. Prediction results are discussed.