Statistical Evaluation of the FHWA Simplified Method and Modifications for Predicting Soil Nail Loads.

A total of 123 measured maximum nail-load data were collected from instrumented soil nail walls reported in the literature. Filtered data sets corresponding to short-term and long-term measurements were used to evaluate the accuracy of the current Federal Highway Administration (FHWA) simplified method to calculate maximum nail loads under operational conditions. The accuracy of load predictions was quantified by the mean and coefficient of variation (COV) of the ratio (bias) of measured load to predicted load. Data in short-term and long-term categories were also investigated according to frictional-cohesive and frictional soil types. Based on the available data, the current FHWA simplified method was found to overestimate both long-term and short-term maximum nail loads on average. The spreads in prediction accuracy measured by the COVof bias were 38 and 52% for long-term and short-term data, respectively. Large spreads in prediction accuracy were also found using data for walls with cohesive-frictional soils alone. With the exception of the frictional soil data set, there was an undesirable correlation (dependency) between method accuracy and predicted maximum nail load using the current FHWA simplified method. A modified FHWA simplified method equation is proposed that has fewer empirical coefficients than the current formulation (i.e., three compared to five) and is shown to be more accurate on average and to have less spread in prediction accuracy for all data sets. Furthermore, hidden dependencies between method accuracy and magnitude of predicted load are not present. [ABSTRACT FROM AUTHOR]