This thesis investigates the capacity and failure mode of suction caissons under inclined loading. Parametric finite element analyses have been carried out to investigate the effects of caisson geometry, loading angle, padeye depth (i.e. load attachment point), soil profile and caisson-soil interface condition. Displacement-controlled analyses were carried out to determine the ultimate limit state of the suction caissons under inclined load and the results presented as interaction diagrams in VH load space. VH failure interaction diagrams are presented for both cases where the caisson-soil interface is fully-bonded and where a crack is allowed to form along the side of the caisson. An elliptical equation is fitted to the normalised VH failure interaction diagram to describe the general trend in the case where the caisson-soil interface is fully-bonded. Parametric study reveals that the failure envelope in the fully-bonded case could be scaled down (contracted failure envelope) to represent the holding capacity when a crack is allowed to form. A stronger effect of crack on the capacity was observed in the lightly overconsolidated soil, compared to the normally consolidated soil. The sensitivity of caisson capacity to the changes in load attachment position or loading angle was quantified based on the load-controlled analyses. It was found that, for caisson length to diameter ratios of up to 5, the optimal centreline loading depth (i.e. where the caisson translates with no rotation) is in the range 0.65L to 0.7L in normally consolidated soil, but becomes shallower for the lightly overconsolidated soil profile where the shear strength profile is more uniform. The reduction of holding capacity when the padeye position is shifted from the optimal location was also quantified for normally consolidated and lightly overconsolidated soil profiles at loading angle of 30 [degrees]. Upper bound analyses were carried out to augment the finite element study. Comparison of holding capacity and accompanying failure mechanisms obtained from the finite element and upper bound methods are made. It was found that the upper bound generally overpredicted the inclined load capacity obtained from the finite element analyses especially for the shorter caisson considered in this study. A correction factor is introduced to adjust the upper bound results for the optimal condition. Comparisons of non-optimal capacity were also made and showed that the agreement between the upper bound and finite element analyses are sensitive to the change in the centreline loading depth when the caisson-soil interface is fully bonded, but less so when a crack forms.

Skewed bridges have exhibited poorer performance during lateral earthquake loading in comparison to non-skewed bridges (Apirakvorapinit et al. 2012; Elnashai et al. 2010). Results from numerical modeling by Shamsabadi et al. (2006), small-scale laboratory tests by Rollins and Jessee (2012), and several large-scale tests performed by Rollins et al. at Brigham Young University (Franke 2013; Marsh 2013; Palmer 2013; Smith 2014; Frederickson 2015) led to the proposal of a reduction curve used to determine a passive force skew reduction factor depending on abutment skew angle (Shamsabadi and Rollins 2014). In all previous tests, a uniform longitudinal load has been applied to the simulated bridge abutment. During seismic events, however, it is unlikely that bridge abutments would experience pure longitudinal loading. Rather, an inclined loading situation would be expected, causing rotation of the abutment backwall into the backfill. In this study, a large-scale test was performed where inclined loading was applied to a 30° skewed bridge abutment with sand backfill and compared to a baseline test with uniform loading and a non-skewed abutment. The impact of rotational force on the passive resistance of the backfill and the skew adjust factor was then evaluated. It was determined that inclined loading does not have a significant effect on the passive force skew reduction factor. However, the reduction factor was somewhat higher than predicted by the proposed reduction curve from Shamsabadi and Rollins 2014. This can be explained by a reduction in the effective skew angle caused by the friction between the side walls and the back wall. The inclined loading did not change the amount of movement required to mobilize passive resistance with ultimate passive force developing for displacements equal to 3 to 6% of the wall height. The rotation of the pile cap due to inclined loading produced higher earth pressure on the obtuse side of the skew wedge, as was expected.These findings largely resolve the concern that inclined loading situations during an earthquake may render the proposed passive force skew reduction curve invalid. We suggest that the proposed reduction curve remains accurate during inclined loading and should be implemented in current codes and practices to properly account for skew angle in bridge design.

Skewed bridges have exhibited poorer performance during lateral earthquake loading in comparison to non-skewed bridges (Apirakvorapinit et al. 2012; Elnashai et al. 2010). Results from numerical modeling by Shamsabadi et al. (2006), small-scale laboratory tests by Rollins and Jessee (2012), and several large-scale tests performed by Rollins et al. at Brigham Young University (Franke 2013; Marsh 2013; Palmer 2013; Smith 2014; Frederickson 2015) led to the proposal of a reduction curve used to determine a passive force skew reduction factor depending on abutment skew angle (Shamsabadi and Rollins 2014). In all previous tests, a uniform longitudinal load has been applied to the simulated bridge abutment. During seismic events, however, it is unlikely that bridge abutments would experience pure longitudinal loading. Rather, an inclined loading situation would be expected, causing rotation of the abutment backwall into the backfill. In this study, a large-scale test was performed where inclined loading was applied to a 30 skewed bridge abutment with sand backfill and compared to a baseline test with uniform loading and a non-skewed abutment. The impact of rotational force on the passive resistance of the backfill and the skew adjust factor was then evaluated. It was determined that inclined loading does not have a significant effect on the passive force skew reduction factor. However, the reduction factor was somewhat higher than predicted by the proposed reduction curve from Shamsabadi and Rollins 2014. This can be explained by a reduction in the effective skew angle caused by the friction between the side walls and the back wall. The inclined loading did not change the amount of movement required to mobilize passive resistance with ultimate passive force developing for displacements equal to 3 to 6% of the wall height. The rotation of the pile cap due to inclined loading produced higher earth pressure on the obtuse side of the skew wedge, as was expected.These findings largely resolve the concern that inclined loading situations during an earthquake may render the proposed passive force skew reduction curve invalid. We suggest that the proposed reduction curve remains accurate during inclined loading and should be implemented in current codes and practices to properly account for skew angle in bridge design.