Since the 2000s, most of the developed nations, such as the United States, European countries, and Japan, have been using reliability-based limit state design. Although the load and resistance factors of concrete tracks, railroad bridges, and track beds are being calibrated with the Korea Railroad Research Institute (KRRI) leading the project, retaining walls have been excluded from the research. While it is fairly easy to estimate the load effects such as the Earth’s pressure and surcharge load on a retaining wall system, once this system is destroyed, it is extremely difficult to actually measure its resistance. To address this issue, limit state design (LSD) methods have been applied overseas for calibrate the resistance factor to realize values appropriate for the safety factors of the existing allowable stress design (ASD) methods. This study aims to determine the applicability of the LSD methods in the field of railroad design by performing reliability analysis of the standard drawings of railroad retaining walls in Korea, load and resistance factor design (LRFD), and evaluating whether these values meet the target reliability index and the LSD conditions.
For this purpose, the differences between the current retaining wall design methods in ASD and the design regulations of LRFD were researched. Through this comparison, it was confirmed that there is a difference in the selection methods of resistance force of friction between the base of the retaining wall and the foundation ground and the method for calculating the coefficient of bearing capacity. Thus, it was confirmed that different designs were obtained by the reliability based design (RBD) and LSD methods.
In reliability analyses targeted towards non-specific areas, the results of major studies on the uncertainties of geotechnical parameters are used to define the coefficient of variation (COV) of the random variable. Because the COV has a large range, the interpretation of the results depends on the researcher’s decision. To ensure the reliability of the results of this study, both Korean and foreign research papers on the reliability analysis of retaining walls were investigated, but only the statistically significant characteristics pointed out by the researchers were incorporated; the median values of such data were chosen, or the results from a more recent study were adopted. In addition, the COV of the load was applied with consideration to the low uncertainty of train loads, in contrast to the vehicle load characteristics in road transportation. This was done because most of a train’s load is attributed to the vehicle itself, and all other stored loads are pre-planned by the train managers.
The target reliability index ( = 2.33) of a ground structure whose key properties follow the minimum safety factor standards of the ASD method was selected as the evaluation criterion for the reliability analysis, and it was compared to the reliability index of the retaining wall’s failure modes. A standard drawing that was unable to satisfy the target reliability factor, both in terms of sliding and bearing capacity failure mode, was identified, and the design variable (random variable) that significantly affected the failure mode of the retaining wall’s standard drawing was considered the angle of internal friction of the foundation and backfill material. It was determined that even when the uncertainty in the angle of internal friction of the foundation, which has the highest sensitivity, was relatively reduced, some standard drawings did not meet the target reliability index. This was attributed to the statistical potential for significant decreases in the resistance force of friction and the bearing capacity factor between the base of retaining wall and the ground, when considering the uncertainty in the angle of internal friction.
The load and resistance factors were determined by applying various combinations of loads, as per the limit state, to the retaining wall. Then, by applying the utilization factor derived from the ratio between factored load and factored resistance, it was determined whether the conditions for LSD were met. The standard drawings for the retaining wall did not satisfy the failure mode of the bearing capacity. Further, it was determined that the disparity between the calculation method of the friction and coefficients of bearing capacity required by the ASD, and the LRFD was a problem, and that future implementation of the LSD should be preceded by detailed review of the design methods.
In addition, to apply the LSD methods to the railroad retaining wall, the uncertainty in the geotechnical parameters representative of domestic ground characteristics should be quantified. In addition, an evaluative study of the reliability levels, and the target reliability index, incorporating economic effects, should precede any such implementation.