Longitudinal bar slip resulting from strain penetration at the end of flexural concrete members will result in a member end rotation and additional lateral deformations in members such as walls, columns and beams. The contribution of these member end rotations can account for as much as 35 percent of the total lateral deformations of these members. Therefore, the deformation resulting from strain penetration should be accurately accounted for when modeling reinforced concrete members subjected to flexural actions.

In order to model the bar slip as well as the associated member end rotation due to strain penetration of longitudinal bars adequately anchored into joints or footings, a local bond stress-slip constitutive relationship is typically required to be incorporated in detailed simulation models to model the interface between the reinforcement and concrete. However, the local bond-slip models available for analytical detailed simulation were developed from experimental tests conducted on reinforcing bars with short embedment length, where the slip of test bar occurred when they were subjected to strains well below the yield strain. Consequently, these models are strictly not applicable to critical flexural regions such as the plastic hinges experiencing significant nonlinear strains.

In order for the reinforced concrete structures to develop ductile response during moderate or severe earthquake excitations, these structures are designed to develop plastic hinges in the critical moment regions at the wall and column bases as well as at the beam end regions. In these situations, the longitudinal reinforcement at the connection interface experiences as high as 25 times the yield strain, causing the rebar to slip over the entire (e.g., beam ends) or partial length (e.g., column and wall bases) through deterioration of local bond at the steel-concrete interface. Given that the objective of detailed analysis should be to produce satisfactory global and local responses, it should be realized that accurately representing the local bond-slip behavior of longitudinal reinforcing bar experiencing inelastic strains is critical. This measure will enable the bar slip and member end rotations due to strain penetration to be quantified accurately.

An alternative approach that maybe suitable for fiber-based analysis is to model the bar stress vs. slip hysteretic response directly as proposed by Zhao and Sritharan (2007), thereby capturing the local and global responses accurately. Their model was based on limited test data derived from pull out tests conducted on reinforcing bars with long embedment length, forcing the bars experienced large inelastic strains.

In consideration of the state-of-the-art summary presented above on the bond-slip behavior of reinforcing steel subjected to inelastic strains, an experimental investigation was designed and completed recently. In these tests, test bars were designed with sufficient anchorage lengths as would be the case for bars anchored into foundations. A total of five bars of two different bar sizes (i.e., #6 and #8) were tested under both monotonic and cyclic loadings. Following collection of quality data from these tests, the model proposed by Zhao and Sritharan was examined and found to be appropriate for modeling the bar stress vs. loaded-end slip relationship at the interface between column or wall and the foundation.

Through an analytical investigation combined with measured data along the embedded portion of the bar length, it was further found that the bond strength reduces as the reinforcing bar experienced inelastic strains. Using the suggestion of Wang (2008) that this reduction could be accounted for through a modification factor, an investigation was conducted by comparing the predicted bar stress vs. loaded-end slip relationship derived from this analytical investigation to the pullout test results. It was found that the modification factor proposed by Wang (2008) was useful in improving the global response. In this analysis process, the strain and local slip distributions along the bar embedment length were examined for when the bar was subjected to strains well above the yield strain. Significant local slip was found to occur along the embedment length over the portion of the rebar experiencing significant inelastic strains, which was consistent with the measured data.

Based on the completed study it is concluded that: 1) local bond-slip relation will be different for a reinforcing bar subjected inelastic strains than those found from bars subjected to elastic strains; 2) the existing local bond models may be modified with a factor such as that proposed by Wang (2008) to account for the effects of inelastic strains; and 3) strain penetration model widely of Zhao and Sritharan that is widely used in fiber-based analysis sufficient will sufficient captures the effects of strain penetration effects.