Statistical analysis of the effective stress method and modifications for prediction of ultimate bond strength of soil nails |
| |
| Authors: | Peiyuan Lin Richard J Bathurst Sina Javankhoshdel Jinyuan Liu |
| |
| Institution: | 1.Department of Civil Engineering,Ryerson University,Toronto,Canada;2.Department of Civil Engineering, GeoEngineering Centre at Queen’s-RMC,Royal Military College of Canada,Kingston,Canada;3.Department of Civil Engineering, GeoEngineering Centre at Queen’s-RMC,Queen’s University,Kingston,Canada |
| |
| Abstract: | Uncertainty in the predicted ultimate pullout strength of soil nails can be significant due to the complexity of nail–soil interactions, inherent variability in soil properties and the effects of nail installation. The paper first presents a statistical evaluation of the accuracy of ultimate bond strength of soil nails using the effective stress method (ESM) equation that has been adopted in Hong Kong. A total of 113 ultimate nail capacity data points from field pullout tests were collected from the literature and used to estimate the accuracy of the current ESM. Based on the available data, the current ESM default pullout model is found to be excessively conservative (on average) by at least a factor of three. The spread in prediction accuracy measured by the coefficient of variation (COV) of bias is in the range of 36–43 % after removing anomalous test data. Here, bias is the ratio of measured to predicted pullout load capacity. In addition, the accuracy of the current ESM equation for prediction of nail bond strength is shown to be dependent on the magnitude of predicted ultimate bond strength and magnitude of nominal vertical effective stress which is undesirable. The paper examines four candidate-modified bond strength equations with empirical coefficients that have been back-fitted to measured bond strengths to improve the overall accuracy of the equation and to reduce or remove the undesirable dependencies noted above. One equation with an empirically corrected stress-dependent term is judged to be the best candidate model based on the mean of bias values, spread (COV) of bias values, lack of dependencies and simplicity. Finally, the relative contributions of random variability in soil shear strength to measurement bias in bond strength (prediction accuracy) for each soil type are computed for the best bond strength model. Analysis of the contribution of soil shear strength to prediction accuracy showed that the combination of variability due to factors other than soil shear strength was greater than the variability in soil shear strength alone, where the latter is defined by the soil secant friction coefficient. |
| |
| Keywords: | |
| 本文献已被 SpringerLink 等数据库收录！ |
|
