Comparison of limit states design with working stress design for shallow foundations

Bridge foundations have traditionally been designed using working stress methods, but the new Canadian Highway Bridge Design Code (draft CHBDC) now specifies a limit states design procedure for these structures. The main objective of this study was to compare working stress design (WSD) with limit states design (LSD) methods particular to bridge abutments. The two design methods have been investigated and compared to a numerical model (developed using the program FLAC). The results of these analyses were compared for reliability and safety. LSD was applied to an existing bridge abutment (the No. 5 Road Bridge in Richmond, British Columbia) which was initially designed using WSD. The two different designs were compared on the basis of factors of safety with the outcome indicating that the structure having been designed using WSD may be too reliable and overly safe. A FLAC model of the No. 5 Road overpass abutment was developed and incrementally loaded to failure in order to determine the capacity distribution of the structure. The resulting normal distribution of capacity was used in a reliability analysis with two different models for loading. This analysis yielded a relationship between mean live load and reliability index for this particular structure. The results indicated that the reliability index at the design live load was higher than the value of 3.5 that was used to calibrate the CHBDC LSD partial factors. The expected displacement during the regional design earthquake was predicted using a FLAC model. The model was run a number of times with various earthquakes and combinations of soil properties. The results of the FLAC runs were combined with joint probabilities of occurrences of soil parameters (derived from a survey) to obtain the expected displacements. The results showed relatively small expected values of displacement which also indicated that the original abutment design may be overly safe in terms of the draft CHBDC. . A sensitivity analysis involving soil parameters was also considered. The soil properties were varied within the FLAC model to determine the resulting variation in displacements, and to ascertain which variables most affect the outcome of the analysis. Friction angle was found to be the critical soil property, as it had more of an effect on displacements than did (Ni)60 or unit weight.Applied Science, Faculty ofCivil Engineering, Department ofGraduat