Model Uncertainties in "Terzaghi and Peck" Methods for Estimating Settlement of Footings on Sand

The Terzaghi and Peck (1948) method was the first for predicting the settlement of footings on sand using standard penetration test blow counts (N values). Over the following 26 years, various modifications to this basic method were suggested. Herein, an extensive database of 426 settlement case histories is used to assess the model factors for this family of methods. The uncertainty of the model factors is characterized using the coefficient of variation. Considering the model factor as a random variable, the common "log-normality" assumption is investigated critically. The impact of model uncertainty on the reliability of settlement estimations is discussed, and criteria are given to assess their accuracy and conservatism. For settlement prediction, the uncertainty from the model factor is most influential. The results indicate a low accuracy for all of the methods considered.