The amount of settlement in rockfill dam is one of the important indices which indicate its performance
and stability condition. In this study a prediction method to analyze the settlement behavior
of concrete faced rockfill dams (CFRDs) after construction is proposed based on the instrumentation
data measured in the fields. The proposed prediction method adapts the artificial
neural network (ANN) technology to estimate the crest settlement from the internal settlements
and uses four input variables such as height of dam, shape coefficient, void ratio and internal
settlement. The comparison of settlements between the prediction method by ANN and the
conventional one was made in terms of correlation coefficients, and indicates that the proposed
method can give a good estimation of settlements in CFRDs.
