Seismic Behavior Of Steel H-Piles In Bridges

In current practice, steel H-piles, with relatively small cross sectional area causing little soil displacement and minimal heave of surrounding ground have been widely used in bridges because of their ease of handling, strong penetration and extension or reduction in length. In particular, in the United States, thousands of bridges stand on integral abutments supported by H piles.For bridge structures, in contrast with huge axial force, lateral load seems to play minor effect in the overall response of steel H-pile groups during the service limit. However, for extreme conditions such as ship impact and severe earthquake, piles can undergo very large displacement. In this situation, lateral load will become more and more important. It is well known that the laterally loaded single pile response can form the basis of researches of lateral pile group response, and the dynamic performances of single pile is always supported by its static behavior, thus, it is deemed necessary to explore the static behavior of laterally loaded single steel H-piles.Firstly, this paper reviews and classifies the tests of laterally loaded piles in the last more than sixty years, which can form the data basis for the future analytical researches. Besides, all the analytical methods are summarized and classified, and advantages and disadvantages in each method are discussed in detail. These methods can provide theoretical basis for the innovative piles subjected to lateral load in the future, and also lay the analytical foundation for the piles under extreme dynamic lateral loading.Then this paper presents the test layout of laterally loaded steel H-piles. Four approximately one-third scale laterally loaded steel H-piles, subjected to monotonic and reversed cyclic loading, respectively, are conducted. Two steel H-piles are loaded in strong axis, and the other two are loaded in weak axis. All the piles are pre-installed, and then sand is placed by vibration compaction.It can be seen from the test results that the simulated H-pile on sand foundation has very stable behavior under large displacement in the strong and weak directions. The capacity of the H-pile- sand system is influenced more significantly by the soil at the initial stage of laterally loading with the pinching shape of hysteresis loop while the response under large displacement is primarily controlled by the steel pile with full behavior of hysteresis loop.The bending moment profiles along the pile depth are similar for both monotonic and cyclic loading, however, the location of the maximum moment is lower in the cyclic tests compared with the monotonic tests. The reason is that cyclic effects result in deeper gap openings, which make the location of the maximum moment move slightly downward. This observation is more obvious for the H-piles loaded in strong direction than that in weak direction. And observed date also shows that the location of the maximum moment depends on the loading direction related to the sectional axes of the H-piles. The location of the maximum moment in the H-piles bending in strong axis is deeper four times the sectional width than the corresponding depth in the weak direction. The test results indicate that the location of the maximum earth pressure agrees well with that of the maximum moment. The earth pressure versus loading displacement curve appears to be parabolic, which represent the high nonlinearity of the soil. In addition, it can be seen from the degradation curve of the stiffness that the lateral displacement of the maximum stiffness for the H-pile bending in strong axis is larger than the corresponding value in the weak axis. The reason may be that the cyclic effect for H-piles loaded in strong axis can make the sand denser than that in weak axis.Based on the test data, finite element model is also built to explore the seismic behavior of steel H-piles in sand. In general, p-y curves generated with the API model and the Reese model reasonably predict the overall force-deformation response of both piles though it slightly underestimate the lateral stiffness and the ultimate capacity. On the basis of lateral force versus displacement curve and moment profile along the pile, Reese model is insensitive to the value of subgrade reaction modulus coefficient. Given that the scatter exists in the relations used to estimate the sand properties and p-y curves, the p-y curves recommended by the API model is slightly modified. The finite element model using the modified p-y curves predicts the lateral stiffness and ultimate capacity very well. Based on the modified finite element model, a series of parametric studies such as varied sand frictional angles, loading eccentricities, axial compression ratios and fixed or free heads are conducted.Finally, the moment-curve equations under different axial compression ratios are employed in the basic equation of the laterally loaded pile in soil.By solving the final equation, the response of laterally loaded steel H-piles under different axial compression ratio can be obtained. In addition, the relationship between displacement ductility and curvature ductility factors can be achieved for steel H-piles loaded in strong and weak directions, respectively, which could provide some insights for further researches on the seismic behavior of steel H-piles.