Multiscale analysis of non-contact splices at drilled shaft to bridge column interface

•Specimens of non-contact splice at bridge column-drilled shaft interface are tested.•Experimental observations are validated by 3D and 2D numerical simulations.•Effect of lap spacing on the structural performance is discussed.•The mesoscopic concrete in the overlapped region is constructed with the RAM.•The shearing and splitting cracks are predicted by multiscale simulation. The contact and non-contact splices of the longitudinal steel bar and dowel bar have been widely adopted as the interface connection between pier columns and drilled shafts in bridge engineering. With the popularization and application of bridge substructures composed of the rectangular column and circular drilled shaft, the suitability and accuracy of the design provision specified in the current codes, which is derived from the investigation of the uniform circular components, are becoming the research hotspots in engineering and academia. In order to further study the mechanical behavior of the pier column-drilled shafts interface with inconsistent geometrical shapes, the experimental and numerical analysis of the interfaces with the overlap spacing of 0, 4, 6 and 8 in. are carried out systematically. The experimental results indicate that with the increment of lap spacing, the load-bearing capacity of the components will be consistently reduced, and the material strength also plays an important role in the structural performance. Compared with the specimen with the contact splice, the non-contact lap splice between the longitudinal bar and dowel bar will lead to additional shearing cracks and splitting cracks, and the inclination angle of the shearing cracks gradually increases with the enlargement of lap spacing as well. The relationship of the lateral displacement at the top of the column and the vertical reaction agrees well with that of the numerical analysis. Moreover, the fracture distribution and the interface separation can be explicitly depicted by 3D/2D numerical models based on homogeneous concrete assumption, which fully validates the rationality of the experimental observations. In order to exclude the influence caused by the variation of the concrete strength in tested specimens, the material properties in 3D and 2D simulation are uniformly specified according to the measured material strength of Specimen 1 and the numerical findings show that the structural stiffness of Specimens 1–4 decrease with the increment of lap spacing. However, the ultimate bearing capacity